Poor Man's v76

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

MeToo2

Well-known member
Joined
Dec 30, 2009
Messages
339
Location
The Netherlands
Having built some valve based equipment I'm getting bitten by the bug.

One potential project I'm thinking about is a poor man's v76.

Why?

You can buy an original, but it costs a fortune and they are difficult to get hold of. Even if you get one you have to recap it and potentially find new transformers etc. as well as a rack mount.

Now obviously you can also buy a TAB version off the shelf for a goodly sum of money (and I have no doubt that it is an excellent reproduction of the original), but I have a hunch that there's interest at a much lower price level.

So why make a poor man's version?

Because V76 have been called the rolls royce of preamps
Because they are pentode based (whereas most diy designs seem to be triode based)
Because they are very quiet
Because of a rising interest in ribbon mics, especially for drums (which require high & quiet gain)

Because it looks tough to do (I've seen other people reporting difficulty getting their build not to oscillate)

Here's a straw man for some parameters that I came up with.....

Main use case: virtually no features except for adjustable high gain / quiet gain usable for ribbon mics for recording drums in small project studio
-> switchable minimum 3-70dB gain or more (aim for 76 and same steps as v76). 106dB S/N ratio or much better (aim for 120)
-> no phantom power in first pass

Cost price: way less than 600 euros for 3 or 4 channels including case PSU etc.
-> Valve choice: no ef804s or ef804 (see cost price) -> ef86 clone or 6j32p level

How accurate: first order reproduction. Using modern components. NOT worrying about metallurgy of transformers etc.
-> Edcor, Sowter or other off the shelf transformers

Stable: No oscillation or motor boating!

General Character: like the V76S studio version or V76/80/120 broadcast version?

What do you think would be good starting points for such a project?
Has this been done before?
Is it a lost cause?

Any other thoughts?

Thanks for your input.
 
MeToo2 said:
...How accurate: first order reproduction. Using modern components. NOT worrying about metallurgy of transformers etc.
-> Edcor, Sowter or other off the shelf transformers

Off the shelf transformers is like doing a Ferrari replica with a Ford engine. It might look cool and it will drive - just not like a Ferrari. In that case it might be better to go for a V72 replica, because it might be easier to get hold hold of high quality transformers with the apropriate ratios. On the other hand, Sowther (?) is doing V76 input replicas? If you get hold of the transformers and chokes then it should 'just' be a matter of doing a good layout. A V76 could actually be a nice project to do, it might benefit from a good dc regulation (and dc heating) and a nice dual layer layout and consequent star grounding. It might also benefit from a straighter layout, which is not ruled by the dimensions of the original housing nad on the way one could replace the 'lytics in the signal path with foil caps - which shouldn't hurt either. I'd probably replace the gain switches with relays and verify if it is possible to drop the hf cut (the internal fixed one, not the switchable) - there are horrible stories on the net that doing so would ruin the sound entirely, but to be frank I doubt that. I'd probably think of a one pcb / 2 unit layout like the DAOC, with the psu regulations on board in order to make things easier to finish and to keep the star grounding easier.

Actually there is a V76m version with a different input transformer, it 'only' does 60 dB gain but has an even straighter frequency response. The 'm' stands for 'measurement', it was intended as a measurement amplifier. I believe the original versions have a 1:30 input and the 'm' has 1:20 (that's from memory, needs verification). This one doesn't have the internal hicut either, but for some people it's not 'the real one' either. To others it's even better in places... I never understood what the 's' version was, any insight on that?

EF 86 should be a workable replacement for the EF804S, if dc heating is used and one is aware of the probably much shorter lifespan of the 86s (the EF804S is a longlife tube afaik, along with other audio optimisations)?

To sum it up, I believe it's perfectly doable and a very nice project - if the necessary irons are available and probably quite some effort is put into layouting. Otherwise it might not be worth the effort and other proven concepts might be more rewarding.

Michael
 
Thanks very much for the input. Appreciate it.  8)

Re 76M and other versions.

There's a well-known resource at http://audio.kubarth.com/rundfunk/index.cgi that has many of the various versions and their associated schematics.
It obviously helps if you can read German.....

