DIY EAR 834P TRLH Phono Stage - Build Guide / Parts List

[spaceistheplace]

I. Intro

Ok, here it goes. Do we have here the L75 of Phono Preamplifiers? If I could be so bold, maybe?

Sometimes it’s nice when life is made a bit easier, right? After all, we wouldn't be here without the epic post from Audiogon by JN. Since something along these lines was what I was personally looking for when I set out on this project, and I've come to believe this represents exceptional value and performance (seeming to be hallmarks of this forum), I figured why not blaze a little path for those coming after me.

The TRLH stands for Thorsten, Robert (user Renchkin)  & Lenco Heaven. I'm just the librarian... organizing it all a bit in a way that I feel will result in minimal casualties and pulsing headaches from reading page by page through the main thread. So, speaking of which, I should take a moment to thank all these people along with all the builders that have come along the way since and put their two cents in.

I guess what I mean by the L75 of phono stages is that like a Lenco, one feels that we have the ultimate starting point and most efforts are toward pushing the existing design further rather than stepping sideways to a different functional principle altogether. That's what I've been considering ever since listening to my first stock design of this phono stage.

By perfect I don't mean it's without any limitation, or that it is the absolute pinnacle of "straight wire with gain" phono playback, but again, much like the L75 Lenco it is engaging and listenable in a way that is so very rare... Lenco lovers will now be nodding their heads in agreement. For the Lenco, the common catchphrase is PRAT, for the 834p I do not have a clever acronym yet to refer to its strengths and I'm not even going to attempt it... any wordsmiths out there wanting to take a stab at it let me now. Suffice to say it's damn good, so why we just cut to the chase and build one?

If you are curious about measurements, scroll down to the last section. Since this is variable based upon your chosen build type, I'll stay away from making global performance statements. However, generally speaking it is very accurate in its response. Noise is low for a tube phono, and it can be improved upon further as you will see shortly.

After building two of these, making some mistakes along the way, and realizing a better path only after I’d gone too far….. I'm compiling what I’d do as a nice compromise between space / value / cost / etc. that will result in a truly outstanding phono preamplifier. If you’d like a bit of hand-holding amidst the staggering array of options discussed within the main thread, this is for you.

My general criteria:

• I made only one brand/type/series choice except when large cost savings could be achieved with other choices.
• Avoided choices with exorbitant Hi-Fi markups lacking a data sheet or measurable benefit in the application- no Teflon Caps and so forth.
• I stayed within a few parts series to minimize the difficulty in reviewing and purchasing all the parts- even if its a bit overkill in a certain location. This is especially important in that the available PCBs from China list value and wattage rather than an identifier, so being able to match wattage even when its overkill is helpful to a novice.
• I also tried to avoid including parts which would require a multitude of retailers... and chose parts that would be available more or less anywhere in the world from Farnell, Mouser, and possibly Tube Depot or Parts Connexion for North Americans and Hi-Fi Collective in UK / EU.
• I avoided parts from the far east or Russia, mostly because of shipping cost (it adds up!) and long wait times.
• I avoided anything NOS (especially advocating against NOS Tubes here if you can't test them yourself) because of the differences inherent between samples, high costs, dubious online sellers, etc..

So, I don’t consider each choice “the best” or "perfect" but very good, sensible, high quality options from my point of view. Obviously I'm not saying: This is what you should do. You wouldn't listen to me if I did anyhow. By all means take what you would like, and leave the rest. This is not the cheapest build, that would be just purchasing the full kit online and cobbling together some on-hand parts. But I believe from a performance perspective this is a keeper.

The total cost I believe with everything and including a little buffer for this and that is approx. $400 USD but it may vary based upon alterations (this is an organic document, which is being updated as more is revealed within the main thread... admittedly less nowadays as I think we've come about as far as it can go). Also, cost depends largely on things like casework choices, transformers, tubes, etc.

Also, before I forget:   WARNING & DISCLAIMER   DIY can be dangerous to your physical, emotional and spiritual health! If you are electrocuted, spontaneously combust, require an exorcism, are eaten by a shark or require past life regression therapy after reading this guide then it’s officially your problem. In the event you are instantly rendered brain dead by the exceptional musical qualities of this preamp upon first listening, we will do our best to notify your next of kin as well as raffle off your 834 within the Lencoheaven membership at the next meet. I will include some safety check information within the forthcoming photo build guide. Don't ignore it! All amusement aside, high voltage is a serious concern for personal safety. I highly encourage all those performing this build to read this excellent thread on high voltage safety. This is more in reference to ultra-high voltage, but many of the same principles still apply. Link Here

After I cover what I feel to be a solid baseline unit I'll add some information on dual chassis configurations and other options, upgrades and so on so no one feels left out, but I don't want to get ahead of myself.

If there is something I’ve missed, or something you feel is very awry, please bring it to my attention... posting it here in the main thread is fine.

Ok, so onto the main event:


II. Selecting Boards

• Buy the Douk board Preamp / PSU Combo: The half-populated version with resistors and some other assorted parts. Purchase Link
• Buy the Xuling Blank Board Set Purchase Link
• Optionally, use Robert's PSU in lieu of either of the above supplied PSUs Purchase Link
• I also have a limited number of stuffed boards available (~10) which will be discontinued once sold out due to time constraints. It takes after Romy The Cat and Thorsten's "End of Life" Phonostage and follows Thorsten's PSU recommendations. See here.

If you want a recommendation on the above, I'd use the Xuling Board and Robert's PSU if building from scratch.

I'd avoid the fully build Douk versions that are making the rounds. The configuration and quality leaves a lot to be desired, despite having an attractive cost.

III. Selecting A Chassis:

1. Buy the Pessante 3U Chassis with 300mm depth and 300mm Baseplate from diyaudio Purchase Link
For European customers you can purchase directly from HiFi2000 here, just make sure you get the same options stated below: Purchase Link

Note: It's drop shipped from HiFi2000 and has free shipping by FedEx for US Customers! Get it with steel covers and the baseplate (Must add this as extra). You won’t need the fully vented top. Get the 4mm front cover to avoid complicated drilling for attenuator and power switch. The 300mm chassis depth is just fine, however there is no harm if you want to go larger I suppose. I think the anti-vibration feet are worth the extra cost. It comes with pre-drilled for an IEC and Fuseholder at no additional cost. The one's I specify below in the parts list fit the cutouts. I think the baseplate with predrilled mounting holes is the best thing since sliced bread... I highly recommend it.

2. Separated Chassis: Really up to you and your location. These will fit the Douk PSU and transformer as well as have room for Robert's PSU: Link . The extra space in building with just the amp board in the pessante gives options for integrating loading switching, mc step ups, etc.

Note: If you must buy a separate enclosure for the power supply, I would still get the Pessante anyhow for the main amp board. It's always nice to not be cramped. Many have expressed that a one chassis build of this particular circuit gives no additional noise if laid out correctly, so substantial $$$ saved from umbilical parts and two chassis can be channeled toward higher quality parts on the boards. I have built both and stand by this recommendation for maximum ROI.

Of course, there are many chassis options out there. Do whatever feels right to you.

III. Selecting Parts

Edit: Unfortunately due to demands of life I'm no longer offering parts kits. I do have a limited number of stuffed boards on custom PCBs, see above for details. Due to this I've begun expanding parts list to use Parts Connection and Hi-Fi Collective as their stocks are more stable and being more general as lots of these have become out of stock over time.

Key:

Parts lists will follow:

Value, Rating, Brand, Series, Type | Part # | Qty

R = RIAA Part - double purchase quantity and match to 0.5%
BC = Bypass Capacitor, optional.
C = This part value may change or be omitted depending on your other choices.
E = This is a part you should have as an extra for later during testing.
Q = Purchase Quantity. If not specified, order 2.

