• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

EAR 834P mod - help identifying components

Hi hi,
I am currently in the process of modding a Douk clone EAR 834P. Not reinventing the wheel here, mostly just upgrades of coupling caps and whatnot .. Thorsten Loeschs Mod guide is well known and I am mostly keeping to it. Most components were easy to identify on the board, but some are really strange. Maybe somebody with experience with the unit can help me.

I have labelled a photo I am using for reference to the markings of the original schematic. I'm having trouble identifying the anode resistors of V1. They are supposed to be 330k, but I don't see them (maybe the two close to C8?..). I don't know where the 12 little 36k Resistors come in .. guessing they are somehow drawn together to substitute for a bigger resistor, but I don't yet know what and why. There is a single stray 68k Resistor close to V1 I can't place... And the strangest thing are two .. well I think semiconductors / transistors close to V1 labelled Q1 and Q2 (see last picture). What would they be doing there?...
Thanks a million.
 

Attachments

  • ear1.jpg
    ear1.jpg
    696.1 KB · Views: 940
  • ear-2.jpg
    ear-2.jpg
    536.9 KB · Views: 1,118
  • EARphonocircuitcopy.jpg
    EARphonocircuitcopy.jpg
    72.6 KB · Views: 1,090
  • ear3.jpg
    ear3.jpg
    969.8 KB · Views: 1,004
Last edited:
The schematic you are referencing is not precisely the same. I’m assuming Douk made some alterations.

Also, complicating matters the 834p schematic often circulating the web from one of the manuals which you reposted here is also not the same as the production 834p. Maybe the first iteration circa late 90’s but there’s notable differences now such as the JFETs you referenced.

I would ask Douk for a schematic of this design in particular or else you risk causing more harm than good.
 
Thanks for your reply.
Yeah, I kinda noticed that it didnt quite fit the schematic around the first tube .. the rest is accurate though. And swapping parts for equal value never caused any harm, so...
Just trying to understand the thing entirely, since learning is one part of the reason I'm doing work on valve amps.
 
Get the schematic from Douk and post it here, then we can actually learn something.

Parts replacement isn’t learning its blind repetition without a real goal in mind.

You will miss an opportunity to really improve performance this way.

JFETs may be the critical component to replace in matched pairs anyhow. Can you get part numbers off them? (Q1 and Q2)

Help me help you.
 
I've written to Douk some days ago, but they are yet to reply. I don't quite count on getting a schematic.

And no, upgrading components to better ones of the same type doesn't teach you much. But first trying to understand everything you touch does. Which is why I'm trying to figure this out in the first place. Also, I want to make this information available. I think this amp is a hell of a bargain right now, and I want to give people the resources to upgrade it without them all having to do all the footwork and without even having to understand it - the same way many great people did for me many a time.

Regarding the Transistors, I will take a closer look when I open the amp up again. I'm needing it in working order these next days. I guess things will get much clearer once I've removed the board and can see the traces. Right now I'm suspecting the Transistors might be remnants of the not installed MC part, which in the Douk version is supposed to be a cheap step up amp encased to look like transformers... Will keep you posted. ;)
 
Looking at the pictures, it appears that Douk even copied the terrible board layout of the original...

I would hope/expect that those jfets are actually signal input, cascoded under the first 12ax7/e83cc. This would be a very good thing if it is the case. It would keep input miller capacitance low, which helps keep most MM cart freq. response a bit flatter after 15kHz.

If you are really interested, you could trace the circuit from the back of the board.

I have a feeling some of those 36k resistors are in parallel. If you don't have a 10k 2W resistor, but have four dirt cheap 36k 0.5W resistors... well, you can figure this out this I think. ;)

From my experience with this circuit, the Thorsten mods are ok, but you need to be careful not to mess up the RIAA correction too much.

From what I can see, the cathode by-pass caps on this board are not so hot. Try Elna Cerafine or Nichicon Muse. Try to get real Elna Cerafine if possible... I was once sold CN garbage fakes (the Hong Kong seller immediately refunded my money when I messaged him). These days I tend to source components from real suppliers, like mouser, etc.

