the pimp / mod the Poor Man 660 thread

[lolo-m]

REAL FAIRCHILD... REAL FAIRCHILD... REAL FAIRCHILD...
First, the real fairchild uses 6386. PoorMan's uses 6BC8... IT WON'T SOUND THE SAME .

You are building a varimu compressor with the Fairchild topology... The smart Fairchild topology is the only thing shared between the two... And this is really great !!! I'm finishing mine and I'm quite sure it is a GREAT COMP... But NOT A FAIRCHILD...

If you want a Fairchild you'll have to spend about 3000€ for a close clone...

Thanks again to Analag for this clever work... I like new designs... Thanks again...

 

[emrr]

What TYPE of capacitors "Should" be used for the Fairchild time constants???

Or should I ask..

What TYPE of capacitors did the REAL Fairchinld use?
or does it even mater?

I'd think it would... tants or electro's?

Films make the most sense as the absolute value is usually a tighter spec than an electrolytic would yield.  Beyond that, I see no reason to have an opinion beyond voltage requirement and footprint.

 

[Moby]

You are building a varimu compressor with the Fairchild topology... The smart Fairchild topology is the only thing shared between the two... And this is really great !!! I'm finishing mine and I'm quite sure it is a GREAT COMP... But NOT A FAIRCHILD...

Of course that sound is not the same, but we are talking about time constants. That's part of topology...

 

[lolo-m]

I do agree, but thinking about the time constants, what about the input transformer inductance ???. The PM670 has got on the tube side a 10K impedance transformer paralled with a resistor (the pot), the Fairchild has got a 50k transformer only... Different transformer means different inductance means different time constants... IMO the input transformer is a part of the Fairchild's time constants (RLC)... Bad news isn't it ?

Don't think I don't understand you, I would like to have those time constants too (plus a few others...). But I think this is impossible. I'm maybe wrong (I hope to). I think that trying different caps is a cool idea... There's certainly a lot to find on this way.
Anyway, I read everything you post, and there's always something to learn...

 

[Moby]

The PM670 has got on the tube side a 10K impedance transformer paralled with a resistor (the pot), the Fairchild has got a 50k transformer only... Different transformer means different inductance means different time constants... IMO the input transformer is a part of the Fairchild's time constants (RLC)... Bad news isn't it ? Sad

Yes, that's true  But I still believe that it's possible to recalculate the constants... Hmmm it's just R/C constant but I was always poor in math... Let's wake up our math experts  If I'm right "thing" will be faster...5 times... am I'm wrong

Anyway, I read everything you post, and there's always something to learn... Wink

Thanks, I'm just trying to think loud and to stay sharp as possible. Hope I don't hurt someone feelings  ;D

 

[lolo-m]

Hope I don't hurt someone feelings  ;D

Not mine, anyway !

 

[Moby]

Thinking again... we are talking about DC time constant. Right? DC resistance of the EDCOR's sec is around 300 ohm not sure about Fairchilds. Termination doesn't change the DC resistance  10k or 50k doesn't matter. ???

 

[khstudio]

You are building a varimu compressor with the Fairchild topology... The smart Fairchild topology is the only thing shared between the two... And this is really great !!! I'm finishing mine and I'm quite sure it is a GREAT COMP... But NOT A FAIRCHILD...

Of course that sound is not the same, but we are talking about time constants. That's part of topology...

EXACTLY.

 

[lolo-m]

What I'm trying to point is the complexity of the Fairchild's time constants.
But copy the time constants schematic of the fairchild. I think you'll get close to them. I don't know if the PM670 can handle it...
Wolker ? Analag ? anyone else ?

 

[Kingston]

What I'm trying to point is the complexity of the Fairchild's time constants.
But copy the time constants schematic of the fairchild. I think you'll get close to them. I don't know if the PM670 can handle it...
Wolker ? Analag ? anyone else ?

What do you mean complex? The schematic only shows parallel caps and their charging time resistors for generating the time constant curves.

Yes the PM670 can handle them, and yes you can use the exact replica of the fairchild 670 time constant knob in your poorman, and yes it will probably even sound somewhat similar.

I personally don't see the point when you have the option to test what values work for your specific compressor usage. Those fairchild time constants were not really designed for the same tasks we see in the average mastering or mixdown session today. Use those fairchild constants as starting points if you insist. I've a feeling those longer time values from fairchild are quite useless in this day and age.

The attack/release method that guy bluebird suggested is a very handy mod. Heck, you could go crazy and add *dual* attack and release knobs if you do it the fairchild way. Two attack knobs per channel! Not the worst idea for the control freaks out there.

 

[Kingston]

About that mid-side transformer winding trick thing.

Who was it that said it was even possible with Edcors with some digging of the windings? How would I go about doing this?

