GSSL HELP THREAD!!!

[metalb00b00]

Why would you expect compression to change significantly with timing?

Because with faster attack and release time, signal amplitude that goes beyond the threshold get chopped off more often, while with a slower attack and release time, it get chopped off less???

At what analogue input level? The comp is designed for levels around +4dB over 775mV ac

My Avid 192 I/O analog lines are calibrated at +4dBu (1,230mVAC or 1.23VAC)
I'm testing the compressor using a generated stereo 1KHz test tone at 0dB sine wave RMS signal level (according to Adobe Audition)

stereo test tone --> 192 I/O analog out 7-8 --> comp stereo in
comp stereo out --> 192 I/O analog in 7-8

Check that you have both timing caps in place for auto (0u47 and 6u8)

Tantalum caps are in place, with correct value and orientation. I have replaced these 0.47uF/35V and 6.8uF/35V four times now, and it still distorts the signal, so I just don't use the auto release timing.

in haste

Appreciates it

 

[metalb00b00]

PROBLEM SOLVED.

Borrowed a digital precision multimeter from my friend, and found out that my line level was at -10dBV instead of +4dBu. I forgot that a few years ago I had wired my analog 7-8 input output to work with unbalanced lines *d'oh! slap face*

Auto release still don't work without distorting the audio signal, despite changing the components several times, hence I adjusted my rotary switch to 2x4 instead of 2x5, making the auto release option unavailable.

Cheers Jakob! ;D

 

[gyraf]

Because with faster attack and release time, signal amplitude that goes beyond the threshold get chopped off more often, while with a slower attack and release time, it get chopped off less???

Dosn't work that way. Sidechain timing is about how fast to regulate all once over threshold.

Tantalum caps are in place, with correct value and orientation. I have replaced these 0.47uF/35V and 6.8uF/35V four times now, and it still distorts the signal, so I just don't use the auto release timing.

There is very little going on in this part of the circuit - switch goes to two cap/resistor pairs in series from ground. Shouldn't be rocket-science to debug.

Jakob E.

 

[metalb00b00]

There is very little going on in this part of the circuit - switch goes to two cap/resistor pairs in series from ground. Shouldn't be rocket-science to debug.

Could it be because I used two 180K resistor in a parallel configuration to get 91K?

EDIT: Found the culprit. Broken cable leading from the rotary switch 5th position to the control board. Replaced with a new one, and no more distortion for auto release

 

[WNStudios]

Sorry if this have been asked, actually i know it has once but he did not get an answer.

However i just bought a fully populated GSSL Rev 7 and picking up where he left off does create some confusion on my part. So i'm hoping someone can point me in the right direction.

The unit is compressing, and it does sound lovely. But even though the VU meter indicates different release times, different attack times and different ratios, i dont hear any difference. It's not behaving like it's supposed to and like it's showing. The VU meter can be slammed, but the sound i get is still a very transparent 2:1-type compression.

Then there's the bypass switch/makeup gain. At first when i put it in bypass i got this terrible noise and extreme volume, but lifting leg 4 on the 2180's seemed to do the trick so that is gone. But the Makeup gain only works when it's in bypass, isnt it supposed to do the opposite?

Also the release caps he put in there are Electrolytic, not Tantalums, would that make a huge difference?

 

[MeToo2]

Sorry if this have been asked, actually i know it has once but he did not get an answer.

However i just bought a fully populated GSSL Rev 7 and picking up where he left off does create some confusion on my part. So i'm hoping someone can point me in the right direction.

The unit is compressing, and it does sound lovely. But even though the VU meter indicates different release times, different attack times and different ratios, i dont hear any difference. It's not behaving like it's supposed to and like it's showing. The VU meter can be slammed, but the sound i get is still a very transparent 2:1-type compression.

Then there's the bypass switch/makeup gain. At first when i put it in bypass i got this terrible noise and extreme volume, but lifting leg 4 on the 2180's seemed to do the trick so that is gone. But the Makeup gain only works when it's in bypass, isnt it supposed to do the opposite?