AFAIK the M version was meant as a "Mess und Anpassungsverstaerker" = measurement amplifier : not necessarily a mic pre.
The S version was for Studio use. The /80 & /120 were for broadcast radio use and hence had narrower bandwidth....

The main differences seem to be the inter-stage filters and of course the construction of the transformers.

I'm sure someone else will chime in with 10000 other differences, but I really want to concentrate on a cheap first order replica to start with.

- relay switching of the gain sounds like a great idea to consider: especially for the input pad. Thanks. You can get relays for a euro or so a piece nowadays and it could simplify wiring.
Not only are 4 pole 12 way switches expensive and difficult to obtain, they're also potentially a major source of unwanted capacitive coupling (and hence stability) in the feedback circuit.
You could house those relays right next to the input transformer and just have the control voltage leave the screening can.

- re DC heating & tubes.
the EF804 is not just long life. It is low noise, low hum, low microphony, and also has low inter-element capacitance. Unfortunately they cost a small fortune now.

Yes I was thinking of DC heating. The particular ef86 version I had in mind is also a long life / low noise / special quality valve and has a DC heating arrangement which is pretty simple to implement with modern components. (I'm not going to reveal the exact type number until I've got myself a decent supply. They're no secret and are already very well known in DIY land but I don't want to drive the price up too quickly!)

I was also thinking on multiple silicon regulated HT supplies rather than going for replicating the original power supply: with a separate one for 1st stage and another for the 2nd. Silicon regs are cheap and the voltages are well defined on the schematics.

- re transformers. Yep this is probably the single biggest thing of course that determines the character.

yes I was thinking of 30:1 replicas of old ribbon mic input transformers. Edcor do one for just $22 and Sowter for gbp 75 or so. The Edcor one looks like a good starting point. Sowter might just fit in a budget.
IMHO You really need a lot of gain here at the front end to have any hope of matching the noise figures and to load the ribbon mic properly.

I also believe the input transformer has the biggest effect on the overall sound, especially as it is completely outside any (linearising) feedback loop.

I have read that the output transformers aren't "that special." Edcor do a very nice cheap matching transformer that can drive 600 lines for only $12 or so. "Heresy" I hear you cry.

As for" Ferrari with a Ford engine": it really is my intention to keep the price down. Yes, I know that is contentious for a lot of people, but I see little point in me trying to replicate something it took Oliver Archut of TAB Funkenwerk years to achieve (and he had access to the original drawings and some of the machinery!). His stuff is still for sale brand new on the open market if that's what people really want. My vision is very clearly a Poor Man's version. Not a full replica. Whether that is anything like the original remains to be seen until it's built.

Decent and cheap plate chokes may well be a real problem to source. I've struggled to find any of these on the market apart from Sowter (and they are 60 gbp each) which would make a significant dent in any budget. Since the plate chokes are within the feedback loop and we are talking pentode amp here (curve depends more on grid2 voltage than the anode voltage) I fail to see how they have too much influence on the overall sound relatively speaking. Others may disagree violently, which is their good right. If anyone can point me to a resource I'd appreciate it. I heard that in the past TAB Funkwerken sold "transformer kits" but I can't see them on their page. Again being realistic they'd probably not come in at the right price level that I'm aiming at (which I can fully understand.)

- re layout

yes I think this is going to be key.

I was going to avoid the "cassette" approach entirely. For a start I can't afford that much mu metal as used in the original.  ::)

I was going to avoid PCB layout too and go for point to point wiring. I have a stack of punches / screened ceramic sockets / tag boards etc. already in stock.

I was thinking on using several (cheap) screening cans together with sufficient air gaps between them (possibly one can for the input transformer, one for the 2 input tubes and associated circuitry, one for the 2 output tubes, one for the power supply) I was also thinking of a very basic mechanically isolated sub-chassis to to avoid microphonic feedback.

- re avoid electrolytic caps in the signal path

yes I agree entirely. You can get some pretty high quality high voltage foil audio caps for a few euros each.

That's as far as I've got for now.
 
