I'm really frustrated - Need help for tube mic pre in a bad way

[emrr]

Doug, I'm not sure if you are commenting on REDD47 sound or about NFB tube amps with stiff PSU in general.  Is this from experience?  Do you like them?

Later era tube preamps in general; virtually all have stiff PSU and 20+dB NFB.  I'm a tube preamp collector, have used probably 30 types to some degree, and am very familiar with over a dozen American types.  I don't own any that would be considered PA or consumer grade, like the later Altec green panel stuff with plug-in transformers.  All are employed in making records at my studio, on client dime.  While I appreciate the later types, which do have their own sort of sound, they tend to be so much closer to SS that I seldom pick them.  And they splat like SS; sometimes worse, I think.  Later types tend to end up on drums that need a softer cleaner sound, like a honky-tonk band doing Bakersfield type country tunes.  In reality, I'll use any of them for almost anything, but when a garage band wants to invoke Little Richard or Jerry Lee Lewis, the front end has got to smack the wall in a cool sounding way, which the later stuff just won't do.  What was this thread about, again? 

Relating to Ian's thread concerning DC stability in NFB preamp design, someone mentioned a comment I'd made about the WE 141A, which is a switched NFB design.  It's harmonic character changes radically with gain adjustment (30 dB range, I recall).  So NFB limits you if you want adjustable gain, simplicity, and remotely similar sound across the gain range.  Many program amps will have an open loop stage hitting a volume pot, then an output amp with loop NFB.  This gets you somewhere in between, but still gives a splat when stage one overloads stage two on unexpected peaks. 

 

[CJ]

V76 also has adj NFB.

 

[baadc0de]

But generally, it's loud, brash and forward.

Ah yes, the seldom heard adjectives for tube amps.  I don't know if I would include brash as a general term but in contrast to the creamy & pleasant sound they can often generate I would have to mention that they can have a very visceral, forward, bold, earthy kind of sound - even call it rambunctious with no NFB to level things out.  None of this is really a bad thing.   

Let me explain "brash" - when it's in its very linear range it's clean anyway.. but we have a lot of clean preamps around that are cheaper, don't need to warm up, that we have more channels of etc.. so we use the REDD when we want to have a huge forward sound and that means pushing it either a bit or a lot and that's where, yeah, it sounds brash to me - in a good way! It's just not something I expected from other reviewers and what tube preamps are known for in general. I'm not sad about this investment at all, it was my first project and it's still used all the time.

I will however, start to experiment with the NFB stage some more.. would like to have a few switches to massage that into less NFB and original. If you folks have any ideas how best to go about it, feel free to chime in or send me a PM so that we don't detour this thread too much..

 

[Kingston]

To the original poster, here is what I would suggest you build. You could even wire this straight into the tube socket, or you could use a turret board. There's just one tube and the most expensive bits are (always) the transformers.

The choice of tube here is very non-critical. It could be 6SN7, 12AU7 or 6N8S. Well those are all roughly the same tube, but 6N8S is especially good value. Stay away from expensive NOS tubes on ebay. You'll just waste money. If a tube for this project costs more than $10-20 you are being ripped off. To the people in the future reading this thread, it might not be true anymore.

This isn't really suited for many ribbon mics because there isn't so much gain, but it will cover most other situations, including "bus warmer" if you use a 20-30dB input pad.

As for PSU use this one from page 3, no modifications needed: http://musiciansgig.com/mila/MILA1-Updated.pdf

You can also use the input pad from there.

Of all projects, I think NewYorkDave's projects are best suited for beginners. They have lots of documentation here on the forum and the designs are so simple that all build errors will be relatively easy to troubleshoot. That's where I also started when I came to this forum some 5 years back.

 

[Kingston]

for someone looking for more designs that take advantage of the fact tube distortion sounds good, instead of trying to hide them here are great candidates.

If you don't want to swim in feedback, or at least want the option to turn it off when needed.

http://www.groupdiy.com/index.php?topic=38252.0

or this:

http://www.groupdiy.com/index.php?topic=42991.0

 

[Kingston]

I'm waiting for comments on G9 sound especially its noise performance.  But from experience please .. rather than " its got PCBs so can't possibly be as good as hand wired P2P by virgins."

What PCBs promise is repeatability.

Yeah I built a G9 using the popular PCB. I can certainly share my experience.

Because it was so much fun.

First thing that happens when you use it you will find oscillation. You end up wiring DI input's from under the board. Looks a little homely and rough already. Then you start measuring the thing. FFT shows nasty rectifier spikes starting from 100hz. You end up building B+ regulator and filtering externally on a veroboard. Now the problem is gone. But wait! I see 3-5dB bass cut on 50hz. What the heck is this people were saying this project is awesome?  Then you end up changing the output tubes, possibly rewiring heaters and/or changing the output transformers completely, externally.