Also the release caps he put in there are Electrolytic, not Tantalums, would that make a huge difference?

Picking up a broken build requires experience of electronic debugging. If you don't have these skills you are going to need a lot of luck.

There are a few ways to debug this. First point is to use a standard test tone on the input for initial debugging: not music.
Doesn't really matter what it is but 1KHz 0dB will work. You'll also need a good high impedance voltmeter, or a scope.
Earlier in the thread I posted some test point measurements from my working build. You can search for that.
It gives test tones at various points through the schematic [to check e.g. that your de-balancer and balancer is working].

In your case since you are passing tones end to end, it's probably best to start by testing the threshold & make up gain controls with a static mono test tone on the input.

Moving the threshold pot from fully cw to ccw should cause DC voltage at point 'f' on the schematic to move from around -12v to ~+0.3V DC.
The TL072 should scale this linearly and add it to any existing control voltage to give a _change_ at pin 1 of approx 56/220 *12V = ~+3V DC as you move the pot from one extreme to the other.

Once you've got that control voltage working, you can make sure that it is having an effect on the side chain VCA.
AC Signal at Pin 7 of the TL074 in the sidechain should vary a lot as you change the threshold pot.



Now for the make up gain. Moving the make up gain pot from fully cw to ccw should cause a change at point 'e' on the schematic of around +12v to ~+0.0V DC.
The TL072 should scale this linearly and add it to any existing control voltage to give a _change_ at pin 7 of approx -100/620 *12V = ~-2V

Check that the bypass switch is wired the correct way around. i.e. in non-bypass mode, DC voltage at point 'e' should change, and you should see AC tone at the pole of the 'ATTACK' switch. In 'bypass' mode, DC voltage at point 'e' should not change as you move the pot, and you should see 0V AC at the pole of the ATTACK switch.

Once you've got that make up gain control voltage working, you can make sure that is having an effect on the main chain VCA.
Signal at Pin 1 of the 5532 should vary a lot as you change the make up gain.

That's the basics: you'll then know that your VCA's are working and 2 controls are working. You'll still have to check timing and ratios.


To quote wikipedia: The tantalum capacitor distinguishes itself from other capacitors in having high capacitance per volume and weight. Tantalum capacitors have lower equivalent series resistance (ESR), lower leakage, and higher operating temperature than other electrolytic capacitors, although other types of capacitors are even better in these regards.

Electrolytics will probably not sound the same in this application. I have to wonder what other shortcuts were taken in the build.

 

[gyraf]

Again, MT2, thanks for a good and well-put answer. if we had an "Employee of the month" award, I'd certainly nominate you 

The tantalum capacitors in the release section are primarily because this type dosen't change capacitance over time (electrolytics will slowly and continuously drop in capacitance when getting old).

Jakob E.

 

[WNStudios]

I have very very little experience, my old man is an electronics engineer but i thought that for once i dont wanna just give it to him to fix it, i wanna be able to do it myself. So i hope i have luck.

The volage i get from F on the controlboard ranges from -12 to +09.52v.
Voltage from pin 1 of the 072 is ranging from -0.700 to +3v.
And on the 7th pin of the 074 i hardly get any changes at all. -0.004 to -0.005 is what i get.

Faulty TL074 or does it indicate something else?

Thank you very much for taking the time to explain this to me, it takes more time than just sending it away to my old man but it's so much more valuable for me

 

[WNStudios]

Ok the bypass switch was indeed wired backwards so now i got makeup gain, works fine on every setting except Auto release, where it distorts the more i gain.

I dont have a scope so i'll have to do my best with the multimeter.

I'll swap the electrolytics for tantalums, but except that all components are up to spec, WIMA's on everything and good quality overall.

 

[Harpo]

... except Auto release, where it distorts the more i gain.

Confirm, the Lorlin end-stop washer is set to pos.5 after dialing this release rotary switch full CCW.
Confirm, you have the 91K resistor in parallel to the 0.47uF cap and the 750K resistor in parallel to the 6.8uF cap (as shown on the silkscreen and contrary to the schematic) and with these polarized caps fitted in their correct orientation.