What other options are there? I mean, are there some quality transformers with correct taps off the shelf? cinemag/sowter etc?

 

[bluebird]

I used 150v to 250v film caps...I'm sure 63v would work.

Moby I would give you some voltage readings and current readings but I can't get my self to pull the thing out of my rack now.

I think a good rule of thumb is to just try and have the smallest resistance you can on the meter resistor. On the scope the sign wave looked great with no resistor (a short) at all.

On a side note: I think before anyone starts having strong opinions on what to do or not to do to the original design they should actually build the unit stock and start from there.

It wasn't hard to lift the board out of my unit and take the 10uF cap out. Same deal with the meter resistor.

don't be afraid to get your hands dirty...

 

[Moby]

Those fairchild time constants were not really designed for the same tasks we see in the average mastering or mixdown session today. Use those fairchild constants as starting points if you insist. I've a feeling those longer time values from fairchild are quite useless in this day and age.

I agree  ;D I never used 5sec release time. But I have an impression that PM has too slow attack time. I'm just thinking because I miss few parts to finish the unit and can't do any measuring or experiment....  :-\

Moby I would give you some voltage readings and current readings but I can't get my self to pull the thing out of my rack now.

No prob, if you find some time it will be useful 

I think a good rule of thumb is to just try and have the smallest resistance you can on the meter resistor. On the scope the sign wave looked great with no resistor (a short) at all.

Yes, that's the point  8) Otherwise voltage will swing together with gain reduction and it's possible that it will drop too much for proper tubes supply 

Heck, you could go crazy and add *dual* attack and release knobs if you do it the fairchild way. Two attack knobs per channel! Not the worst idea for the control freaks out there.

Oh, what an overkill ! But who knows maybe it's fancy to have a lot of knobs  8)

 

[khstudio]

Good info everybody, thanks for all the diverse input

Those fairchild time constants were not really designed for the same tasks we see in the average mastering or mixdown session today. Use those fairchild constants as starting points if you insist. I've a feeling those longer time values from fairchild are quite useless in this day and age.

I would have to agree... they're quit long release times.

For all the time & money involved in this project, I just felt the sidechain (time constant section) was a little limited & possibly under designed.
I know it's not fair for me to say this without finishing my own build & hearing for myself FIRST...but I do understand HOW it works & there doesn't "appear" to be enough control in the original PM670 design... this seems to be confirmed (to some degree) by Bluebird.

Bluebird,
Thanks for sharing!!!

 

[lolo-m]

What do you mean complex? The schematic only shows parallel caps and their charging time resistors for generating the time constant curves.

I mean there's a RC network followed by the inductance of the input transformer. Don't you think this inductance will change something in the time constants ? I'm maybe completely wrong ... ;D

The attack/release method that guy bluebird suggested is a very handy mod. Heck, you could go crazy and add *dual* attack and release knobs if you do it the fairchild way. Two attack knobs per channel! Not the worst idea for the control freaks out there.

I think that this is a really more interresting mod... But this is just my opinion...

 

[Moby]

I mean there's a RC network followed by the inductance of the input transformer. Don't you think this inductance will change something in the time constants ? I'm maybe completely wrong ... Grin

I think that transformer resistance and inductance doesn't react with time constant. C4 is pretty close to the ground through RV6 and that's the resistance he is "looking at". So, that's something like 660 ohm.   :

 

[Kingston]

What do you mean complex? The schematic only shows parallel caps and their charging time resistors for generating the time constant curves.

I mean there's a RC network followed by the inductance of the input transformer. Don't you think this inductance will change something in the time constants ? I'm maybe completely wrong ... ;D

The effect of inductance on the time constants would only come to play here if the design would use some advanced transformer tricks that would skew the frequency response (say, pre- and de-emphasis tricks) with each i/o stage (much like an EQ in the sidechain would).

But that's not a fairdhild/poorman thing, and the frequency response (in conjunction with the inductance and loading) should be flat everywhere in the unit.

 

[emrr]

I said something about M/S hacking of the Edcors in several places, none of which I can point out in a quick manner.  I have not tried it, but it should work fine.  I've had to repair a number of Edcors straight from the factory; windings connected to wrong tabs. 

 

[Kingston]

I said something about M/S hacking of the Edcors in several places, none of which I can point out in a quick manner.  I have not tried it, but it should work fine.   I've had to repair a number of Edcors straight from the factory; windings connected to wrong tabs. 

they're cheap so I'll just hack away. What exactly am I looking for in there, and any suggestions for a tool for the task?

Damn this cheap mobile phone camera. I can't get accurate enough pictures to post in here.

 

[Moby]

I've had to repair a number of Edcors straight from the factory; windings connected to wrong tabs. 

How do you determine wrong windings connection? By measuring DC resistance or inductance.. or something else. It will be great if you can give info about proper connection and how to measure that