Do you want to see how a PCB looks like after something like this? No you don't, it's like a hemorrhoid correction operation, when it hasn't healed.

G9 is long overdue for MKII revision! It's a great sounding preamp when it's fixed and running perfectly, but not a beginners project, especially if said beginner starts seeing all these errors and tries to fix them.

While PCB's do provide repeatability, the problem comes when they repeat absolutely everything, including design errors.


Don't get me started on the redd. How bad does a design have to be to need a local (plate-to-grid) and global feedback running at the same time? I know I know, the people in the sixties didn't actually want to hear distortion, but still, redd is pretty damn inefficient.

 

[leadbreath]

cool, thanks for those threads kingston, very interesting tube projects going on there..

 

[emrr]

Collins 6Q with interstage volume mod, or Gates SA-70. Or lassoharp's recent three stage circuit, or RCA BA-2, both of which require a DC tolerant output transformer. 

I also have original RCA BA-21A transformers in the BM if anyone wants to go there. 

 

[ricardo]

Thanks for this Kingston.  Is this the Director's Cut of NYD's masterpiece?  Where can we find the best P2P or tube socket layouts for this?  Any idea of the gain?  And what are the recommended transformers?

I suppose I'm asking which thread is definitive or at least guaranteed to work well.  The original 36 page thread is more than slightly confusing to newbies & gurus alike.

To the original poster, here is what I would suggest you build. You could even wire this straight into the tube socket, or you could use a turret board. There's just one tube and the most expensive bits are (always) the transformers.

The choice of tube here is very non-critical. It could be 6SN7, 12AU7 or 6N8S. Well those are all roughly the same tube, but 6N8S is especially good value. Stay away from expensive NOS tubes on ebay. You'll just waste money. If a tube for this project costs more than $10-20 you are being ripped off. To the people in the future reading this thread, it might not be true anymore.

This isn't really suited for many ribbon mics because there isn't so much gain, but it will cover most other situations, including "bus warmer" if you use a 20-30dB input pad.

As for PSU use this one from page 3, no modifications needed: http://musiciansgig.com/mila/MILA1-Updated.pdf

You can also use the input pad from there.

Of all projects, I think NewYorkDave's projects are best suited for beginners. They have lots of documentation here on the forum and the designs are so simple that all build errors will be relatively easy to troubleshoot. That's where I also started when I came to this forum some 5 years back.

 

[ricardo]

First thing that happens when you use it you will find oscillation. You end up wiring DI input's from under the board. Looks a little homely and rough already. Then you start measuring the thing. FFT shows nasty rectifier spikes starting from 100hz. You end up building B+ regulator and filtering externally on a veroboard. Now the problem is gone. But wait! I see 3-5dB bass cut on 50hz. What the heck is this people were saying this project is awesome?  Then you end up changing the output tubes, possibly rewiring heaters and/or changing the output transformers completely, externally.  ..... 

G9 is long overdue for MKII revision! It's a great sounding preamp when it's fixed and running perfectly, but not a beginners project, especially if said beginner starts seeing all these errors and tries to fix them.

Thanks for this Kingston.  Has gyraf replied to any of these criticisms?

While PCB's do provide repeatability, the problem comes when they repeat absolutely everything, including design errors.

Absolutely!  One hopes the designer also debugs the PCB layout.  On a low noise device, this may take MUCH longer than the circuit itself.

 

[Kingston]

I suppose I'm asking which thread is definitive or at least guaranteed to work well.  The original 36 page thread is more than slightly confusing to newbies & gurus alike.

There is no definitive one. People have built whatever they feel like. When you go p2p they kind of automatically become the ultimate modular gear. Add/subtract options as you go. I think I've seen about 10 different versions this far. Not two are alike.

They are all guaranteed to work just as well as the builders own quality control.

 

[emrr]

Has gyraf replied to any of these criticisms?

Yes, read that thread. Plenty of varying experiences, it appears.

 

[emrr]

Just heard the Everly Brothers "Bye Bye Love".  Distortion all over the vocal peaks, the kind healthy NFB will deny you. 

 

[ruffrecords]

It's simply a mu follower using a 6cg7. Add a 1:10 or higher input mic amp and you have around 40dB gain... need more, add another stage for about 60db gain total. But it's going to want to see 10K at whatever you plug it into...

Yes, that is basically right. The 6CG7 mu follower gives a good 25dB gain so two of them, with a pot between them plus a 10: input transformer will give you up to 70dB gain. The output will happily drive into a 10K load or if you really want to drive 600 ohm loads you can tack on a 10K:600 transformer and lose 12dB of gain and headroom. The mu follower does not use global feedback but does allow the 6CG7 (which is really a 6SN7 in a smaller bottle) to achieve close to its intrinsic distortion level which is pretty low (0.4% THD at +26dBV into 10K and almost entirely second harmonic). The simple two fixed gain stages means the first stage gradually gets over driven as the input level rises but it will usually achieve +30dBu before serious distortion sets in which would require an input level of about -15dBV. To avoid the distortion at high input levels you can simply switch in a 20dB pad.