III.A. Preamp Board Parts List

Note for all RIAA parts: This is where it counts! I bought multiples of each and used the two closest values here to get Thorsten's recommendation of better than 0.5% parts matching in the RIAA. The part list only shows 2 of each, so if you have a high resolution multimeter and would like to do the same you'll have to increase these quantities yourself. 4 of each should be sufficient to get an extremely precise match.

Purchase all parts in sets of two unless otherwise noted. Quantities in Purple.

• R - 270pf 500V CDE Silver Mica Capacitor | Part #: CD15FD271FO3F | Qty. 2
• R - 33pf 500V CDE Silver Mica Capacitor | Part #:  CD15ED330GO3F | Qty. 2
• R - 100pf 500V CDE Silver Mica Capacitor | Part #: CD15FD101FO3F | Qty. 2
• R - 787k 1/2W Resistor .1% 15ppm | Holco Part #: 279-H4787KDYA | Qty. 2
  
  • Note: Or 2x Each 680k and 110K resistors to get 790k. 1/4 or 1/2W ok here.) Note: If you want to splurge on a nice resistor this is the place… Rhopoint or Tepro or whatever floats your boat… To get even closer to the 790k ideal you can use the second option of two resistors in a series. These two resistors should be of the same type as they will be placed in a series to attain 790k.
• 1uf 400V Vishay Roederstein MKP-1839  | Part #: MKP1839510401
• .15uf 400V Sprague 715P | Part #: 715P15454LD3
• 100uf 16V Elna Silmic II | Part #: RFS-16V101MH3#5
• 1uf 160V Vishay Roederstein MKP1839 | Part #: MKP1839510164
• 47uf 400V Panasonic EE Series Electrolytic | Part #: EEU-ED2W470 | Qty: 6
• .47uf 400V Sprague 715P | Part #: 715P47454MD3 | Qty: 6
  
  • Note: Thorsten originally specified 100uf. If you want to go further you can try 6x Nichicon LKX 100uf  450V | Part#: LKX2G101MESY35 and 6x Panasonic 1uf 400V Part#: ECW-F4105HL
• O.1uf Panasonic | Part #: Film Capacitors 63V .1uF 10% Part #ECQ-E12104KF | Qty: 3
  
  • Note: Across tube heaters
• C 680k 1% Dale RN60 (2M Resistor Replacement if using an NOS tube in V2) | Part #: RN65D6803FB14
  
  • Note: 680k for NOS Tube in V2 -
• 2M 1% Vishay Dale CMF/RN60
  
• Note: For current production tube in V2

[/li]
[li]68k Dale CMF60 Resistor | Part #: CMF6068K000FHBF
[/li][li]Note: Only required if using a stepped attenuator[/li][/list][/li]
[li]100R Ohmite Little Demon Carbon Composition | Part#: OD101JE
[/li][li]Note: Thorsten Mod. Carbon composition, while usually not used in audio, it is important here for taming oscillations[/li][/list][/li]
[li]51k Vishay / Sfernice | Part #: RCMS0551001FHS14[/li]

[/list]


III.B. PSU Parts List

• Nichicon LKX 100uf  450V | Part#: LKX2G101MESY35 | Qty: 1
• Elna Silmic II  2200uf 16V | Part #: RFS-16V222MK9#5 | Qty: 3
• Panasonic EE 47uf 450V | Part #: EEU-EE2W470 | Qty: 3
• Nichicon LKX 330uf 450V | Part#: LKX2W331MESB45 | Qty: 1
• Choke Hammond | Part #: 154E | Qty: 1
• MUR4100E Diode | Part #: MUR4100EG | Qty: 4
• 1N5373 to tweak B+ Voltage | Part #: 1N5373BG | Qty: 4

** 10/10/18 Sorry parts list is presently incomplete I'll fill this out soon when I can check stocks of various retailers ****

III.C. Tubes:
After rolling a whole ton of tubes personally through the 834, these would be my recommendation. I've used the best mix of NOS and New production tubes I'm aware of at this time selected for each location. As I said before I would not trust NOS tubes - they're expensive and the quality can be variable unless from a trusted source- in which case they're even more expensive. If you don't have your own equipment to test and verify, I would avoid anything not new production. They are just as good, if not better, in my experience.

Note: From any vendor I would specify to "select for low noise and microphonics" even if the tube is already stated as a low noise/mic. tube. This is how tubes are selected for the commercial 834p: V1 selected for low microphonics, V2 selected for low noise, V3 "non-critical"

• V1: SPAX7A or Ruby 12AX7AC7 HG+
• V2: Tung Sol ECC803s
• V3: Ruby 12AT7ACZ HG  (Note: Or 12ax7 or 12au7 Ruby depending on your choice of cathode follower. I'd recommend 12at7.
• 3x Gold Ceramic Solder Type 9-pin Sockets | Purchase Link
  
  • Optional Tube Covers with PCB legs  | Purchase Link
  
  
  III.D. Transformers:
  Transformer (115V / 230V)
  
  
  • Antek AS-05T280 - 50VA 280V Transformer Purchase Link
  • Include the steel cover/can - Part #: CA-050
  
  Note: Antek very infrequently updates their website. Even if these items show out of stock, or are not there at all still email John at Antek directly as it's likely he will still have the transformer and cover in stock to sell or have one made for you, you just won't checkout through the website. Everyone who has attempted this seems to have been successful.
  
  Update: Antek now has these on eBay complete with metal can, making things easier. Search for the part #.
  
  Another US option is an EI Core from Hammond:
  Hammond 369ax: Datasheet Link
  
  International Options:
  
  To to shipping costs, feel free to look locally for a high quality Toroid or R-Core Transformer with the following specs: 6.3V 2A / 260-280V with Cover (Toroidal ideal for PS in same chassis, a steel cover is also recommended for toroids).
  
  Here are a few few options international builders have used with good results:
  
  • http://www.ampmaker.com/store/20W-toroid-power-transformer-PT06.html
  • http://www.ebay.co.uk/itm/301308565631
  
  
  III.E. Attenuators, Switches and Miscellaneous
  
  • Valab 50k Stepped Attenuator (optional for output level matching) | Purchase Link
  • Mono / Mute / Stereo switch (optional) | Purchase Link
  • Power Switch | Part #: HS13X
  • 6x Neutrik NF2D (1 set input, 1 set loading, 1 set output) | (3 x Part #: NF2D-B-2, 3 x Part #: NF2D-B-9 )
  • 1x Fuse Socket 6.3mmx32mm | Bulgin Part #: FX0415/S
  • 1x .5A Slow Blow Fuse 6.3mmx32mm | Part #: 0313.500HXP
  • 1x IEC Connector (screw type not snap in or integrated with power or fuse if using Pessante) | Schurter Part #: 6100.3100
  • These heat sinks are a perfect size match for the 6.3V rectifier: Link
  • Roll of copper tape for RFI Shielding the chassis: Link

III.F. Things you should already have

• MG Chemicals Heavy Duty Flux Remover or comparable
• Pack of ESD safe brushes for flux removal
• Assortment of heat shrink (get it to match your cable sizes)
• Assortment of wire in different colors stranded tinned for Power (Rated above 400V) Purchase Link
• Assortment of wire in different colors small for signal Purchase Link
  
  • Note: I used solid core mil specPTFE for signal Purchase Link , Stranded PTFE mil for power)
• Cable Ties + Sticky Cable Tie Bases (Local hardware store)
• m3 Nylon Standoff assortment with screws
• Assortment of small nuts and screws for the RCAs, IEC, etc.
• • Knobs of your choice
  • Multimeter rated 600V+
  • Stepped Drill Bit Set
  • Assortment of regular metal bits
  • Cutting Fluid
  • Soldering Iron
    
  • Note: If you are unfamiliar with proper soldering techniques, read this fun comic book guide! or watch some youtube videos before you get started
• High Quality Solder and Flux (Kester 44, Cardas, etc.)
• Soldapullt Desoldering Pump Link

[/list][/list][/list][/list][/list]

[spaceistheplace]

**** Part II: Main Board Assembly - in progress ****

2.1 Board Modifications

Component Modifications:

This is kind of a combination of Thorsten's original mods and Robert's further RIAA mods, but I will keep them together as it's all in the same board section.

