Calibrating the G-SSL with THAT2181 - share thread !!!

[Emperor-TK]

What did you guys do on the sidechain VCA for the pin 5 resistor (Iset)? The Vee on this pin is -12V, vs. the -15V on the signal VCA's. It sounds like the original 3K9 resistor might be right here due to the lower Vee. But I can't see how the 5K1 resistor is derived in the first place. I get more like 6K3. Anyone have any thoughts?

 

[martthie_08]

nothing :grin: But you have a point there, I didn't bother looking at the sidechain vca yet. If I see it correctly, the sidechain vca is Ec- driven, so again page 6 of the design note should apply. According to page 2 the lower control voltage should only be relevant for the resistor R3.

I think the 47R resistor between pins 2 and 4 is not necessary (according to page 1 of DN137) but should pin 4 be connected to ground then?

regarding the resistor on pin 5 (R4): I think it should be 5K1 and I am not sure about that 10K resistor between pin 5 and pin 3 on the vca.

Maybe I will find some time upcoming weekend to do some testing. cheers, Marten

 

[Emperor-TK]

My ohms law calc giving me a 6K3 resistor of course is wrong. I was tired last night and assumed all V drop across the resitor. Duh. Anyway, I don't know how to spec the resistor here since I dont really know what's going on inside the VCA.

Marten, I already deleted the 47R. Pins 2 and 4 are connected inside the VCA through an internal 51R resistor. Just chop the 47R out and you are good. I did the same adjustments to the sidechain VCA as discussed in this thread until I noticed that it was running on 12V rails. Now I'm sort of stuck for the "right" thing to do until someone with better understanding chimes in. It's likely that these adjustments won't affect the sound of the unit, so I have no idea how to distinguish right from wrong myself. I guess I should just try the two different resistors and then measure voltage drops, hence Iset. I'm still afraid that I am missing the bigger picture though.

Has Jakob ever commented about the original designs? They don't exactly follow the 2150 datasheet either (10K between pins 3 and 5, 3K9 resistor on pin 5). The CV VCA was designed for the 2150 and not for the 212.

-Chris

 

[martthie_08]

My ohms law calc giving me a 6K3 resistor of course is wrong. I was tired last night and assumed all V drop across the resitor. Duh. Anyway, I don't know how to spec the resistor here since I dont really know what's going on inside the VCA.

:sam: You may be correct about having to adjust this resistor for different control voltages. In the datasheet it says the resistor on pin 5 determines the DC current bias - input / output signal current.

Here is my take on this, please correct me if I'm wrong: If for -15V the 5K1 resistor yields 2.4mA (according to the datasheet), then the voltage drop across this resistor would have to be 12.24V. So, to get a 12.24V - 3 = 9.24V voltage drop across this resistor, it would have to be 9.24 / 0.0024A = 3850 R, which is damn close to 3K9! So I think everything is fine with this resistor value after all - it is just "incorrect" for the signal path VCA's!!

Marten, I already deleted the 47R. Pins 2 and 4 are connected inside the VCA through an internal 51R resistor. Just chop the 47R out and you are good.

hmm, are you sure it's ok to chop the 47R out leaving pin 4 "floating"? I had a look at the 2181 datasheet, but still have some difficulties in understanding everything. I guess that the trimming works by applying a small offset voltage, so shouldn't pin 4 have a 0V referenced ground connection for this purpose? The internal resistor is only 25R by the way...

Has Jakob ever commented about the original designs? They don't exactly follow the 2150 datasheet either (10K between pins 3 and 5, 3K9 resistor on pin 5). The CV VCA was designed for the 2150 and not for the 212.

yes, it would also be interesting to see the original SSL schematic, how they implemented the sidechain 2150 VCA (with regards to the 10K resistor - what is that sucker for anyways :?.

cheers, Marten

 

[Emperor-TK]

Marten, I think you are correct. I messed a couple of bits of info up. I agree, pin 4 still needs to reference ground. I'll probably just leave the 47R. Also, the internal resistor is 27R. 51R is the value of the external resistor used in the 2150 designs.

Your derivation of the 3K9 resistor also works for me. I'm tired. Need to get some sleep... no more public mistakes :wink:

 

[Purusha]

:roll: 3K9 or 5K1? Which one do we put in at the end ???

 

[martthie_08]

in my opinion: for the sidechain VCA 3K9 for the signal path VCA 5K1

still not too sure about pin 4 of the sidechain VCA - to ground or not to ground :roll: maybe Jakob or SSLTech or someone else can give an expertise on this question?

The original sidechain circuit of 82e26 looks quite familiar, but there is a 202c in there, of course. So maybe the 10K is part of Jakob's 202 emulation? I get the feeling, that these issues should be resolved by ears rather than math, but whatever :?

 

[Greg]

There's been alot of talk about changing various resistors in the VCA emulation circuit or what's connected to pin 4 of the VCAs, etc... this is fine by me as it's good to try and improve things. But as you're changing resistors, are you performing listening tests or performing distortion calculations, etc. Basically are you testing to see if these changes are even worth the trouble?

I removed the resistor connected to pin 4 in the sidechain when using 2180LB, since the datasheet said pin 4 NC, in case anyone is interested. It just seems that the GSSL "mods" as they've been called are getting beat to death pretty hardcore.

 

[martthie_08]

sorry Greg, I have to disagree. The title of the thread states that this is about the trimmable 2181 series, it is correct that if using the pretrimmed 2180 series that pin 4 is NC.

