the Poor Man 660 support thread

[lewilson]

maybe a small cap from cathode to cathode on the first amp  will bring up the top a little on some of these units?

 

[lewilson]

I tried a .047 from r1 to r2 cathode side and got a nice little boost from about 10k up. Noise floor is reduced some im sure, but its a little sweeter.

 

[bernbrue]

My version:
3 U, Frontpanel made by Frank/NRGrecording (thank you very much, Frank).
Meters are old Sifams.
I built it the original way, no modfifications.

Very special thanks to Rowan/Analag for sharing.
I really like his design ideas, I really like my D-AOCs, I will build a second PM 670.

Thanks to Volker for the PCBs.

Holger,
what a beautiful compressor! Your taste concerning frontpanel layout makes me a little bit jealous . Your built will be difficult to top. Congrats!!
regards
Bernd

 

[Holger]

Thanks Bernd 

 

[lewilson]

Holger, Where did you buy the tor. power trans?

 

[Holger]

It's one of the Rondo/Müller transformers.
O got two from Bernd's / Volker's fist shipping, then I killed one, a little bit later I jumped on another group buy to get a second.

 

[Kingston]

I managed to get cross-over distortion to acceptable level. It was a case of tube mismatches, as suspected earlier. And thanks to all the hints people. I had not matched the tubes at various grid voltages, and that did the trick.

I have a box of 20 of 6BC8/6BZ8 and was just barely able to get an acceptable set. Lesson learned, next time I'll know to order at least 40 or more.

I decided I needed to feed at least two different grid voltages to the amp stage - for two sets of measurements -  to find out how much each triode side swings, and if the triode sides are mismatched or not. So I picked silence, and 50hz sinewave at 0dBU as the grid voltages. When I get more tubes, I will add a third step. Maybe some really loud +12dBu pink noise to get everything matching perfectly at the very highest levels as well.

I could really see I had a wildly varying badly matched set of tubes. Probably not so desirable rejects pushed down the line by the seller to an unsuspecting customer (me). Ah well, armed with my multimeter I overcame his ill will.

Now I have a pretty decent sounding poorman base model on the bench. Still not happy about the transformers. The edcors just aren't awfully good in this application. There's this weird phasing in the high frequency area, that makes all "S" area and hihat sounds like they have spit and sand on them. Not even really good as an effect, so the "base poorman" is basically an ok - but not brilliant - tracking compressor.

Since I didn't want another tracking compressor, I'm going to start looking for a good set of alternative make up gain amp transformers.

Also, I don't really agree with the "base" time constant set up either. There's barely any range for a compressor this powerful.



Luckily there is a full rainbow of mods to do, and I suspect this project will eventually turn into into the turbo compressor I always wanted.

 

[lewilson]

So i suspect it would be out of the question to get another power trans? I might build another pm if the boards become available.
Kingston, I 've been listening to mine today quite a bit. My high end is good. Have you tried the Russian tube that replaces the 6bc8?  its the 6n5p. its a much better tube I think.  Buy 20 or so from ebay. 
  I guess if you had another transformer that was close , you could sub it in on the first position on one channel and see if there is a clear difference.

 

[lolo-m]

I have a box of 20 of 6BC8/6BZ8 and was just barely able to get an acceptable set. Lesson learned, next time I'll know to order at least 40 or more.

Kingston, maybe it's a worthless question but did you try to match your 6BC8 by pairs ? I mean current pumped by the 6BC8 plates must be the same on the 2 sides of the push-pull stage (and it is the addition of the plate current of 2 triodes of 2 different tubes). So the match can be achieved by 2 tubes with both mismatched triodes... I hope I'm clear but my not so good english doesn't make me sure I am  ;D...

 

[Kingston]

Kingston, I 've been listening to mine today quite a bit. My high end is good. Have you tried the Russian tube that replaces the 6bc8?  its the 6n5p. its a much better tube I think.  Buy 20 or so from ebay. 
   I guess if you had another transformer that was close , you could sub it in on the first position on one channel and see if there is a clear difference.

No I have not tried the 6n5p, didn't know of its existance, thanks for the tip.

about the transformer again,

there's nothing broken about them, and I certainly did not mean to convey that. They are all up to spec, and work fine. The sound quality problem I described earlier is subtle, but still something I would not want on a mastering compressor. The test is really easy, there's no to even use poorman for it. Just strap the transformer directly on a balanced line, and it can be heard, easier on "trebly" material. Especially when compared to an approximately equal Cinemag or Lundahl. Most here probably don't care about differences that subtle.

Kingston, maybe it's a worthless question but did you try to match your 6BC8 by pairs ? I mean current pumped by the 6BC8 plates must be the same on the 2 sides of the push-pull stage (and it is the addition of the plate current of 2 triodes of 2 different tubes). So the match can be achieved by 2 tubes with both mismatched triodes.

I saw that from the schematic, and that's indeed how I eventually found a good match. My box of 20 was still all over the place.