Below is Robert's original notes with original RIAA values. He bridged the 0.1uf cap location with wire. To reduce likelihood of oscillation use a 100ohm 1/4W carbon composition resistor in this position. Other information worth noting is the resistor to be changed if you intend on using an ECC83 in V3. Thorsten advised a 12AU7 in this position to reduce overall gain.

Here are the different values in the 68k position for various tube options in V3 post Thorsten Mods:

12ax7 needed 1mA (120k 3W Replacement)
12at7 and 12au7 ideally 5mA (22k 3W Replacement)
12ay7 required 3mA (36k 3W Replacement)

Leaving the 68k for the au is only 2mA so a bit light and so while it was previously recommended to be left unchanged for 12AU7, follow the above for best results.

Note: After a while of experimenting back and forth with a 12ax7 and 12au7 I would personally leave a trio of 12ax7.

To quote Robert: "In green shows the caps that have the greatest effect on sound quality and good quality caps within your budget should be used here. In yellow shows the RIAA caps that should be silver mica and at least 1% to get a close flat RIAA response." More on the "green area" caps later.



Below is with Robert's optimized RIAA values and Thorsten RIAA mods fitted to the PCB.
Note: the red Sangamo is a 100pf silver mica, just not the dipped type as is common. If you purchase from the parts list you will have a CDE similar to the yellow ones.






Summary of changes:
The 680k and 110k resistors are in a series in the 750k resistor position.
The 270 and 33pf silver micas are wedged in the 330pf position in parallel.
The 110pf silver mica has been replaced with a 100pf silver mica.
The 2M and 3M3 resistors are removed entirely. They do not require bridging.
The 0.1uf capacitor is replaced by a 100 ohm 1/4W carbon comp. resistor.
The Elna Silmic 2 100uf 16V capacitors are placed in their positions- take note of polarity.
The second 2M resistor that was located above the 2.2K resistor has been replaced with a 680k resistor.

Note on the 2M replacement resistor: This area may effect bass response under 100hz with certain tubes. We are working on a solution to this, looking at more tube data and may have a recommendation of either an optimized value for the tube you are using, or a variable resistor which you can adjust if you do a lot of tube rolling.

The general consensus is:
680k for NOS tube examples
2M for new production chinese examples
3M3 for 6N2P-EV

A diagram utilizing the Xuling board (Click for full size):




Physical Modifications: Cutting the Xuling B+ / B- traces:

I'll quote member drat:

Along with the B-, the B+ is also connected to both sides on this (Xuling) board.  Both traces are easy to get at on the edge of the board between the high voltage terminal block connections.  B- trace is at the top of the board, B+ is at the bottom directly below. I just ended up using a dremel to cut out a small section containing both traces and then of course tested for continuity very carefully.

The second option is to use an exacto knife on each side and carefully cut out and peel off a portion of the trace for both B+ and B-. This is the more "surgical" approach, but reduces the likelihood that a connection will be made with the ground plane. Also, it keeps the board looking intact.

Either way, the key here is to make sure the cut is very clean and no shards are connecting other board layers. There's 285V there, I encourage you to be cautious and double check your work.

Here is a close up of the traces to disconnect:





2.2 Non-RIAA Capacitors

Main Films

Let's take a moment to consider what some options could be.

My personal experience is that using a high quality, sensibly selected capacitor it a critical location can make an audible difference. However, there's a place of depreciating returns in which you descend into madness thinking capacitors will solve all your woes and the money just gets stupid. I like focusing on areas that have tangible attributes that can be related to one another in a manner that isn't full of sociological pitfalls.... not because I don't think there's ambiguous, little understood phenomena that can be at play with various minute differences in capacitor constructions and types.... but because I'm just not going to let myself get led out into the desert on a horse with no name. I'd rather stay somewhere with identifiable goals and paths, like better understanding the circuit as a whole.

I'd stay within Polypropylene (Vishay MK1839, Sprague, Wima MKP10), Film Foil types (like Ampohm, etc.) and PIO (Mundorf, Russian stuff)- take care to not choose something massive which will have trouble being attached to the board (too large of coupling caps has also been known to create parasitic issues)... choose something reasonable, within your means, that doesn't exceed for example what you might have to spend on one tube. Tubes are an area were the investment is realized clearly, and where the possible margins of error are larger than that of a capacitor.

However, the main drawback of this approach is that your board will not look nearly as dazzling, and so your ability to win friends and influence people when they're looking inside your finely tuned phono chassis after a Saturday evening highball is severely diminished.   .....sad, I know.

Basically, pick whatever makes you feel good. That's what's important here.  

For those using capacitors with metal bodies, use an exacto knife or file to scrape a small area exposing bare copper, from which to run a small wire to ground the casing. Also, use some heat shrink around the cap after soldering the wire, or some teflon or electrical tape underneath to prevent the capacitor housing from making contact with the PCB Holes. Do the same with some heat shrink to keep the leads from making contact. Here's an example of how it might look at the 0.15uf location (notice, they are grounded to the spare negative pcb hole in the 33/47uf electrolytic location).



Thorsten Mods - Cathode Bypass Capacitors

Thorsten advised two different alternatives, based upon the space in your chassis and thickness of your wallet.

Option 1 is most appealing in terms of cost and space which is bypassing larger electrolytic with film caps. The reasoning is that a small value film can compensate for the limitations inherent in an electrolytic capacitor and deal with some of the high frequency "grain" that main associate with them.

If you'd like to read a more detailed discussion of this practice, see here

Note the bypass caps are assuming 47uf electrolytics on top... if you are using films in this location you can omit them.  We can go higher than these values in which case you will also adjust the bypass caps. The general rule of thumb from Morgan Jones' book is 1/100th of the value of the electrolytic... So, if you are using 100uf, then 1uf bypass is appropriate, etc.

Photo of underside of board with Electrolytic Bypass caps & caps across the tubes (explained in next section) affixed. Also worth noting in the photo is the area for the 68k resistors if you are using an attenuator.



Put them in parallel with the electrolytics (same pin locations) and then you can adjust them so they are perpendicular with the board to assist in clearance with your standoffs like so:



Option 2 is more expensive, takes up more space, but if you want to "go all the way" this is how you do it:

Like I said, films for anything other than the .15uf and 1uf film caps are going to be pretty large. Keep in mind you're going to have to find a way to mount these things! Figure that using 400v axial film types throughout the amp board you might need a full second amp boards worth of space to mount everything. Thorsten recommends this, he also recommended 2x33uf Ansar Supersound in each location, to be specific... which just isn't feasible for me space wise (and a bit crazy on cost) but worth noting. Also, he only had 2 large cap locations on his schematic, we have three on ours... I think 33uf x6 is about as far as you want to go for space, cost, etc.

Probably the cheapest solution for an axial at the values we need for main amp board are Audyn's available from Parts Express at ~ $9 ea
You can also try your luck with the PP types from mouser made by Vishay etc. but these are substantially more expensive than the above option and will be more difficult to mount.