There isn't much mumetal in the original, just the input transformer has 3 layers. Do you have a unit to spy on? I'm aware of the differences between the 80/120 (very minor, just the filter frequency) and the 'm', but I never came across any further info about the 's'. I havent checked Kubi's site on this revcently, maybe he added some stuff (actually quite some of the Braunbuch stuff there came from me originally). If you know what the difference is and you could put it into some words it would be great!

I'd anyway do a pcb if I was doing this project because if I do something once and it turns out to be great I might like to repeat it - and then a pcb will most likely save a lot of time. And it might even save time for the first unit, having a logical check between the schematic and the pcb, especially if many relays get involved. But that's just my personal opinion of course.

Yeah, I also wouldn't drop the switchable filters, they're very handy when the box is used as a distortion unit  ;D

Another thought (actually a clarification of what I wrote before): if you want to go for a budget version, you might be better off with a V77 / V72 variant. Simpler, probably cheaper high quality iron, just 2 tubes, less parts in total, simpler gain switching but not as much gain (not sure for the V77). And in my opinion it sounds even better than a V76, the mids are clearer. It is a matter of personal preference though, the V76 wil shine in different situations. For me the V72 is more hifi (but with great character) while the V76 rocks more.

Michael
 
My feeling is that whatever unique sonic properties these amps afford is to be found mainly in the iron and circuit design - which for the v72 entails a tricky plate choke.

http://www.groupdiy.com/index.php?topic=36618.0

So what you're left with when you move that far away really isn't a V anything and I think whatever inspired the moniker "Rolls Royce . . " will be lost.

The tele-Vs more than a lot of other tube pre's seems to be an all or nothing thing as far as 'cloning' goes.
The Mercury 'clones' may be a good place to get ideas from. Lundahl certainly makes a nice variety of plate chokes.

The WE141-A uses a pentode on the 1st stage and CF output with no output transformer and can do ~70db into 600ohms

I don't mean to dampen your enthusiasm - If it's more about finding a 'cheaper' solution then it's certainly worth the research effort. If it's more about the practicality of owning a good sounding tube pre using a pentode at reasonable price I'd say look towards a different design.

Also, if you choose to go the EF-86 route I'd recommend  giving the 5879 a try.  From reading forum posts from experienced users of EF-86s It looks like they're prone to being noisy and microphonic just like any other tube which means sorting. I've stayed away from these because of that - 17.00/tube is an expensive price to sort a dozen of at. Make sure the vendors will take them back.

Good Luck

 
That's what I meant. The mojo is - first of all - in the iron. The plate choke of the V76 will probably not be less difficult than the V72 one, but I haven't compared them in detail.

If anyone had experience with the Sowther replicas it would be great to know how they compare to the originals. All the historic value aside, just for working units I'd rather diy a new one with the outlined aproach and full 'new and modern'-parts reliability than spending days on restoring old units. It really isn't that much fun digging through 50 years of dirt, trying not to hurt dedicated, but weak aged parts. IF it sounds the same or better like an old unit, otherwise there's no alternative to an old one for me. I'm more a sound than a diy freak...

Thinking of it, changing the psu will certainly severely change the overdrive characteristics. Even replacing the old selen rectifiers with solid state ones does so. But it might improve things in linear mode.

Michael
 
lassoharp said:
Also, if you choose to go the EF-86 route I'd recommend  giving the 5879 a try.  From reading forum posts from experienced users of EF-86s It looks like they're prone to being noisy and microphonic just like any other tube which means sorting. I've stayed away from these because of that - 17.00/tube is an expensive price to sort a dozen of at. Make sure the vendors will take them back.
Know what you mean about ef86's being overpriced on auction sites. I think I can get hold of sufficient new old stock ef86 tubes at 2 euros per tube for NOS tubes from a well-known manufacturer to be able to hand select a working set of 9-12 tubes (for e.g. 3-4 channels) for a very reasonable price. I'm not into spending $17 or more on a suspect tube that may not even be new.  :-[  Also E83F seem not to be problematic at the moment. I got 12 @5 euro per tube for NOS e83f tubes from a well-known manufacturer, and they had more available. You only need 1 E83F per channel.
 
