[BUILd] PM660 / PM670 SCAmp Booster build and tips help thread

[lolo-m]

Edited first post with diferent links and soundsamples...

Eddie, is that what you're looking for ?

 

[lolo-m]

Changed the soundsamples on first post... 

 

[lolo-m]

PIMPING MY OWN PIMP !!!
Waiting the boards to arrive at home, I played with my prototype... And pimped it  ;D ;D ;D.
Here are the results :
  I tried at the begining of this mod to reach the fastest attack times of the Fairchild ( the more difficult thing to reach)... but I forgot one thing : the slower attack times...
  So, playing with the prototype and custom TC network, I realised I forgot one resistor  :-[. But you can easily put it on the board ;D ;D ;D :
On the footprint of the 1N400X diode, solder a 75 ohm serial to the diode on "ring" side of the diode.
With this resistor you'll get :
- approx 0.2ms of attack time with a 2,2uF (no paralleled cap)
- approx 0.4ms of attack time with a 4,4uF (2,2uF paralleled cap)
- approx 0.8ms of attack time with a 6,9uF (4,7uF paralleled cap)
- approx 1ms (or maybe more because it gets difficult to measure this) of attack time with a 15,8uF (2x 6,8uF paralleled caps)

To know what resistor you must parallel for a defined release, use the basic RxC calculation (result is in second). Don't forget you already have a 221K already parralleled...

If you plan to use the Fairchild's TC network, use the 75 ohm resistor serial to the diode. It will work  .

I really suggest to use this 75 ohm resistor... Without it it's working, but you will always have approx 0,2ms of attack time with various compression attack figures depending of the caps you'll use in the TC network...

I sould reffer to my note to self a little bit more  :

 

[Kingston]

Regarding bigger and more flexible networks,

how do you suppose it should be handled, when there will be significant amounts of LARGE film caps for each channel?

with the possibilities you listed there combined, it's going to be one FAT time constant expansion board in an already cramped space.

 

[radiance]

Lolo, just checking weither I understand your side chain booster.

Higher voltage in the side chain means that the caps that handle the attack time can recharge faster (or that you can use different value caps that do).
Soo, can this concept be applied to the GSSL side chain as well so we can have ultra fast attack times?

Or can the VCA's not generate them?

Thanks!

 

[lolo-m]

Regarding bigger and more flexible networks,

how do you suppose it should be handled, when there will be significant amounts of LARGE film caps for each channel?

with the possibilities you listed there combined, it's going to be one FAT time constant expansion board in an already cramped space.

Yes, my TC switch is quite big !!! I used a 12 position, 4 rows switch per channel. I have 6 different release times on each of the 4 different attack times ... If I cannot have what I want with this, I guess I'd better use another comp  ;D !
The switch measure about 8 cm !!! But my case is 40cm deep  !!!

 

[lolo-m]

Lolo, just checking weither I understand your side chain booster.

Higher voltage in the side chain means that the caps that handle the attack time can recharge faster (or that you can use different value caps that do).
Soo, can this concept be applied to the GSSL side chain as well so we can have ultra fast attack times?

Or can the VCA's not generate them?

Thanks!

Yes, I had a look at the TC part of the GSSL... The TL074 isn't a really powerfull op-amp...

I don't know if the VCA can handle ultra fast attack times so I can't say it'll work or not  ???.

Changing this opamp for a more powerfull one, or maybe to had a buffer before the TC network could give faster attack times.

But the network isn't the same at all:

- The Fairchild TC network on position 1 2 3 and 4 is a cap charged by the current of the scamp (attack) and discharged through a resistor (release). To change the attack times, you've got 2 solutions: changing the cap or having a more powerfull scamp. As the release resistor is the grid resistor, it can't be increased at more than 220K to keep the tubes in a safe operation. So, a smaller cap will give you a quicker attack but also a quicker release... The sc-amp ability to charge the cap fast is the main difference between the different varimus. Fairchild is the more powerfull so it is the quicker...

- The SSL network is a cap charged through a resistor (attack) and discharged through another one (release)... The charge resistor and the cap values are certainly the main way to get the attack time changing... I think it would be a lot simpler to reduce the charge resistor or the cap value to have faster attack times. But changing the cap value will change the release value too... But in the GSSL, there's no tubes, no grids, and you don't have the same limits in the release resistor choice...

  So to be clear , IMO the pimp on the GSSL to have faster attack times would be more to change the TC cap for a smaller one... and maybe to change the attack resistor, but you may find the TL074 limits ...

 

[radiance]

Thanks Lolo!