For radial and mounted types you have the ASC and other Motor Runs (NOT motor start) - You can mount these on the top of the chassis.

I'm currently attempting this path with 6x Intertechnik / Audyn Q4 33uf 400V.

To give you an idea of how the cathode bypass films might look:



I'd be interested to hear anyone's thoughts who has successfully implemented some of these options and what the result was.

Also suggested for those who have to mount films a bit of distance from the board:

"The only thing I can think of when it come to the large film caps is there proximity to where they are needed. Their purpose is to act as reservoirs for transients and as a blocking filter between stages. The long leads add inductance and as such stop the transients from being seen by the caps. They still act as filers. A thought.. if they don't sound as good as you liked you could try to add some small films (.1uf/.47uf 400V) in the original cap holes. These would dump the transients not seen by the big caps."

I might attempt this as well, as I have some 1uf 400V films I had on hand from bypassing the electrolytics


Caps across the Heaters

This was a suggestion by forum member xdanx and I think it's worth putting within the main modifications. Also, it makes my photo make more sense / reduces confusion.

These were Panasonic film capacitors that I had on hand and so I used them and was quite pleased that they stayed in place rather rigidly... I've used Wimas as well and they were a pain with little leads and heat shrink ... The Panasonics are film caps, the bypasses are PP, the .1ufs are polyester... they benefit from the strong leads which allow for a much easier installation than the box style Wimas, so I said screw the Wimas.

Here's a photo which shows their orientation (same for all 3, only 2 shown)

[spaceistheplace]

****Part III: Power Supply - in progress****

Note: If you are considering using Robert's improved power supply when it becomes available, be sure to extend the area around the PSU board so that you can replace the original PCB with a 130mm x 130mm replacement. Also leave a little room on your transformer and PSU cabling to accommodate a possible different point of attachment

3.1 - B+ Assembly

Because I know I love a visual aid:



Double visual aid!




When placing / replacing diodes be sure that the marked line on the casing is on the same side as the tip of the arrow marking (purple and red stars). Also be sure they aren't laid flat again the board- give them room to breathe they get hot.

Hammond choke replaces the resistor that was in the position marked by the green rectangle like I said before. The choke is non-polar it can be connected any way you choose.

The electrolytic capacitor polarity however is essential to get right. This is marked on the boards as well but I've also marked them here with the blue hexagons. Note the side with the dashed line indicates the negative lead!

If you are using 1N5373 and 1N5374, their order is not relevant in the chain just their presence... most seem to require 4x 1N5373 to get under 300V. This is variable depending on what transformer you are using.

If these aren't already soldered to your board try 4x 1N5373 first and then measure (more on that later). If they're already soldered leave as is and measure. These are difficult to desolder and you are likely to mangle them getting them out if you haven't been through this before.

The largest cap is the reservoir cap (330uf nichicon) and the one to it's left is the 100uf cap. All of the other caps match the values printed on the PS circuit board (look underneath if you're having trouble!)

The transistor, although not picture, should be placed in the obvious location... the only place with 3 holes for the only part with three leads   The MJE13009 is no longer manufactured. The replacement is FJP13009. If you bought from the parts list you should have a heat sink and mounting kit as well. Attach this BEFORE soldering to the chassis. The heat sink should be on the side closest to the reservoir capacitor.

I know my board is looking a bit rough around the edges... apologies... don't mind the errant piece of zip tie or assorted other muck / gunk.....  

Always use two wires for + and two wires for - as laid out on the PCB. Why, you might ask?

"Signal flowing through a channel of the amp causes currents of the same frequencies to be drawn from the B+ and dumped into the 0V. The capacitors in the supply filter out most of these signal currents but what is left particular transients that the caps cannot filter can get onto the other channels supply and reduces separation of the channels and loss of openness and clarity. If seperate power and rertun wires are used then the currents can only mix at the point of joining at the lowest impedance points and closest to the PSU filter."

This is an important consideration if you are going for a two chassis build, as the connectors can be pricey and many attempt to go for a 5 pin plug instead of a 7. Don't do it, imho. More on that later.

3.2 - 6.3V Section Assembly

If you have the room I'd highly recommend bolting the 6.3V Bridge Rectifier directly to the baseplate. Mount it upside down (NOT backwards) so that it travels beneath the PS board and the leads form a 90 degree angle up to the pcb holes, then use a screw and nut to affix it . Do not solder the rectifier until AFTER this is complete, to ensure a snug fit. Also, make absolutely sure no portion of the rectifier leads are touching the chassis. This configuration will minimize heat flowing right onto those resistors nearby. Other options include heatsinks (see parts list above), or just rolling with it. I worry about the proximity of the bridge to the resistors there, and could imagine the resistor value drifting from heat but I don't have the math to back it up, however I noticed some minor fluctuation/instability in my own and my own hunch is pointing toward the rectifier position and the fact that I live in a very hot climate.

Bend the rectifier to point up, the totally flat side down. Use a piece of painters tape to mark the hole location like so:



Make sure the screw and nut of the baseplate can move freely. Then make your hole and affix with an m3 screw and nut.




Solder and snip the excess only after you are sure you have a snug fit and that no piece of the rectifier leads are touching the chassis.

You might feel the urge to affix the bridge to the baseplate with thermal adhesive... the specified rectifier does not get hot enough to really require this, a tight connection to the plate is more than enough. Mostly, this is to keep the heat off of the 6.3V resistors. If you really feel the need, don't do so until after everything is done, because the rectifier will likely be permanently adhered or be destroyed if you can manage to remove it.

Don't wire up the DC from PSU to Amp board yet we are going to measure and make sure we are in a general range.

This is also good as if there is a problem you won't torch any parts on the amplifier board even if you've done something really, really silly.

3.3 - Wiring it up, Safety Check, Determining voltages etc.

This will be a preliminary check of the power supply so this does not need to be a final configuration within the chassis. Only concerned with getting the PS wired up safely.

Follow your Transformer Datasheet to determine leads for each voltage, wrap them tightly and wire them to the AC inputs. Make sure any extra leads are tucked away and there is no exposed conductor. The best way to do this is fold it back and cover it with heat shrink. Wire the transformer positive and neutral to the IEC outlet.

For the IEC outlet ground: The IEC connector MUST connect directly from the ground prong to the chassis using a preferably thick piece of wire and stay there PERMANENTLY. No ifs, ands or buts. Your safety depends on this. All other possible grounding schemes must follow AFTER this is taken care of.

Link the transformer static shield (if fitted, some transformers may not have this) to this location as well as a wire from the point labeled GND on the power supply to this common chassis ground location. Now we can plug in and do a voltage check. We do not need the power switch yet, we will desolder the positive lead and run the power switch later.

Here is an example wiring diagram for the AS-05T280 (Datasheet):http://www.antekinc.com/content/AS-05T280.pdf

This is assuming a single fuse and a filtered socket (like this ) so I will no address X or Y caps or multi-fuse arrangements.

The full 4A is not required for 6.3V (total current is 0.9 x 1.6 = 1.4A so 2A is plenty) but the extra buffer cannot hurt and makes for a cleaner wiring.

260V is used instead of 280V as extra would just be wasted heat, and is sufficient to give us the 285 we are shooting for at output. 250V could also be used.

Note that the actual colors on the second 6.3V were Brown and Orange, not blue and green as indicated on the data sheet photo.

For 115V Countries:



For 230V Countries:




Wiring Filament Voltage to the Xuling Board:

The Douk is only configured for 6.3V, however the Xuling accommodates both.