When I re-designed and built this circuit for myself, I used amtrans AMCH for the RIAA correction (C3 and C2). Its probably not easy to get these in the correct values anymore though, so silver mica is a good choice. Even better is polystyrene.... but be really careful when soldering polystyrene - don't get them hot or their values can change. In any case, the C3 on your board looks like it is silver mica, and is probably ok.

I also replaced V3 with source followers using ZVN0545A. Let me see, I also did a completely different power supply since the original design was quite modest. These days I would build a statistical regulator as described by Morgan Jones, but that will require more space than this little box.

Ian
 
Last edited:
I had very much feared someone would say that. Although I searched extensively, I have not found a schematic with that topology. Which advantages might this design offer over the original?

cascoded jfet input is something that Allen Wright did too. It is a nice way to keep miller capacitance low.

Genera_and_Goldring_2500_all.gif


The purple guy in the above diagram is not what you want. I don't know how many times I hear an MM phono set-up that sounds like purple-guy. Shrill sounding... I can hear it right away. Very quick listening fatigue...

I sometimes think this is what gave MC a commercial boost over MM. However, after getting mr. miller in check, I sold all my *expensive* MC cartridges on that auction site and have never looked back. :)

The phono cable alone is probably 100pf. Measure it if you feel inclined to do so. 12ax7 or any other high-gain input valve will have by definition, very high miller capacitance. Some (very few) MM carts (shure mostly) actually need extra input capacitance, which can easily be added. ;)

These days I am cascoding E88CC on my phono input and am pretty happy :)
 
Last edited:
Thanks for you fruitful input, Ian!

I would hope/expect that those jfets are actually signal input, cascoded under the first 12ax7/e83cc. This would be a very good thing if it is the case. It would keep input miller capacitance low, which helps keep most MM cart freq. response a bit flatter after 15kHz.[...]cascoded jfet input is something that Allen Wright did too. It is a nice way to keep miller capacitance low.
Thats a good improvement then. My hate on what I thought was a repair on something not broken slowly subsides. :D

If you are really interested, you could trace the circuit from the back of the board.
I'm planning on doing that. They peaked my curiousity. Gonna draw up a schematic also.

I have a feeling some of those 36k resistors are in parallel. If you don't have a 10k 2W resistor, but have four dirt cheap 36k 0.5W resistors... well, you can figure this out this I think. ;)
I kinda figuered something like that.. Crazy what they do to save a few cents... Altering the board just for that.

From my experience with this circuit, the Thorsten mods are ok, but you need to be careful not to mess up the RIAA correction too much.

From what I can see, the cathode by-pass caps on this board are not so hot. Try Elna Cerafine or Nichicon Muse. Try to get real Elna Cerafine if possible... I was once sold CN garbage fakes (the Hong Kong seller immediately refunded my money when I messaged him). These days I tend to source components from real suppliers, like mouser, etc.

When I re-designed and built this circuit for myself, I used amtrans AMCH for the RIAA correction (C3 and C2). Its probably not easy to get these in the correct values anymore though, so silver mica is a good choice. Even better is polystyrene.... but be really careful when soldering polystyrene - don't get them hot or their values can change. In any case, the C3 on your board looks like it is silver mica, and is probably ok.
I'm not going to play around with the RIAA stage values. Some people did, but I decided against it. Just got some really good silver micas with a very tight tolerance and same for the two resistors. I figure the micas are propably really decent, but since I'm ripping the thing apart anyway and plan on keeping it quite some time, it's not big trouble to put in the best things I could find.

For the cathode bypass I ordered some Elna Silmic II and soldered some 1uF MCaps as a bypass to them. I'm doing the Coupling Caps with Mundorf Zn, as suggested by Thorsten. I have no references for this unit, so I just went with it. Also gonna replace a few resistors with some nice and ridiculously pricey alternatives. Replace, enlarge and bypass PSU elkos. Replace the diodes. Bypass heaters. Throw out the caps and resistors between V2 and 3 and put some 100R Carbon composites there. Overkill has it's own rules... :D

The phono cable alone is probably 100pf. Measure it if you feel inclined to do so. 12ax7 or any other high-gain input valve will have by definition, very high miller capacitance. Some (very few) MM carts (shure mostly) actually need extra input capacitance, which can easily be added. ;)
I think I remember I made it 100pF when I changed it. I am running a Nagaoka MP500 cartridge, which technically is a moving iron as far as I remember, but yet it has proved to be really unproblematic.
 