MeToo2 said:
yes I was thinking of 30:1 replicas of old ribbon mic input transformers. Edcor do one for just $22 and Sowter for gbp 75 or so. The Edcor one looks like a good starting point. Sowter might just fit in a budget.
IMHO You really need a lot of gain here at the front end to have any hope of matching the noise figures and to load the ribbon mic properly.

Unless I overlooked a new tranny in Edcor's product line, I think you're confusing things here. If you're talking about the RMX ribbon transformer, that's an *output* transformer for a ribbon mic, not a mic input transformer for a mic pre. It is designed for a source impedance of a fraction of one ohm (typically about .2 ohms). It won't do for a source impedance of 200 ohms.

The bad news is that constructing a good high ratio mic input transformer is much more difficult (and more expensive) than a ribbon transformer of similar (or higher) ratio. If you have to go cheap, you might wanna look for a 1:30 Beyer mic input tranny. But the're tiny and will likely distort at higher levels.
 
type in V76 CJ in the google custom search to start. 

If you search here you might find stuff.
 
MeToo2 said:
Because V76 have been called the rolls royce of preamps

Hmmm; Stephen Sank has been calling the RCA OP-6 the HOLY GRAIL for long enough that it's beginning to stick too.  And Western Electric is a much bigger Rolls Royce / Holy Grail than either.  I find it a funny designation of impossible promise, and I think those proclaiming anything to be THE RR/etc are doing so without listening to all available evidence; it's simply a prevailing wind.  

Because they are pentode based (whereas most diy designs seem to be triode based)

Plenty of pentode based preamps of old if you dig around in schematic land.  In practice I find pentodes used without feedback to have too high an initial gain, requiring an input pad up front nearly at all times.  I have changed vintage front end pentode stages to triode on many occasions at users request to get initial gain into a manageable range for modern sources.   Or you see pentodes are used in circuits with loop NFB to increase the gain within the loop; more feedback.  I think a lot of DIY being triodes is of a practical nature; loud modern sources.  

Because they are very quiet

As are plenty of the aforementioned available vintage circuits

Cost price: way less than 600 euros for 3 or 4 channels including case PSU etc.

Please prove me wrong, but now I think you're dreaming.   Especially if chasing the next item:

General Character: like the V76S studio version or V76/80/120 broadcast version?

transformers transformers transformers, coils, and transformers.  spice with unobtainium cap, resistor and tube flavors.  

You may build something that works reasonably well like a V76, you may not.  It seems best to keep it a lose target and inspiration at most, or you'll be better off just saving your pennies for the real deal, if you are that certain of your love.  

Good luck for the project!

 
Hi,


    I also agree wholeheartedly that the v72 sounds better than the v76. I sold my v76 and bought v72's. FOr most rock and pop recording, I still needed a pad! I had some 600ohm constant impedance attenuators for this. You would definitely need more gain for ribbons! I also loved them over a mix. They are a bit light at the bottom as stock, but I am sure I have read how to mod them to run much lower - just a cap change iirc. Sadly, I had to sell 'em . . .


      ANdyP
 
Hmmm plenty to think about still. Even at a very basic level.

@strangeandbouncy
Which version of V72 did you own and like particularly?

V72a V72b or V72s (or even the v72t transistor version)

The various V72 schematics I have seen appear to use *completely* different tubes and iron ratios ...... they're not even close.
Which makes me wonder when I read descriptions of how a V72 "sounds" if everyone is even talking about the same thing.

BTW I just scored two original V series input transformers and a power transformer for 10 euros for all 3. May be junk. May be usable. May be interesting just for comparison purposes. We'll see.
 
I believe that they used EF40(?) when they were first built, and EF804, and EF0804s later on as the new tubes were developed. iirc, they are interchangable, (not sure about pinouts) with the only real difference being the noise. Please someone shoot me down if I am wrong!


 I had EF804s in all of mine.



  http://www.tab-funkenwerk.com/id9.html
 
I never had a V72a (which was built by TAB), but I did own a V72 (Siemens).
Schematically wise, the V72a is very different to the V72, indeed.
From what I heard from a dutch tech experienced with this tube stuff (he used to work for the dutch broadcasting company NOS/NOB), the V72A has a much clearer sound.