Most will stick with 6.3V @2A with the various supplies we've been using, however the wiring to the board is slightly different. If using 12.6V you'll require half the current. (click for full size)



This is what mine looks like bridged for 6.3V. I used a spare piece of wire taken off the 6.3V transformer secondaries.



If you don't bridge pins 4/5 you will only be heating one side of the tubes and giving you only one functional channel.

Before Powering on for the first time

The two most useful settings are measuring resistance (and on buzzer range) and voltage ranges. (The resistance measurements must be made with a fully discharged pcb if the unit has been powered and discharge can take a long time. So if you have powered up the unit before reading this, power down and check voltages are 0 before using ohms.)

Before powering for first time check that there is no shorts on the supplies. Connect the multimeter in ohms range (or buzz range) and check accoss the 6.3 and B+ wires. You will see a low resistance or even a buzz for a second or so while the big caps are charged a little by the meter, but that should go away and reach several kohms heading to a meg or so (valves not fitted when checking 6.3) also check the output terminals they should be open (very high resistance)

Check chassis and any other external surfaces show 0 ohms to the mains earth for safety

If all is good connect power and do a quick power up and down. No bangs or obvious over heating then it's good to power again for longer. Without touching any components just check for any possible overheat.

Voltage Check

Make sure you have a 600V+ rated multimeter! 250V doesn't cut it!

Plug in the power supply

Measure DC Voltage negative lead to negative B+, positive lead to positive B+ and note figure. Do the same, one at each 6.3V terminal.

At the 6.3V terminal you will have a bit higher than expected since you are measuring without the valves fitted. When I checked mine I had 8.36V, then with valves fitted it was almost a perfect 6.3V

These with valves fitted should be less than 300v and ~7v respectively, however we'd like to get closer to ideal if we can.  Without the amp board connected the values will be different. As long as you are somewhere close to those numbers and nothing has exploded you have succeeded.

Turn off the power supply and give it some time for the caps to discharge before handling (the LED will turn off when caps are fully discharged). Don' get nervous if the LED stays on for quite a while after you turn off, as without the valve section attached the LED has the full responsibility of discharging the caps. With the valve section attached the LED will turn off instantaneously.

Don't proceed to the 6.3V adjustments section until after we have completed the chassis and amp board connection with valves fitted.

If you are having issues look at the troubleshooting section.

3.4 Adjustments

Adjusting B+:

Anywhere between 280-300V is ok. If it is in that range leave as is.

To drop B+ Add one 1N5373 and check, To increase B+ add 1N5374 and check, and add more if necessary.

If you have the light blue board, the zener adjustments should be printed on it:



To give you an example I was (if I remember correctly) ~310V and replaced 2 1N5374s with 1N5373's and came down to about 285.

Adjusting 6.3V Voltage: Wait until you power up the board with the valves attached before going through this step!

Allow the unit to warm up and measure DC voltage at the 6.3V inputs on the amplifier board. If over time the voltage becomes established then you can use this formulae to adjust the 0.22R resistors in the 6.3V section of the power supply.
     Radd = (Vmeasured - 6.3) / 0.9

Then add that amount of resistance to one of the 0.22R or split it and change both. Purchase whichever resistors you need... I'd go for 3W-5W rated due to the heat in this area from the rectifier.. just to be safe.

3.5 - Troubleshooting

Troubleshooting B+

I. This will require a systematic approach starting with resistance checks
WARNING doing resistance tests requires the caps to be fully discharged this can take a long time (> 10 mins) and checked by measuring the volts on the B+ and 6.3v outputs are (<0.01V).

1. Disconnect the PSU from the amp
2. Recheck the resistance at the 6.3 output. It will start up low as the meter charges up the large output cap but should rise to a high value > 100k
3. Do the same for B+. If the resistance does not go up then there is a fault
4. Next measure resistance of the 6.3v on the amp board it should be about 1 ohms when cold
5. Reconnect the amp board.
6. Power up and note the 6.3V with the meter. Also measure one pin of the 6.3V to gnd about 35V depending voltage from transformer. The current can be calculated as VPSU / 7 (aprox when hot)  and should be about an amp.
7. Check the heat in the diodes.
8. Carefully measure the AC volts across each diode of the bridge it should be about a volt if higher then the diodes are poor and will increase the heat in the bridge if you measure 3.5 or 7 then you are measuring across more than 1 diode and seeing the load. The power dissipated in the diodes can be calculated as the sum of the power in each diode (Vdiode * Icalculated) . The best way is with the PSU running then you have a totem pole of 4 zeners. the bottom of the totem-pole is connected to GND it is very noticable as its connected to the large GND plane. if you measure the voltage connecting the - probe to ground and the + probe to the marked (pointy end of the diode Kathode) you should see a voltage of about 70V, if you move the voltmeter + probe to the marked end of the next zener it should read about 140V, moving on again 210V and finally to the last and it should be about 280V. If any measurement jumps higher than expected you can measure that zener directly moving the - probe to the non marked end and the + to the marked end that will give the breakdown voltage of that diode.

It is not easy to measure these diodes out of circuit as it would require a voltage source > 70V to get a breakdown, a meter diode setting will not do that, so these measurements have to be done live.

Shorts
Some when separating PSU from the AMP pcb have created a short by a less than totally clean break.. can be removed with a file along the edge
To check remove the gnd wire between the boards and measure resistance between gnd and 0V on the AMP it should be open this will also check that there is not a short to the chassis

[spaceistheplace]

****Part IV: Chassis, wiring and Final Assembly- in progress****


4.1 - General Mounting Practice

Not relevant if you purchased the isolated neutriks in parts list but very relevant if you purchased others. Either way important to know as this is fairly standard practice so we can keep our grounds just the way we want them, rather than everything getting linked by chassis by default.



One of the plastic rings has a little collar. It is the one which is already stuck onto the connector in the photo above. This one stays on the connector when it is inserted into the plate, so it's visible from the outside. The hole in the plate needs to be big enough so that this collar extends into it. This way it is ensured that the RCA connector is isolated from the metal plate which is important to avoid ground loops. Then the second plastic ring is put on. This ring isolates the connector from the under side of the plate. Then everything gets tightened to the plate with the first of the two screws. The ring for the ground connection is placed between the two screws.



It is important that the screws are secured very tight to avoid that the connector loosens over time. Many people have the habit of twisting the connector when they insert or remove the interconnect. This can loosen the screws over time. To make sure the screws stay in their place after they are tightened, they are secured with some solder. This should be applied such that it covers the ground part of the connector, the screw and the ground ring. This way it will stay in place even if the cable is twisted a bit and a secure ground connection is made. Be careful to not shorten out the ground part with the inner pin for the hot side of the signal.




4.2 - Mounting the Neutrik RCAs

Firstly, before measuring, note that the chassis top has a small lip!! If you don't take this into account (like me), you might mount the Neutriks too high and cause an issue closing the top... and then have to shave down the cover for it to fit.

Put the Neutriks in and tighten the outer screws. Then, loosen the inner screws to allow for the shield tab to be adjusted for easier wiring.  Speaking of which they don't come with the outer screws. I used #4-40 x 3/8ths in and they fit perfectly. Double check up close when your done to make sure a shield wire that is stuck through isn't rubbing again the back of the chassis or touching one of the Neutrik machine screws. Do not link any shields or tie any to your ground post. After tightening the screws and having everything in place, solder across the screws like in the standard RCA photo above to make sure these stay secure over time. Install the ground post similarly to the RCA photos above, taking care to leave enough room for the inserts and to have the post making contact with the chassis. I always additionally wrap a bit of plumber's teflon tape or electrical tape where the post goes through the chassis plate.  Like so:



Keep the loading resistor jacks as close as possible to the input jacks. We will be setting up our loading resistors inside RCA plugs so that we can quickly swap out depending upon the cartridge we are using.