 

[WNStudios]

... except Auto release, where it distorts the more i gain.

Confirm, the Lorlin end-stop washer is set to pos.5 after dialing this release rotary switch full CCW.
Confirm, you have the 91K resistor in parallel to the 0.47uF cap and the 750K resistor in parallel to the 6.8uF cap (as shown on the silkscreen and contrary to the schematic) and with these polarized caps fitted in their correct orientation.

The lorlin was set to pos 6 so i changed that.
It was indeed both 91k and a 750k, the 6.8uf cap was also right and i have now swapped all of the 47uf's on the controlboard to tantalums instead of the electrolytics. When on auto release the makeup gain makes no difference until almost full, where it starts to sound distorted and loud in a beat, the more i gain the faster the beat goes. On all other settings i have no problem.

I also swapped the TL074 and TL072 just to make sure they werent broken, same thing. I took out the resistors suggested by this site: http://diy.fischerworks.com/gssl_vca.shtml but no change. I've done new measurements (and correct ones, i was a bit too tired the last time) and i dont see much changes at all no matter where the threshold pot is set.

 

[Harpo]

...and i have now swapped all of the 47uf's on the controlboard ....

Maybe a typo, but there is no 47uF on the controlboard. These are 0.47uF (=470nF).

 

[WNStudios]

...and i have now swapped all of the 47uf's on the controlboard ....

Maybe a typo, but there is no 47uF on the controlboard. These are 0.47uF (=470nF).

Ah! That might be it then, the electrolytics was 47uf so i stupidly thought that was what i were to put there. Oh man i hope the old man has a stash of them at work

 

[Harpo]

The volage i get from F on the controlboard ranges from -12 to +09.52v.

You already double checked ALL resistor values? FI your measured voltage range at 'F' indicates, the 47K resistor connecting to the threshold pot more likely being a 4K7 resistor instead. This might not be the only one with a digit error...

 

[WNStudios]

The volage i get from F on the controlboard ranges from -12 to +09.52v.

You already double checked ALL resistor values? FI your measured voltage range at 'F' indicates, the 47K resistor connecting to the threshold pot more likely being a 4K7 resistor instead. This might not be the only one with a digit error...

Checked half of them so far and all of them have been right, so far. The 47k on the pot i'm 100% sure is right because i soldered that in myself as he had not I am gonna swap pots though because he did put some flimsy ones in there.

 

[Spaturno]

6dB difference sounds suspiciously like only half of your output driver is working on one channel.
The left one looks dodgy.

When you say you measured 0v on the XLR relative to pin 1 I presume that was DC (otherwise you wouldn't hear any signal at all).
Have you measured AC?
A signal should be easily measurable with an AC multimeter of quite a few mV when you send a 100Hz or 1KHz test tone at -10dBm.

If you've got a scope, trace out the signals, especially looking for differences between the two channels for (ac) measurements made on pin 6 & 7 of the two NE5532's.
That will tell you if the problem is on the input (10K from the +ve phase) or on the output (-ve phase).
Check for shorts and open circuits around the NE5532 (I remember the board is pretty tight in some spots there)
Check the wiring of your output XLR's to make sure the hot (pin 2) & cold (pin 3) are wired correctly, and are not shorted to pin 1 (multimeter in resistance mode with gssl powered off) and that they are connected to the 100uF electrolytic output caps.
Check the 100uF electrolytics are inserted the correct way around.
You could also try swapping the two NE5532 chips over to see if one of them has failed (presuming you've used sockets).

Thanks MeToo2

I swapped ICs - no change

Checked polarity of electrolytics - good

I measured AC at the output XLRs aginst pin 1 and read:

(L output) pin 2-1.905 VAC and  pin 3- 1.950VAC
(R output) pin 2- 1.965VAC and pin 3- 0.001VAC 

but from the header pins I measure:

(L output) pin 2-1.891 VAC and  pin 3- 1.896VAC
(R output) pin 2- 1.847VAC and pin 3- 1.843VAC


With the multimeter in resistance mode with gssl powered off, I checked for shorts and have shorts between pins 1, 2 and 3 on all inputs and outputs. They seem to rise slowly over time in the Mohm range (capacitance?)