Cheers

Ian

 

[ricardo]

Kingston, I apologize.  I've spent the last 3 days going through 51 pgs of the official G9 support thread.  It is definitely not a project for newbies.

With apologies to Jakob, IM very HO, I think these are the features which make it a difficult project.  I hope you will consider addressing these in a future revision.  This revision should be detailed in http://www.gyraf.dk/gy_pd/g9/g9pd.htm rather than in prodigy-pro as the info needs to in an obvious, easy to find, place.

• The PCB layout fixes the important sensitive paths with one exception; the DI input.  This is a major problem cos it adds more than 30cm of poorly shielded and uncontrolled signal path in a very sensitive region.  It would be better to have the input circuitry and transformer near the front panel to minimize this path.  Taking the mike input via balanced mike cable from the back XLRs is less of a problem.  It would also minimize the path lengths to the sensitive gain switch.
• There needs to be a very clear recommended earthing system.  Detailed diagrams and list of things to do.  This must include wiring of off-PCB capacitors, transformers & mains wiring.  It needs to say stuff like "twist these wires together and push them into the edge of the chassis."  I say more on this at http://www.groupdiy.com/index.php?topic=32559.msg586162#msg586162
• The PSU, which makes it one of the Lowest Noise tube preamps, is overly complex.  It is impossible to test by itself.  Taking the HV B+ via a reverse transformer is not a good idea.  In particular, putting 15V into a 12V winding will lead to saturation, large losses & heat.  There is too much voltage going into the 12V regulator which is why it gets so hot.  It might be better to use a 12VAC  winding and add another stage to the voltage ladder to the 48V supply.  The recommended transformer ratios are not clearly specified.  From what I can see, the secondary for the B+ supply needs to be 220x15/12 = 275 VAC.  You won't get this if you use 230V:15V feeding 12V:230V

I realise, this is a major revision which Jakob is unlikely to ever find time to do.  Especially as it really needs building up, testing,  revising and building (at least) one more to check the changes.  But some stuff, like the earthing recommendations and transformer specs can be done now.

Jakob, would you let some guru, who has already built one, to put these on www.gyraf.dk?

 

[emrr]

I think with an eye towards context, Jakob designed this at a time when there were virtually no DIY resources out there, few specialized parts suppliers at moderate cost, etc.  I do believe he states somewhere that they are essentially abandoned, but still available for reference.  To bitch and moan about it now ignores this context; be smart enough to build something else, or foresee the need to adapt and modify. 

I remember not too long ago when the rev D 1176 board came into being, and the output transformer was a huge stumbling block.  There was a one-off group buy that seemed like it might be the only shot.  How easily we are spoiled and forget. 

 

[ricardo]

I think with an eye towards context ....... loadsa pertinent comment.....  How easily we are spoiled and forget.

I'm fully aware of the HUGEamount of work that even converting an existing commercial design to a DIY project entails and we should all be grateful to Jakob for doing this.  But IMHO, any design needs to address issues like earthing, clearly and accurately.

This info is somewhere in the 51 pgs of the "Official support thread" or its links but hardly in a form for anyone contemplating a build.  I'm not bitching.  My request to Jakob is really if he will let someone else maintain the definitive info on G9.  It may need to be on a different website but should only have prescriptive info.  The discussions can stay on prodigy-pro.

This is so the gems that emerge from the discussions can be highlighted in the right place rather than buried in 51 pgs of "Can I use a capacitor made by Ozzie virgins instead of WIMAs?"  It took me 3 days to go through it.

But I like to think, and I'm sure Jakob hopes too, that the design will evolve and solve it's weak points while retaining its strengths. 

For this, a definitive statement of the proven status quo is necessary.  It's really just good Drawing Office & Design practice.  Just as useful for DIY as it is for commercial designs.

Maybe I should post this on the "official support thread" but I'm reluctant to add to 51 pgs.

 

[emrr]

G9 is one of many that at this point could use editorial oversight, if only it were possible.  Sum down the threads and nuke it all, give a link to the whole bloody mess in the wayback machine. 

 

[gemini86]

The g9 is getting a face-lift...

the 51X version could also be used for a 1ru build and I've been working to address the current issues with the layout, including adding relays to shorten the input paths and keeping the different stages of audio away from each other...

It's hard to want to want to get to work on these open projects when people grumble about your free gift to the community.

But I stand by my original opinion, build a NYD 1B or Ian's Mu follower.