Here are my loading plugs for example:



One you are done, place a piece of heat shrink and lightly close it up. Note: I generally prefer the RCAs with LESS metal a-la the Eichmann Bullet Plugs, but I had these on hand. You can also get Neutriks with similar construction from Mouser which will ensure a snug fit with your jacks. I just used what I had on hand.

This is a 1k loading resistor pair for the Denon DL-110. You can find further information on what you might need for your cartridge in the further reading section below.

Also, using one of those label makers, some label stickers or even a piece of painters tape its helpful label what value each RCA plug is to avoid pulling out the multimeter later.

4.3 - Routing the power supply cabling through the baseplate:

Keeps everything nice and tidy and the power away from signal cables. Leave extra length just in case until you are totally done and then come back and make everything snug. I had much less room to work with as I made the mistake of getting the 2u chassis and had 6 giant poly caps to wedge in there, so you just might be fine routing it along the normally. Either way make use of the baseplate holes and use some zip ties- but don't bundle all the cabling together- each on its own track. I'd also recommend using heatshrink for at least the areas traveling through the baseplate holes to avoid scraping and abrasion.. those holes can still be quite sharp.

Note this picture is before I changed the B+ neg lines to PSU 0V (see grounding section below)





4.4 - Grounding

It's a lot of text so it will look small here to not break forum image rules. Click the image to see full size.

[spaceistheplace]

****Part V: Pushing things further: More alternatives and additional doodads ad infinitum****

5.1 - Cartridge Types and Optimizations

Additional Reading for Cartridge Loading:
http://www.lencoheaven.net/forum/index.php?topic=15147.0
http://www.lencoheaven.net/forum/index.php?topic=18231.0
http://sound.whsites.net/articles/cartridge-loading.html
http://www.hagtech.com/loading.html

MC Step Ups
http://www.rothwellaudioproducts.co.uk/html/mc_step-up_transformers_explai.html
http://www.kandkaudio.com/applications-guide/

Suitable MC Step Ups

For good value new production step ups to integrate into a chassis ( does not come plug and play) there are two / three very good options:

1. Cinemag Link here - About $180 for the pair of 3440AH

2. Lundahl Link Here - Basic kit is $325

3. Silk Audio - Supermalloy SQ Annealed Core MC Step Up Transformers - Link here $260-300

These have caught my attention and come from the east (Thailand?). If I did not already have a step up I'd spring for these. I would think they are a better value than the Lundahls due to their location of production. "It can be wired as 1:10 or 1:20 with guaranteed better FR response than even Ortofon T-3000 (+$1,600USD). beautiful stainless 5x7cm can with mounting brackets." Further reading in a pairing with a tube phono preamp from Troels Gravesen - Link here


5.2 Tube Rolling & Tubes of Note

Again, tube objectives for each location: V1: Low microphony, V2: low noise, V3 is non-critical.

If you are interested in reading my complete findings on this topic, most is contained here in the main thread. I've kept things more brief here to save room.

Main themes are:

1. New stock tubes have horrible quality control, so it's not necessarily about the production factory, but what third party is buying from the factory and what they are willing to keep vs reject. There are however some fine examples of new production tubes which I have now recommended within the parts list above.

2. NOS tubes generally test better, but can also be variable and have issues from mishandling over the years. Do your homework, and if you can, spend a little more to have an expert select one particularly for your use case. There are also microphonic, noisy NOS tubes.. just because it's NOS doesn't mean it's great.

3. Either way, it is essential to purchase tubes from a reputable seller who has a quality testing apparatus to get the best performance for RIAA applications.

A note on tube types: I've found that using different tubes with different gain levels and operating parameters may appear to sound better on first listen, but without taking pains to make circuit modifications to accommodate these different tubes, we are really just listening to something akin to a "warp in a record"... you may like the wobbly sound a few beers in but that's not the point of an RIAA stage... adjust color later with a tone control if you need to. So, let's brush 12au7, 12at7, 12ay7, 5751 and so on right off the table from the beginning... better to find the right trio of 12ax7 first before messing with the design. 12ax7 would be the family that includes ECC83, 7025, 6681, cv4004... same gain factor, different assemblies and special use cases.

I will say before I begin that NOS tubes are not just a marketing / fashion thing- they still routinely measure better than anything manufactured anywhere in the world today regardless of price! (Even inexpensive examples- i don't mean just $$$ telefunkens). At first I was not convinced of this, and felt that perhaps NOS tubes were just a gimmick like many things "audiophile". However, once I started experimenting myself I was quickly eating humble pie and looking at NOS tubes.

New Production 12AX7 Options:

An analysis of New Production Tubes by an expert with large sample sizes has shown that new stock has a much higher reject rate for high fidelity applications. This makes it essential to use a reputable vendor and get tubes selected for Low Noise and Microphonics.

Here's one person's findings which has been the highest quality technical resource I've been able to find on new production tubes and how they stack up:

The big kahuna: http://la-economy.blogspot.com/2013/05/current-production-12ax7-tubes-ten-of.html

And a post from 2012: http://la-economy.blogspot.com/2012/03/vacuum-tubes-today.html

And a selected chart:



If you are looking for a high-quality, low cost alternative to the 12ax7 then the 6N2P-EV is the ticket:

6N2P-EV: The Russian 6N2P-EV is a military tube with almost identical performance characteristics to the 12ax7. The main appeal is their substantially lower cost (about $4 shipped from Russia) to NOS 12ax7s. I have not used them but the cost savings on tubes is deamatic enough to beg inclusion here. The main caveat is that is will require some modifications to the Douk PCB and going back to 12ax7 will require new tube sockets (you'll have to clip the 9th pin and run it to ground). The other change is the grid resistor that we changed before from 2M to 680k earlier. Now this value will be 3.3M. An RN60 will be fine. Robert ran a simulation with the 6N2P and it seems everything checks out. It will just be 1db lower gain. As said before it will require a modification to V2 grid resistor for most accurate bass performance under 100hz. Douk provides a 2M in this location, for 6N2P there should be a 3.3M there.

Here are the adjustments to the pinout:






5.3 - Two Chassis Solution

Approx. dimensions of separated PS Board: 80x130mm
Appox. dimensions of separated Amp board: 130mmx200mm

You want at least 90mm height if you are going to use an EI type transformer like the hammond.

Total height of amp board with standoffs and tubes installed approx. ~70mm
Douk Boxes from Ebay ~ W260×H90×D311mm

These will fit both PS and Amp board: Link

Regardless of choice, make sure you get one with the IEC cutout only, and one with the RCA cutout only, not the two in one. Keep in mind the douk faceplates are nice and thick and will require some clever drilling for mounting of switches, attenuators and so forth.

Umbilical Wire: 7 conductor, rated for 400V min.

