the pimp / mod the Poor Man 660 thread

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Here are the samples ( just a part of a live track of a band I'm working with, choosen for its nice dynamic ). They are 24bits 48KHtz wav files. Same channel used , same settings, hardcomp means hard compressing, and soft comp soft compressing:
http://pagesperso-orange.fr/NBNS-Studio/PM670/ORIGINAL_01.wav
http://pagesperso-orange.fr/NBNS-Studio/PM670/ANALAGS-HARDCOMP_01.wav
http://pagesperso-orange.fr/NBNS-Studio/PM670/LOLO-HARDCOMP_01.wav
http://pagesperso-orange.fr/NBNS-Studio/PM670/ANALAGS-SOFTCOMP_01.wav
http://pagesperso-orange.fr/NBNS-Studio/PM670/LOLO-SOFTCOMP_01.wav
 
I do think the same but the other mod give really interresting results... Because there is a +17/-17 rail, I wonder if I should use this to get more close to the Fairchild compression style. The only question is how much negative goes the CV in the peaks of compression. If it's less than -17V an other mod is possible, but it won't be a all tube comp... But who really cares ? The idea is to keep all the sidechain amp as is, and to buffer with an opamp the CV after the Time Constant network... It will be possible to use resistors with a more than 1M value...

Note to everybody: in Analag's design, the faster attack is with the 10K pot set at 10K... The lower it's value, the longer is the attack or to be precise the softer is the attack... The release time is fonction of the program which charged the network. More clearly, if the comp compress during a short time, the 10uF will be charged during a short time and the release will be quite fast. If the comp compress during a long time, the 10uF will be charged more and its discharge (release) will be longer... It is an adaptation with really interresting attack/release of the position 5 and 6 of the Fairchild.
 
I wonder if I should use this to get more close to the Fairchild compression style
Hmm, with 6BC8 i doubt it's possible. Of course if you are talking just about T.constants maybe it's possible... I think that 6BA6 is better  because of varimu curve, but that will be pseudo triode sound signature. I'm still waiting for my 6BA6's . Also I'm not sure about wiring in triode mode. Direct?
 
My idea is this :
To have a 0,2ms attack time the SC-amplifier needs a 100nF cap. If we want a release of 300ms we need a 3M resistor in parallel. This 3M resistor isn't good for the 6BC8 (or 6BA6) grids.
If the voltage out of this network is not going lower than -17V, we can had a buffer to adapt the impedance of the network to the grids needs with an op-amp  ;D

Moby said:
I think that 6BA6 is better  because of varimu curve
You can't go out of the varimu curve of the 6BC8 in the PM670. The PM670 is a varimu comp.
To change the 6BC8 will change the way it compress but won't give you really more compression... 6BA6 is one of the possibilities... But the main "problem" if we can call this a problem, is the Side Chain amp IMO.
 
But the main "problem" if we can call this a problem, is the Side Chain amp IMO.
Yes I understand what you are talking about. I just wanted to say that 6BC6 has real varimu knee  ;) Regarding time constants, yes, some buffering will probably improve thing. I'm not sure about maximum DC SC swing needed  :-\ Did you tried to measure voltage at max Gain reduction? I;m away from my unit  :(
 
Hey, moby I've been away for a while...

Yes I wired the 6ba6 in triode. Pins 6,5,2 all together for the plate. and remember I boosted the voltage up to about 210 volts.

Like I said earlier the 6ba6's really made this thing rock. they can go into deep compression without the wierd crossover distortion the 6bc8 was getting.

I do think the time constant thing is more important than the tube change though.

I've been using the 2uf and 68K in parallel setting the most. this is closest to the fastest fairchild setting. its great on bass drums vocals, just about anything. I'm sure its not the exact same because of the weaker sidechain but I'm loving it.
 
Yes I wired the 6ba6 in triode. Pins 6,5,2 all together for the plate. and remember I boosted the voltage up to about 210 volts.
Great, I thought that it's dirrect triode connection  ;D About voltage boost I can see in the first post. Thanks  :)
 
I've got another working mod to make the PM670 be able to handle the Fairchild's time constant networks... and so many more... Cheap and easy to build !!!
Interrested ?
 
No, absolutely not. Of course damn it! ;D
I´m still waiting for my tube sockets to arrive so I can play with attack and release times myself. I want to control them separately for better flexibility. I´m really thankful for any information and advice, so please share your experience!
Thanks, Emre
 
Here it is !!!
The main problem in the PM670 is the output power of the SC-amp. The idea is really simple :
- Not enough power = let's add a power amp !

OK it is not a all tube comp, but the mod isn't in the signal path, so I don't care  ::) .

My main problem was to know were I could get something easy to deal with. I thought about having the time constants network buffered by an op-amp but the high impedance of the resistors ( up to 3M ) was a real problem. So I decided to make a big balled follower of the rectified signal after the Sc-amp.
More simple than what I was trying to deal with at the begining !!! The simpler the better !!!

It is working as is but there's a few thing I'm still working on. The extra diode makes the control voltage a little bit lower than original. There's two solutions to solve this, to had another polarisation pot for this mod, if you want both stock and modded mode, or to use a low forward voltage diode which can handle high repetitive currents during short periodes...

;D ;D ;D Any idea is welcome  ;D ;D ;D!

boost-analags-sc-amp.gif
 
Nice Lolo... so, boosting the Sc will can use the Time Constant's network from the original PM670 design? isn't it? is that the idea?

Cheers,

Eddie :)
 
I can't help to wonder where the "Designer" of this project is & his input on the "Side-chain" subject or even an explanation of WHY he chose to build it this way.  ???

If nothing else... it could help those of us who wish to have more options with the sidechain understand a little better what direction to go.... or not to go. ;)

Rowland... you out there bro? :-X
 
nice hacking lolo-m!

I guess the original PSU has plenty of power for all this.

so now you grab the "extra" +17V to your opamp, but also lose it for 12V gadgets (meters, relays)? Do you branch (regulate) the +17V down to 12V somewhere there as well, or did you just strap some resistors to your meters/relays get 12V?

Also, where do you grab the "additional V+" and how much did you intend this to be?
 
Moby said:
Did you tested the "thing" or it's just idea?
Yes, it is tested ! And it is working !  ;D

SaMpLeGoD said:
Nice Lolo... so, boosting the Sc will can use the Time Constant's network from the original PM670 design? isn't it? is that the idea?
Yes, the idea is to have the possibility to plug the real Fairchild's model 670 time constants network... And many more !!!
Faster attack is possible with a 1uF but you'll have a faster release too...

Attack time is about 0,2ms with a 2uF, about 0,1ms with a 1uF...
The release time is simple to calculate in the 1 to 4 positions : R* x C* !!!
A lot of things to experiment !!!  ;D

Kingston said:
nice hacking lolo-m!

I guess the original PSU has plenty of power for all this.

so now you grab the "extra" +17V to your opamp, but also lose it for 12V gadgets (meters, relays)? Do you branch (regulate) the +17V down to 12V somewhere there as well, or did you just strap some resistors to your meters/relays get 12V?

Also, where do you grab the "additional V+" and how much did you intend this to be?
You've got it, I use the +17V/-17V of the PSU board. Because my relays ( and later leds... ) are not supplied with 17V, I had a regulator to feed them ( 78XX where XX is the voltage you need !): this is the additional V+.
 
Hey guys... sorry if I missed it but what is the Fastest Attack Time with the "Stock" circuit?

Actually, what are the ranges of the Attack & Release with the stock circuit?
 
Back
Top