I don't know if the VCA can handle ultra fast attack times so I can't say it'll work or not  ???.

I think there will be distortion with faster attack settings. Did some tests with the fastest available attack and release and when compressing heavily I got some distortion in the low end. Faster will probabely be worse...

 

[Kingston]

I think there will be distortion with faster attack settings. Did some tests with the fastest available attack and release and when compressing heavily I got some distortion in the low end. Faster will probabely be worse...

AFAIK it's the VCA crapping out. It's not really designed for such fast speeds. I forgot a name of the company, but there's one very recent design SSL "clone" compressor out there that has discrete VCA's to overcome some limitations, a mastering unit.

 

[crisotop]

Vertigo VSC-2

 

[radiance]

Well, that Vertigo VSC-2 has the exact same attack times as the ssl...I would have expected something faster....

 

[crisotop]

Printed on the panel, it feels a lot faster (not sure if the ssl timeconstants are correct)

 

[analag]

Good God...It's time to release my ultimate Mu Compressor...the mind blower...and if you want sound samples I will have my man Volker host some serious crush, attitude, color, tube sound...fer yah. Seriously I don't want to give out that one.

 

[Kingston]

Sorry for off topic.

Well, that Vertigo VSC-2 has the exact same attack times as the ssl...I would have expected something faster....

That front panel is only some engineers' definition of what the attack time is. There is much more to attack than what is printed on front panel.

Printed on the panel, it feels a lot faster (not sure if the ssl timeconstants are correct)

Indeed. They probably only used the SSL standard time constants on the panel because their product topology is so heavily based in it.

But the attack shape might be completely different, resulting in far less distortion, and at least the illusion of faster speeds. This might not fit the original SSL designers definition of attack speed at all. He might have labeled a product like this very different.


I think similar needs were the starting point of this whole SC amp thing. Analag posted some pictures showing the much more even exponential slope a more powerfull SC amp could provide for time constants, again resulting to less distortion.

 

[lolo-m]

Good God...It's time to release my ultimate Mu Compressor...the mind blower...and if you want sound samples I will have my man Volker host some serious crush, attitude, color, tube sound...fer yah. Seriously I don't want to give out that one.

Do you have a problem with mods done on your stuff ?
I thought we were in a DIY forum  :... Or maybe my poor english is a problem to understand  what you mean...

 

[analag]

Good God...It's time to release my ultimate Mu Compressor...the mind blower...and if you want sound samples I will have my man Volker host some serious crush, attitude, color, tube sound...fer yah. Seriously I don't want to give out that one.

Do you have a problem with mods done on your stuff ?
I thought we were in a DIY forum  :... Or maybe my poor english is a problem to understand  what you mean...

I love it, but if you step up the design I'll come up with something else...and we keep it going. I push you and you push me, and in the end who knows.

 

[SUPERMAGOO]

you could do some tests with drums?.
I would like to hear the different mods with Buss-drums or just
bassdrum+snare.

eso lo grabaste en españa? saludos!

 

[lolo-m]

you could do some tests with drums?.
I would like to hear the different mods with Buss-drums or just
bassdrum+snare.

eso lo grabaste en españa? saludos!

Here it is ! On first post  ;D...
Sorry for being so slow to release those samples, but I had to fix a hum problem. This is out topic, but have a look with a scope at what's happening in the filament PSU. Push-pull technology is meant to cancel filament hum problem, but in mine it isn't  :'(...

 

[Moby]

This is out topic, but have a look with a scope at what's happening in the filament PSU. Push-pull technology is meant to cancel filament hum problem, but in mine it isn't  Cry...

Did I understand you correctly? You had a hum issue caused by filament supply? Did you solve that? How?

 

[lolo-m]

This is out topic, but have a look with a scope at what's happening in the filament PSU. Push-pull technology is meant to cancel filament hum problem, but in mine it isn't  Cry...

Did I understand you correctly? You had a hum issue caused by filament supply? Did you solve that? How?

Yes, you did understood correctly.
I tried all the possibilities I had. Unregulated and regulated. Regulated was better but not perfect. The transformer is roughly used on filament PSU. We are pumping 5,2A constantly, and the caps haven't got enough guts to handle that. There is something like 2 or 3 volt sawtooth on the DC...
The drums sounsamples are made with a regulated PSU based on 2 LM350. I had to add a cap (330uF) to the adjust resistor and went with four 4700uF caps ( and not 2 ) to limit the oscilation. Of course, I played with the transformers place and put a metal shield between the toroid transformer and the electronics...
  And the hum is still here  :-\, small but here...