As far as I can tell they are connected tot he 100uF caps.

I have the same problem measuring AC on the output XLR, i double checked the wiring, and cables (all chain is balanced), on the header, 100r resistor up to pin1 on 5532 i read the same value 0VAC. i have allreadx try to swap cables and 5532 with no result.
Somebody could help me? thanks in advance.

 

[MeToo2]

Sorry for the break in the thread here. I tried to scour for the info I am looking for but after 200 pages or so I got a massive headache. My question is simple, but I am not a tech, so can't figure it out for myself.

Can the GSSL be used with just one channel connected. i.e. for tracking mono instruments?

Yes. That will work fine. Either left or right. Not important which. You'd probably want to plug a dummy source (1Kohm resistor across XLR pin 2-3) in the other channel to stop any noise.

and if it has Turbo board, can that also be used with only one channel plugged in?

thank you much
jon

Do you mean GSSL + Turbo board, but only one mono input connected thus compressing a single mono instrument? Yes, that'll work fine. Just as above.

Or do you mean GSSL + Turbo board acting as a dual mono, with the two mono inputs acting entirely independently to compress 2 different mono instruments like a UA 2-1176 in "dual" mode? No, that won't work.

A compressor consists of the main chain, a side chain (rectifier/detector) and a timing chain. GSSL has 2 main chains, 1 side chain 1 timing chain.

The Turbo board provides a second side chain for detection for "improved" stereo operation, but there's still only one single (common) timing section that controls both main channels.

You could theoretically also duplicate the timing section for dual mono mode operation, and switch the control signals in as appropriate, as well as adding a Turbo board, but then it wouldn't be anything like a GSSL any more.

 

[Harpo]

I have the same problem measuring AC on the output XLR, i double checked the wiring, and cables (all chain is balanced), on the header, 100r resistor up to pin1 on 5532 i read the same value 0VAC. i have allreadx try to swap cables and 5532 with no result.

Your problem seems different with 0VAC on the header. Member wahfreak only seems to have a crimp-connector not making contact to a pin of the header, so the signal didn't arrive at the output XLR.
Double check for shorts. (maybe at the 10k resistor leg close to the left side bal.output +/- connection). Using a magnifying glas might help to spot it.

 

[aldaraia]

So I'm currently putting together a mouser cart for the GSSL and plan on building it over the course of the next few weeks as parts come in and I have the money for components, but I've come across a bit of a snag.

Right now I'm thinking about having the compressor mimic the look of the bus compressor on the original SSL consoles: gray face, gray/gray+blue knobs, buttons that light up (currently thinking red but I may go for yellow, like the original compressor). The problem is that I discovered that illuminated switches require a different method of switching and powering, with relays and such, but after hours of research I have not been able to find any specifics on HOW to accomplish it. Haven't found a schematic, haven't found the right switches for the job, or appropriate relays. The nearest I've come to solid information is the bypass n' more PCB that was hinted at but apparently never followed through with.

I am just looking for some definitive information on wiring this. This is the last part of the process for me before I actually get the parts and components and start putting the thing together and it is definitely the most frustrating part because I cannot find ANYTHING on it. Any information would be helpful, but total information would be the greatest thing ever to happen to me.

 

[Spaturno]

I have the same problem measuring AC on the output XLR, i double checked the wiring, and cables (all chain is balanced), on the header, 100r resistor up to pin1 on 5532 i read the same value 0VAC. i have allreadx try to swap cables and 5532 with no result.

Your problem seems different with 0VAC on the header. Member wahfreak only seems to have a crimp-connector not making contact to a pin of the header, so the signal didn't arrive at the output XLR.
Double check for shorts. (maybe at the 10k resistor leg close to the left side bal.output +/- connection). Using a magnifying glas might help to spot it.

Yes my problem is different, i found 3 short and now the signal is well balanced, thanks for the tips. Only the hum at about -50db is still there, maybe is the print tranformer. Do you think a Toroidal Tranformer has less hum?