Umbilical Connectors:

Why go through all the trouble and try to sneak by with 5 instead of 7 wires? Silly idea imho.... also, don't mess with safety and proper ratings!! Do NOT cheap out on something that can't run 400V+ safely. If you are wondering what the benefit is of running 7 versus 5 and won't take our word for it, review the explanation in the B+ section of the PSU Assembly portion of the guide. I've found and evaluated myself a much less expensive (than the previously recommended mil amphenols) solution that handles this voltage safely in 7 pin configuration. You can find 7-Pin mil connectors (M+F Pairs) for $9 USD each pair with free shipping from China. $18 vs $70 is a no brainer. Link Here

Here is them in a better photo in use (green locking connectors) on a different tube phono stage (the PH16)




Here's a couple nice photos of two box solutions from the original thread:







5.4 - Other RIAA Capacitor Types - Air Gap / Vacuum Capacitors

Micas seem to be the only logical in the values we need for the RIAA. Thorsten likes them very much at higher voltages. Some people criticize them. Also, some people are just obnoxious and looking to be critical, so there's that. The dissatisfaction is based on a commonly held belief is that dielectric absorption is a main contributor to the quality of sound one gets from a capacitor. Mica falls behind Polystyrene, PP, and Teflon, however try finding 1% Polypropylene 100pf capacitors... Polystyrene you can find, however they will apparently melt at high temperatures and other issues that make them not suitable for tube amps. No one seems to have much respect for ceramics, and while the new ones are touted as having excellent characteristics, for our purposes in the RIAA I think Mica is the only sensible choice as it is very stable at high temperatures. No one is stopping you from trying alternatives. If you do, stop in and let us know how it went.

Oh, right.... this is about air caps.... SO, there's also something called air variable capacitors which are the theoretical ideal as a dielectric, and they come in the values we need!! BUT, like all the greatest of things- they are really huge and some can be really expensive.

Also, you will need a multimeter capable of that kind of find resolution at very low values. If you are currently wondering whether yours does, the likelihood is it does not... womp womp.... but still check anyway! It's uncommon for meters to go below 200pf with any kind of accuracy and it's usually a feature of expensive units. The DE-5000 I think is a reasonably priced unit that goes that low. I don't think the old ham radio type air gaps are as stable with their values over long periods as the vacuum types, and from what I've read on the new ones on their data sheets they can vary with temperature and humidity fluctuations... So, I think i would need something sealed in my climate? I can't say definitively, this is really just thinking out loud / speculation. If you have more experience on this I'd love to hear it. Also, all the new stock "air trimmers" seem to be "mica dielectric" and "PTFE dielectric" anyhow? So, you'll probably be stuck with NOS examples. Or, if your REALLY want to impress lots of people on the internet and have some time to burn, you can build your own Air Variable Caps Guide Here . Maybe I will get one and put it inside my build just to keep it toasty and see if I notice any changes... until then, I think it's wise to stick with the Micas unless you are after that last fraction of a percent- but those kinds of people already new that / didn't need me to tell them and have already gone out in search.

Romy (from original mod thread) thinks it's the ticket and seemed to not mind the air types. Romy also knows where he wants to go and doesn't mind building a gargantuan chassis.... His goal is approaching what he feels is perfection and he has the added flexibility of arrangement since doing a point to point build. I admire his perseverance and "code" and i'd like to try that maybe down the road, but for me not now I don't have the space in the current chassis.

Here's what he did at first with the open air gaps:



And then the vacuum types he settled on:




5.5 Other mods

This is a nice compromise before going to a film cap solution: Here, member xdanx used 330uf caps up front (this cap can be any size, really) and 47uf caps following. He also bypassed the Elna Silmic II cap with a 1uf Mundorf. I'm not sure exactly how that would play out in terms of audible effect or cost/benefit as I have not tried it. Many feel that bypassing electrolytics removes some of their limitations. This is a contentious topic, so I'll let you be the judge for yourself in your own system.

In the gold standard book on tube amplifiers by Morgan Jones, he believed in this practice and suggests a 1/100th value of the electrolytic. This would equate to .47uf, 3.3uf for the big guys (both rated 450V+. For the Silmic, a 1uf 63V will do fine.

His build pics (Click for full size)

[spaceistheplace]

******Original Circuit Full Description ********


Ear Notes




Roberts circuit explanation:


Below is the schematic of the standard build of the 834 amp as used by Douk and others (IE not the Thorsten or other mods) and I will use this to give a rundown of the design keeping as simple as possible and no math!



The design is in three stages each using a single valve (my schematic uses U1-3, instead of V1-3 but are the same thing) and a RIAA filter in the feedback (known as an active RIAA filter)

Stage 1

This consists of:

    The input / grid resistor R9 (standard 47K loading for MM carts)
    U1(V1) a ECC83 (12AX7)
    The cathode resistor R4 (2.2K)
    The anode resistor R1 (330K)
    The AC coupling capacitor C3 (0.15uf / 150nf)
    And shared with this stage the U2(V2) grid resistor R8 (2M)

The purpose of this stage is as a voltage amplifier, the coupling cap C3 and the second stage grid resistor R8 form a high pass filter and is part responsible for the bass roll off below 20Hz (becomes more important when the Thorsten mod removes the rumble filter) Now the hard bit!!

U1 has a flow of current from its anode to its cathode due to the grid to cathode voltage and the anode voltage, this current develops a voltage across R4 raising the voltage of the cathode above gnd by about 1V. This sets the grid voltage at about 1V negative, and that provides an anode to cathode current. This same current flow through R1 setting the anode voltage, the combination of the grid voltage and the anode voltage defines the absolute current though the valve and as each effects each other it is a form of negative feedback, stabilising together to produce the bias. Signal voltages are also effected by this feedback so limiting the overall gain of the stage.

The voltage gain is dependent on the amount of current change for the amount of grid voltage change and hence the change of voltage across R1. The larger the current change the larger voltage change. This current is purely dependent on the valve and each valve will have an amplification factor in the case of the ECC83 this is high at near 100, and the actual voltage gain depending on the chosen R1 and R4 values.

For this stage a 4mV cart voltage is amplified to 185mV or 46 times

Stage 2

This consists of:
    The grid resistor R8 (2M)
    U2(V2) a ECC83 (12AX7)
    The cathode resistor R5 (2.2K)
    The anode resistor R2 (330K)
    The cathode bypass cap C8 (220u)
    An attenuator made from R14 (2M) and R7 (3M) and a bypass capacitor C4. R7 is shared with stage 3 as U3(V3) grid resistor.

Again this is a voltage gain stage with similar bias conditions to the first stage, but this time there is the extra C8 across R5 and a potential higher anode voltage due a higher supply voltage. This pushes the gain of the stage to 73 times. How? well I mentioned that there was feedback between the cathode to grid voltage and the anode voltage in the first stage, and this is still true for DC (bias) but for AC (C8 chosen to pass all frequencies above a few hertz) all the current that is due to signal is dump to gnd and as such does not get fed back negatively to the anode voltage so greatly increasing the gain of the stage.

R14, R7 and C4 form a high pass filter where at low frequencies the output of U2 is attenuated by 2/5 and at higher frequencies the whole of R14 is bypassed and so no attenuation. This forms the main source of roll off at the low end and acts as a rumble filter. [The Thorsten mod is to remove this filter and join the output of U2 directly to U3]

Stage 3

    The grid resistor R7 (3M)
    U3(V3) a ECC83 (12AX7)
    The cathode resistor R6 (68K)
    The anode resistor R3 (100R)
    The AC coupling capacitor C5 (1uf)
    And the load resistor resistor R13

This stage is different to the previous two in that it is not a voltage amplification stage but a cathode follower (ignore the very low anode resistor R3). The current generated by the cathode bias flows directly from the supply through the cathode resistor R6, this produces a massive negative feedback back to the grid voltage level so much so that the gain of this stage is slightly less than 1. Also as R6 is large, then the voltage across it is large and so the cathode voltage is high (74V) creating a high grid/cathode voltage -2V that with the high anode voltage due to no real anode resistor, in turn produces the high current (2.5mA). The signal voltage presented to the grid modulates this high current producing a high current output signal. This translates to low impedance as virtually no current on the grid gives a high current on the output (high impedance -> low impedance)

As the output voltage has a high DC component C5 is used to AC couple to the amp output. R13 forms another high pass filter with C5 but again its way down in the small hertz. R13 is mainly there to discharge C5 to gnd when the power goes off.