But there's another thing what I've read, and what I want to add a note to:
That same person told me he sometimes replaced the choke with a 47K ohm resistor, and organised a listeningtest with a bunch of people that are known as having "golden ears". Guess what ? They all agreed that the one with the resistor instead of the choke sounded better.

I had a defective choke in mine, so I replaced it with a resistor, it works very well. I never had a chance to do a shootout though.

But if you read a lot about the V72, it's primarily the inputtransformer that adds the 'mojo'.

Personally, I do like the V72 a lot, but I don't dare to call it a clean, it's colouring as hell !  I had to sell my V72, but I'd like to build a Drip V72 one day. But I've got a lot of other projects running that I want to finish first though. (One of them is a G9.)
 
I remember a comment from CJ that he swapped out the choke with a resistor & ended up with a very similar sound. He did not say what value he used...

I've never heard a V72 but are the transformers that important? Would Jensens (for example) make a big difference?

I know all transformers have their own sound, but how good are the V72 transformers? Are they that much better than our best off the shelf units today?

I await your comments!

Peter

 
peterc said:
I know all transformers have their own sound, but how good are the V72 transformers? Are they that much better than our best off the shelf units today?

"colouring as hell" says the person before you. Would indicate they are much worse than the high end we have available today.

You're not going to have a V72 with any of the Jensen offering, but likely something much better.
 
Kingston said:
You're not going to have a V72 with any of the Jensen offering, but likely something much better.

Which may just ruin everything, and make everyone hate it. 

Just how did grandma make her spaghetti sauce? 
 
Don't mean to start a nuclear war on transformers.  ::)

The answer of course is "it depends."

Do you believe in mojo?. Do you believe in electronic engineering?
Are you actually looking for a clean sound, or a coloured sound?

Sticking to the V76 circuits I have seen, both sub-sections have pretty large amounts of negative feedback in them.
On the first stage there's an overall feedback loop into the input which is how the overall gain is tamed back to be adjustable between 3-76 dB.
The second stage is limited by capacitor 2uF C 31 and resistor 80K R 75 and is fixed (fed back into the cathode.)

Pentode response (when wired in classic pentode mode) is governed largely by the screen grid voltage & grid voltage and not by the anode voltage much at all (rp is very large for any decent pentode and it behaves like a current source). Plus heavy use of feedback tends to linearise the response and also reduce output impedance, greatly limiting the effect of any anode choke. Output impedance would generally by Rp || rp.
Since internal rp is so large the effective load would be dominated by the anode choke Rp (aka Zp). Otherwise, if the feedback was not doing its job, the anode load (and hence stage gain) under high z external load would be directly proportional to frequency, which it clearly isn't. So IMVVHO, as long as the choke is pretty big it shouldn't be able to influence things too much.

In the V72 circuit I have seen, the 2nd ef804 is wired in classic pentode mode, with linearising feedback taken from the (anode) choke through C40 C39 80K R18 R20 etc.
Input transformer is 2*1:20.

In the V72a circuit I have seen, the E180F is wired as a triode cathode follower as far as I can see: again with linearising feedback taken from the (cathode!) choke through C5.
Input transformer is 1:2*4.

There couldn't be much more difference between a V72 and a V72a circuit schematic if you made it up. Never mind who wound the iron or what wire or core they used.

The real exception I can readily see and understand from an electronic engineering point of view is the input transformer.
It is outside of any feedback loop, and transformers behave in a very non-linear manner when saturated.
They will also present very different loads back to the source depending on their frequency response and ratios.

Hence my earlier comment that I believe the input transformer is the most important factor in determining the sound. I don't mind spending money here.
(which rather luckily seems to be backed up by the previous remark in the thread about replacing the anode choke with a simple resistor and no-one hearing the difference :) )

So getting back to the v76, I think I'm going to experiment with a very cheap anode choke versus a simple anode resistor and see what effect that has. The anode choke seems to be the component that is toughest to buy as a modern equivalent at a reasonable price. I would imagine that a cheap anode choke would highlight any potential differences to the sound and let me know if it is worth investigating further /spending more cash. So I'll wait and see.

But what do I know?
 

Latest posts

Back
Top