 

[1954U1]

about the transformer again,

there's nothing broken about them, and I certainly did not mean to convey that. They are all up to spec, and work fine. The sound quality problem I described earlier is subtle, but still something I would not want on a mastering compressor. The test is really easy, there's no to even use poorman for it. Just strap the transformer directly on a balanced line, and it can be heard, easier on "trebly" material. Especially when compared to an approximately equal Cinemag or Lundahl. Most here probably don't care about differences that subtle.

I've made the same "crude" test with various transformers [edcor marinair carnhill lundahl UTC cinemag sowter haufe altran peerless], found that:
- the edcors XSM step-up treble problem isnt that subtle IMO.. cleary audible yes on cymbals hi-hats sibilants
- the edcors  have by a great distance the worst [measured] phase shifting behaviour.. [and non-linearities past 20khz]

I dont know if the issues are related.. perhaps with a proper loading and tubes matching this treble problem will disappear?

 

[lewilson]

Sounds like you found the trouble then.
   I think its time people start posting pm670 sound samples.  Maybe we need another thread
I have some extra edcors here. Maybe I will do some square wave testing  today. Maybe somebody has done it already?

 

[emrr]

There are Edcor freq/phase/loading comments in some dedicated threads.  Somewhere here......

 

[lewilson]

I wonder who makes OPAMP LABS transformers? They have a 600/10k model T-610 for 50.00 usd . I happen to have one here I took out of something.  Maybe that could be a good solution for the input tranformer. there made to go into a octal tube socket. Anyone use these?

 

[emrr]

The one I measured was very poor on the bottom end, unless driven with a very very low-Z source.  Check it first to confirm or deny. 

 

[lewilson]

O.k Ill do some transformer testing this week.  
  Ive been putting some drum  machine through mine as well as some pre-recorded material. Im not hearing any smearing. I just squashed some Sinatra "thats Life" I pegged the comp meters till they were standing at 20db and it sound great! The only siblence I hear is that old u47 on the original recording.I can adjust the attack time and hear the difference in the high end change, as well as the string bass attack. It changes the whole mix without mudiness. Its cool.
If you try it you can hear a bunch of 120cps at the end fad out of that song.

 

[1954U1]

Ive been putting some drum  machine through mine as well as some pre-recorded material. Im not hearing any smearing. I just squashed some Sinatra "thats Life" I pegged the comp meters till they were standing at 20db and it sound great! The only siblence I hear is that old u47 on the original recording.I can adjust the attack time and hear the difference in the high end change, as well as the string bass attack. It changes the whole mix without mudiness. Its cool.
If you try it you can hear a bunch of 120cps at the end fad out of that song.

The "bad" highs I'm hearing on standalone XSM 600/10k edcors arent bad because smeared or muddy..
Its somewhat the opposite..
Like having little particles.. how can I say.. like touching with your fingers a wrongly washed silk.
Maybe it has to do also with wrong past 20khz content, and subsequent IMD in next stages.. I dont know.

 

[Kingston]

More fighting with poorman, standard setup. I have now highly matched tubes, and have gotten both channels behaving equally.

But there's still massive distortion (same on both channels), and I'm definitely not seeing even close the 18dB amplification that should be present with the make up gain 6BC8 amp, I have about 10dB, and even then extremely distorted. I can't get clean audio from input to output even when I set the unit to unity, and with no compression. At max amplification, and modern levels at output, it sounds just nasty. It's like there's no headroom at all!

The distortion I have is very odd looking. Not cross-over distortion, but not the type of crunch associated with lack of headroom either (flattened waveform)

This is the setup:

1. feed clean 6.9dBu 100hz sinewave to poorman input.
2. Set poorman make up gain to full blast, no attenuation
3. Set compression to full attenuation (no compression)

This way I get about 12-13dBu sine at poorman output, but badly distorted.

The reason I use 6.9dBu at input is because I would have to otherwise attenuate the output going to my soundcard so it won't clip.

But really what all this means, is that I have absolutely no way to listen to clean audio even at unity gain through poorman, with no compression! when I start to compress, I have to use more make up gain, which means more distortion from the make up amp.

Notice the odd looking (but symmetrical) distortion I'm getting. It's like the sine wave is "thinned out".

Any ideas what's going wrong? As far as I know there should be FAR more headroom than that?

Could someone else simply feed about 7dBu 100hz sine to their poorman, and tell me if there's any distortion?

Also, what is the best way to disable sidechain amp completely, and cleanly take it out of the signal path? I'd like to isolate this. I'm quite positive the problem is with the 6BC8 amp, but just to be sure...

 

[lolo-m]

Kingston, what do you use in the meter path ? Did you bypass R6 (22k) with a smaller resistor (1k or less) to feed the vu ?

 

[Kingston]

Ah yes, the meter is paralleled to the 22k resistor with a trimmer. I have a 4,7k trimmer, set somewhere below 1k I think. That's how you zero the VU's in the first place, and they seem to respond fine (if a little slow, but that's just the meter ballistics). This one: http://www.canford.co.uk/Products/86260/58-379-SIFAM-AUDIO-LEVEL-METER-AL20SQ-Retro

Taking the meter completely out of the circuit has no effect on distortion, just on the level very slightly.