Last edited:
Moderator
Joined 2011
cascoded jfet input is something that Allen Wright did too. It is a nice way
to keep miller capacitance low.The purple guy in the above diagram is not what you want.
I don't know how many times I hear an MM phono set-up that sounds like purple-guy.

Many others have, like Audio Research. There's lower noise too. Are those actual
cartridge response curves? Usually there is a broad dip before the peak.
 
Last edited:
Nagaoka MP500 is a fine cartridge. So you definitely do want to keep an eye on the miller capacitance on the phono input. once checking, I would leave in the jfets.

I would not bother by-passing the cathode caps with film... but ok. your choice.

For coupling caps I have used the really CHEAP ugly yellow plastic ones that are on your board before. They were in fact quite decent. But I imagine in Germany it is almost a 'must' to use Mundorf caps somewhere... :D

I would not bother replacing resistors. I doubt you will hear any difference. Serous. Someone will dispute me here for sure. I still doubt you will hear a difference. Nichrome film caps are IMHO tough to beat. Even very, very cheap ones...

The original board was not so good. Hum was built in with the terrible layout. Maybe today's version is better than the 1990's version in this regard. I doubt it as it looks the same and uses that same hopeless metal divider.

The only complaint I will make is that with 12ax7, the RIAA is influenced by the 12ax7... Not much you can do about this though.

I found cheap china (thick glass, ugly red or white print) 12AX7A to be just as good "sounding" as telefunken flat plate 12ax7's (Dynaco label, etc). but you had to tube roll a bit. Older tubes will have a higher Rp, so the bass will start to be more prominent...

The power supply change made a big difference for me. Back then I used a well shunted supply. Today I would probably build a statistical regulator like Morgan Jones...
 
Be careful with the 100R Carbon comp mod. It removes a low pass / subsonic filter which in some situations could damage your speakers. Do some homework before blindly implementing that..... kind of like the suggestion to use the air cores for riaa caps... seems a bit overboard and asking for a short or some other problem.

I’m not sure what the actual cutoff point is or if it could be optimized though... anyone have any ideas? It should only cut off the inaudible range anyhow. Sure you lose a cap, but a good .1uf film shouldn’t be detrimental.

Yes the FETs are a good thing if implemented well.... I’m just skeptical Douk used closely matched devices or the same as ear. Do you have a way to test the FETs? The ideal RIAA values would be 787K, 100pf and 303pf.

Also special care should be given to the 2M feedback resistor. Placing a grounded can over V1 will also make a big difference. Also using channel matched low noise and microphonic selected tubes is critical. I’d recommend the 12ax7EH or SPAX7A for new production.

Your model doesn’t have MC correct? Can you fit a Salas shunt regulator on the ps side? If so I’d gut what’s there.
 
Last edited:
Nagaoka MP500 is a fine cartridge. So you definitely do want to keep an eye on the miller capacitance on the phono input. once checking, I would leave in the jfets..
Yes the FETs are a good thing if implemented well.... I’m just skeptical Douk used closely matched devices or the same as ear. [...] Do you have a way to test the FETs?
I am planning on leaving them, you have made a compelling argument. I have no way to test or match FETs. I had rigged something about seven years past to test FETs for a symasym build I did, but it was a breadboard and was quickly taken apart after... I hoped I did not have to do that again...

I would not bother by-passing the cathode caps with film... but ok. your choice.

For coupling caps I have used the really CHEAP ugly yellow plastic ones that are on your board before. They were in fact quite decent. But I imagine in Germany it is almost a 'must' to use Mundorf caps somewhere... :D

I would not bother replacing resistors. I doubt you will hear any difference. Serous. Someone will dispute me here for sure. I still doubt you will hear a difference. Nichrome film caps are IMHO tough to beat. Even very, very cheap ones...