:roll: I guess this is where all the talk comes from. There are quite a number of VCA options in the meantime, both modern and "vintage" and I think it is important to think about what needs to be done for the model VCA you are using in your compressor. When I first got my GSSL up and running I didn't find it to be very nice sounding, this is what got me started. As stated, I haven't tested the latest results from this discussion, but the initial changes I descibed on http://groupdiy.silentarts.net/martthie_08/gssl/ did improve things, just "testing" with my ears and applying recommendations by THAT.

For me this is particularly about the 2181LA, trying to get a transparent compressor that really has that SSL "stamp" characteristic on it. If using other VCA's you are in a different boat. The neat side effect of this research is that I am slowly beginning to get an idea of how this ciruit works. All this discussion may be a little confusing, but I am doing my best to document the important stuff on that web page for everyone to see.

edit: btw. that sidechain filter board is really nice, a friend of mine used the gssl with sidechain filtering extensively on a recent rock / country style mastering job and he was really happy with the results. :thumb:

cheers, Marten

 

[Emperor-TK]

[quote author="Greg"] are you performing listening tests or performing distortion calculations, etc. Basically are you testing to see if these changes are even worth the trouble?[/quote]

I'm more afraid of running the VCA out of spec and causing a premature death, latch-up problems down the road, etc. I don't want to have to revisit the "redesign" in a year when my brain turns back into mush. Particularly with the SC VCA, I don't expect any audible differences. What happens if Iset is too high, I don't know. Not being proficient at electronics, I don't know if any of my concearns are justified, but it's probably worth chatting about.

[quote author="Greg"] I removed the resistor connected to pin 4 in the sidechain when using 2180LB, since the datasheet said pin 4 NC, in case anyone is interested. It just seems that the GSSL "mods" as they've been called are getting beat to death pretty hardcore. [/quote]

I already built 2 units with 2180's which sound fine. There are a couple of extra issues with the 2181's though, as well as some still unanswered questions regarding the discrepancy between the original 202 to 2150 substitution and the datasheets. I'm still curious what the reasons are. Half the fun about building these things is trying to figure out what makes them tick. Your ear can't always help you there.

-Chris

 

[Greg]

Well put, Marten and Chris... :thumb:

I think I was kind of grumpy this morning and probably just didn't feel like reading about the GSSL.

 

[martthie_08]

Greg, :guinness: , I'm feeling a little grumpy myself today, just one of those days where I just cannot drink enough coffee... anyways, I think I can answer one of my "questions" myself, after poking around on the THAT corp webpage a little more (excellent reading, see DN124 and DN127).

still not too sure about pin 4 of the sidechain VCA - to ground or not to ground Rolling Eyes maybe Jakob or SSLTech or someone else can give an expertise on this question?

OK, if pin 4 is put to ground this will short out pins 2 and 4 and also the internal 27R resistor, bad idea. So Chris was right about ripping out the 47R resistor on pin 4 for the 2181 in the sidechain and NOT shorting it out.

A new question has come up though, from reading DN124:
The 100p caps on the opamps following the VCA could be as low as 22p, has anyone tried this out or is this a bad idea (oscillations)?

I am still not sure about the function of the 10k resistor between pins 5 and 3 on the emulation circuit, does it apply an offset voltage to the negative control pin :?:

cheers, Marten

 

[Purusha]

[quote author="martthie_08"]Greg, :guinness: , I'm feeling a little grumpy myself today, just one of those days where I just cannot drink enough coffee...
cheers, Marten[/quote]

Just chant a few rounds of Hare Krishna maha mantra and you'll be happy... it works every time, believe me :wink:


OK, just to be sure, we are talking about the red 47R below?

So we just rip this one out, no grounding or using jumpers, yes?

 

[martthie_08]

So we just rip this one out, no grounding or using jumpers, yes?

yes, I think that is correct, the red 47R gets taken out for the 2181 VCA - please report your results, I am curious whether this will make a difference!

cheers, Marten

 

[Purusha]

Thanks Marten. I'll do it in the next few days and report back if I notice any difference in the sound.

 

[mathflan]

my problem :


On my SSL when I put no compression, threshold at +20 and I put Make up at 0, the sound when the compressor is active is lower than when the comp is bypass

do you know where this problem come from??"

...is resolved

In fact , Purusha had right, I have trouble with +4Dbu and -10 Dbu setting.

When I connect My soundcard (set to +4dbU) to the SSL and connect the SSL to My monitoring speakers, I don't have Loss volume.
But When I put my SSL to a Mixer inputs (-10dbu) I have a loss volume, LOGICAL....

:grin:

Thanks to everyone,

Pics of my SSL soon....

 

[Emperor-TK]

OK, I'm calling my GSSL done. I did the adjustments discussed earlier in the thread for the signal VCA's. For the SC VCA, I deleted the 10K and the 47R on pin 4. I also used the original 3K9 too. Grounding pin 4 was not an option. It caused my GR needle to peg.

As is, everything seems fine with the build. The two units I built with the 2180's are not here for comparison, but from memory, it seems like I can push this one a little harder without crapping out.

-Chris

 

[Purusha]

Hi Chris, so this is how your VCA sections look right now:

Signal VCA

SC VCA

 

[martthie_08]

ok, I have tested the calibration steps and they all work fine.

Regarding the sound, it seemed a little more transparent at comparable gain reduction readings on the meter, but I'll have to check if the meter is still giving me a precise measurement. Not a big change, maybe it is easier to compare if you have two units side by side.

I finally got some webspace, so I'll move the site from Volker to my new place: http://homepage.mac.com/marten.thielges/gssl/

cheers, Marten

 

[Purusha]

I wonder, can anyone answer to my question about the VCA sections above?