The RIAA circuit

This consists of:
    C7 (110pf) R10 (750k) and C6 (330pf)

These values are chosen to provide the 3 poles of the RIAA filter at 3180us, 318us and 75us the effect is to reduce the gain at the high frequencies by 20db and increase it at the low frequencies. by 20db with 0db @1khz hence the gain of the amp and the cart are always at 1Khz as this is the frequency where RIAA gain is zero. The result depending on these values should be a flat response 20-20,000 Hertz.

The factors that effect the RIAA beyond that of the RIAA components are impedance of the driving and loading circuits. In a perfect world the load on the RIAA would be infinite impedance, no resistance, capacitance or inductance, and the drive would be able to supply an infinite amount of current (zero impedance) any non perfection effects the circuit. The high grid impedance offers the best load and the low impedance of U3 output provides the best drive.

Finally the power supply.
R15 with C9 provides the first supply and powers U3 with almost the full 284V. C9 acts as a DC reservoir supplying local demand and topped up by the main PSU PCB, it also dumps signal currents that are caused by the current demand changes with signal dynamics.

R12 and C2 form another supply to power U2 this time the voltage has dropped to 265 and again C2 is used for local demand and signal dumping.

Finally R11 and C1 forms the last supply as clean as possible and not effected from the demands of stages 2 and 3 at 200V. As before C1 does the job of local demand and signal dumping.

This chain keeps the supply to each valve as low impedance as possible, as rock solid as possible and filter out as much noise (unwanted signal) as possible adding to a cleaner sound.



*******Measurements*****

RIAA:

Robert's comments:

"With the RIAA values recommended in the guide a quick sweep of frequencies between 20-20K shows it to be flat within 0.5db. That is well within measurement error from a fast sweep.  I will be doing a slow sweep over the weekend where I can compensate for the errors and hone in. I expecting it to be within 0.1db between 100hz and 20k and depending on valve also flat below 100hz down to 20

Oh and it sounds really good too, huge detail and air between everything. And this is before the new PSU"



*******************
Some space reserved for my personal notes / reminders:

Panasonic EG Metal Oxide 3W or Vishay PR03 for amp board?

Find appropriate TO-220 heat sink and mounting kit #s for MJE13009.

[hatehifi]

Hi John, just curious as to basic ball park cost. I take it this is for MM? The Internet addresses would be convenient. Thanks!

Cheers!

[Chris65]

Hi John, just curious as to basic ball park cost. I take it this is for MM? The Internet addresses would be convenient.

Have you somehow missed the 50+ pages of the main DIY EAR834p thread John? All the information is there.
http://www.lencoheaven.net/forum/index.php?topic=20984.0
Search that auction site for Douk Audio Phono. Cost? Around €200 or so would give you a basic set, & on up from there.

MM only. Extremely difficult to implement an MC phono stage with tubes due to the gain required.

You've essentially had one of these before (the so-called one-off phono) .

[willbewill]

Jay's build

http://www.lencoheaven.net/forum/index.php?topic=25054.0

David's build

http://www.lencoheaven.net/forum/index.php?topic=22195.0
http://www.lencoheaven.net/forum/index.php?topic=23945.0

netphreak's component questions

http://www.lencoheaven.net/forum/index.php?topic=25494.0

[hatehifi]

Ay, mates, so right you are and one-off sure did sound good, these better I assume. Thanks for the clarification and please pardon the ignorance.     

Cheers!

[ZZMoko]

No worries a comprehensive guide is well worth while...

Just wondering why you need a stepped attenuator...this is supposed to be a phono stage ( a fixed line output device) which should feed a pre or integrated amp at the right level...I know some others seem to think it's a Pre amp...it's not it's supposed to do just one job and that's to take a tiny phono signal from the cartridge and amplify it to the correct level for a Pre / Intergrated amp.

[spaceistheplace]

Moko,

What level is "correct" is dependent upon the other devices you use, the gain of your preamplifier and it's relationship to your amplifier and speakers.

For example, unless you have matching preamps, power amps and speakers from the same maker/series, or have considered this matching in advance, the likelihood is high that these components are not ideally matched (in most cases they are close enough and not problematic, for example efficient speakers and class A low wattage amplifiers etc.).

The stepped attenuator provides the ability to dial in what's most appropriate by adjusting output resistors to what is optimal for your system.

The 834 naturally has very high gain. Perhaps this is part of it's magic.

It's common to use a 12AU7 in v3 to reduce that a bit, but even still many have issues.

The commercial 834 had an attenuator presumably for a few reasons:

1. It could be used to directly drive an amplifier without any other additional preampfification if one was using just one source.

2. When connected with a line stage it's output could be adjusted to better match the signals of the other connected units. This is so when switching for example from phono to another source or vice versa you don't go from quiet background music to skull splitting.

3. The 834 originally incorporated step up transformers, and presumably many (myself included) would like to incorporate  MM/MC switching. Since cartridge outputs vary and step up configurations / gain boosts vary, there is no supremely "right" answer for what the level should be when it leaves the 834. So, you may find that a 103R requires 20db gain, but this forces you to keep your line stage volume at somewhere around 7 o'clock if you don't want to blow the house down.

"Volume" is not a good way to think about it.

There's gain, and then there's attenuation which are two different concepts. Adjusting gain can be more difficult and changes the behavior of the circuit. Attenuation is simple and straightforward with a stepped attenuator at the output.

You can absolutely refrain from including one if you desire, but you may later find you have a lack of range and control at your line stage. If you notice this is a recurring theme in the main thread for many. I wanted to avoid others building this unit and then running into this problem / frustration.

Since we are dealing with an infinite number of potential systems for this to be utilized in, I think it's a wise choice. Moreover, a stepped attenuator avoids the issues inherent in carbon pots and the like. After finding the "ideal" level for your system, you can also remove the attenuator from the signal path and directly solder resistors that give you the best results. However, I question if there will be a clear audible difference between having the attenuator in the chain vs directly soldered resistors.

Gain and level matching is a complicated and multifaceted topic, I've read quite a bit and still much of it does not register well.

So, reason for inclusion in short was for simplicity over complexity and to make it easier to create synergy with the 834 and the rest of one's setup.

Hope that helps.

[spaceistheplace]

Hi John, just curious as to basic ball park cost. I take it this is for MM? The Internet addresses would be convenient. Thanks!

Cheers!

Yes this is for MM, but if you so desire you can include a step up transformer in the build to make it MC compatible.

I will be uploading a parts list in excel form when it's done, as well as links... but still a bit more fiddling to do. Still trying to take the pictures and write the content for the other sections... man, what did I get myself into?!

The cost of chassis: $104 USD, Parts (including chassis parts): $156 USD + Douk Board (~$56 USD) + Cost of Transformer + Cost of Tubes + ~$20 for attenuator + Assorted junk + Shipping = probably approaching $400 USD when all is said and done.

I've updated the main post to include rough costs.

Chris' answer I believe was right on target... I'm not sure what the Euro is at now though.

[spaceistheplace]

Jay and others: hold the phone until I run the parts list by a few knowledgeable folks, maybe we can pull the cost down a little while also optimizing for some critical areas.

I don't believe mouser does group buys but you might want to run the parts list by hificollective in the U.K. Or partsconnexion and see what they say?


John - USA

[ZZMoko]

I priced up the fancy bits I need for the RIAA section on Mouser UK came to a not bad £17 but they want £12 for delivery of somebits that would fit inside a tiny envelope 

So I would be interested in some sort of buying deal..the free delivery point is £33.