I had the MCaps kicking around for more than half a decade without actually ever using them. Some People recommend bypassing cathode resistors .. I just don't know. No qualified opinion. I think they at least won't hurt.

For coupling caps I mostly used old F&T polystyrene caps in the past .. but the biggest value I could ever get my hands on was 47nF, and paralleling that many caps is a bit of a pain. They were rather cheap and sound really good, but they were old stock and I have none left. So I decided rather randomly on supporting the german economy. :D

Regarding the resistors, I tend to think you are right. Although I have swapped all resistors in power amps for carbon composites before and thought that I heared a transition to a smoother, warmer tone, I think it's mainly the ones directly in signal path making the difference.

The original board was not so good. Hum was built in with the terrible layout. Maybe today's version is better than the 1990's version in this regard. I doubt it as it looks the same and uses that same hopeless metal divider.
I've tested this unit for a couple of weeks now, and even when listening to music at unchristian volumes, the amp is extremely quiet. There is no relevant hum whatsoever. I am very impressed with that.

The only complaint I will make is that with 12ax7, the RIAA is influenced by the 12ax7... Not much you can do about this though.

I found cheap china (thick glass, ugly red or white print) 12AX7A to be just as good "sounding" as telefunken flat plate 12ax7's (Dynaco label, etc). but you had to tube roll a bit. Older tubes will have a higher Rp, so the bass will start to be more prominent...
Also using channel matched low noise and microphonic selected tubes is critical. I’d recommend the 12ax7EH or SPAX7A for new production.
I am not buying telefunkens anymore.. They have multiplied in price in the last years, and I think it's getting ridiciulous. I have some good pre war EIs I got for around 20 bucks a piece, which were made on the telefunken machines. They are really good, but also getting scarce and pricey these days. Right now I have some old JANs in V1 and V2 and a NOS Mullard CV4003 (12AU7/ECC82) in V3 and I think they made a relevant improvement over the stock JJs. When eventually my stock runs out and all NOS tubes cost fortunes, I will look for good new production. I hear Psvane makes really good stuff....

The power supply change made a big difference for me. Back then I used a well shunted supply. Today I would probably build a statistical regulator like Morgan Jones...
Can you fit a Salas shunt regulator on the ps side? If so I’d gut what’s there.
I think it's safe to say that the PSU is critical in pre amps and propably the biggest improvement you can make to an existing circuit. I have early on made the decision to not rebuild the PSU, because I just don't have the time anymore. Initially I only wanted to change the coupling caps, but the whole thing escalated once again...

Be careful with the 100R Carbon comp mod. It removes a low pass / subsonic filter which in some situations could damage your speakers. Do some homework before blindly implementing that..... kind of like the suggestion to use the air cores for riaa caps... seems a bit overboard and asking for a short or some other problem.

I’m not sure what the actual cutoff point is or if it could be optimized though... anyone have any ideas?

It should only cut off the inaudible range anyhow. Sure you lose a cap, but a good .1uf film shouldn’t be detrimental.
[...]
Also special care should be given to the 2M feedback resistor.
This is supprising to me. As far as I can tell many people did this mod and noone reported negatively about it.

The ideal RIAA values would be 787K, 100pf and 303pf.
Maybe next time I order from hificollective I'll get 300pf silver micas and a something to build a closely matched resistance.

Placing a grounded can over V1 will also make a big difference.
Where would you place what?

Your model doesn’t have MC correct?
Correct.
 
Last edited:
The valves/tubes are in a closed metal enclosure, so no need to do extra shielding.

There is honestly only so much to get out of this topology. It can be pushed pretty far, but at some point you might want to go for something different.

The riaa is tricky... easily changed by a few picofarad. The input 12ax7 can easily change it too - checking input jfets is probably not necessary, but you can run a jfet off a battery supply (which makes setting up a testing rig easy). Jfets from the same batch tend to exhibit similar characteristics. Mosfets however are a different story... From my experience there will be far greater potential variation between input 12ax7 triodes than the jfets.

What is the psu transformer like? The original ear834p was quite modest. Be happy you have no hum.

Honestly, if it sounds good, don‘t mess it up too much. If it is really not good enough, think about a different topology...
 
Last edited: