GSSL HELP THREAD!!!

[AmosMoses]

Hey all, first post.

Have tried to read lots and lots of posts and it made me understand quite a lot about how these electronics should or shouldn't work and at the same time made somewhat more confused so I will give a go and try to describe my problems and hopefully it will all be ok (meaning that I didn't miss any search stuff )

Have built with PCBgrinder pcb.

1. The unit powers up, does not compress (the meter is always at 0), and all the controls have no impact on anything, bypass, sidechain switch, attack, release, ratio, treshold and makeup do nothing.

2. The unit passes sound, but at a very low volume, I have checked the audio signal with audio probe and the signal is ok till it reaches 27k resistors before going into VCA. and when it hits the output it's a much quieter 30 or more -dB and it sounds like something from a reverb tank and a big low pass filter on. When powered touching pin 1 on VCA there is no sound but unpowered there is none tha same on sidechane THAT and pin 1 (trought the audio probe).


3. And interestingly if I put it into audio interface the input dB levels are normal near 0 in the daw but as I said the sound is very quiet and the input can easily oscalate even on the low imputs, i can only connect to imputs with pads pushed otherwise it instantly oscalates.

I'm using THAT2181, and this layout http://www.gyraf.dk/gy_pd/ssl/ssl.pdf. My THAT's are not pretrimmed and I have no way currently to trim them, maybe this is the problem?



1. All the voltages according to this http://www.gyraf.dk/gy_pd/ssl/ssl_sch.gif  -12 or +12 and -15 and +15 are correct.

2. Tried bridging the vca pins 1 and 8 without THAT, the output and the sound is the same as with THAT inplace

What should I do, where should I look, I suppose first thing I should get is a good signal then i can continue to the compression problem.

Sorry if I missed a post where someone had a similar problem.

 

[Script]

2181, OK. A? B? C?
Definitely not 2180s?

Look at 'Matt Fischer DIY Audio' page for correct implementation of the 2181X. You might be missing resistors around the VCA.

 

[AmosMoses]

I'm sure.  THAT2181 A

I'll look into and try to install all those missing parts. What kind of a trimmer should I use? regular 50k, is it linear, log or what? What about the sidechain THAT2181? there is no trimmer there, should I just order the 2180 and clip the 4 pin?

 

[Script]

Ordinary trimmer. Linear. Not even sure there are log trimmers.

Correct ICs installed everywhere? Type and orientation.

Your output wires are  connected correctly?

And the absolute classic with this board:.  check for solder joints / connectivity with meter. Some wrong or missing joints are close to impossible to see, especially in the denser parts of the board. I use Chinese magnifying googles from eBay for that

While at it, check for connectivity and maybe even replace that resistor sticking out at the button left of your picture, close to the audio output.

 

[AmosMoses]

Installed a jumper wire, 680k resistor and a trimmer to the VCA circuit according to the 'Matt Fischer DIY Audio' page. Now I get unison level in one channel output as in inputt but with low pass filter and in the other I get a distorted as if playback is through blown speaker, but connection on the xlr output nr 2 is amplified and without the low pass and middle pin gets the same level as the other channel and with low pass too but same as connection nr 2 very much distorted.

To sum up, my changes did this

1. channel one, unison level input and output but with low pass
2. channel two, sound is distorted as if the speaker would be blown. pin 2 on the output amplified without low pass, pin 3 same as input, but distorted and low pass.

I checked the signal before the pins on the output jacks it's the same.

All that said, the distortion meter doesnt move and any of the controls have no effect on the sound.

Maybe I have to instal one more 47r resistor near the sidechain for THAT2181, currently there is only one 47r and 10k.


Good news I think, at least theres something new

 

[AmosMoses]

Had some progress since last time.

Checked all the connections on the board with magnifying glass, changed THATs on VCA and SC and now I get compression and all the control board works + the SC and BP switches except the treshold potentiometer.

The output is still quiet and with the damn low pass filter... but not distorted anymore. I tried changing every IC with a new one except the input ones (because, i get a good signal until the VCA) with no difference.

I get too much compression, only when using slow attack and fast release i get compression around 10 -dB using simple drum samples, on other settings the needle stays at 20 dB with some movement bellow. When i turn on the the hpf the compression reduces drastically about 50%. but I guess this has something to do with the treshold settings, and until i get them working it's really hard to tell anything.

Any tips how should i proceed next?

 

[Script]

Check your THR pot for connectivity and observe voltage on wiper when turning the pot.

The threshold is kind of sensitive. It can be adjusted to need though. Use the search function.

Only you know what your HPF frequency is. But the more bass you filter out of the SC signal, the less there is for the comp to act on, meaning less compressing action.

 

[AmosMoses]

An update to the situation.

with a big red face i admint it was a bad orentation of one of the capasitors. The one that was in the way to the +12 sidechain. Replaiced it with a new, putted in a new VCA replaced opamps and what do you know the compressor works. all the controls and everything is working as it should it think

Now I have one problem left. I still get the low pass applied to the the signal. The signal is good until it reaches the 27k resistors to the VCA, on the output the signal has a low pass filter. Tried replacing opamps, tried bridging the pin 1 and 8 with sidechain disconected or connected it doesnt matter it still has a big low pass filter on. And this could be related maybe, my one output is lower, the one on the bottom left corner - output signal is much lower then the rest of them.

To sum up, signal after the VCA is with LPF with or without VCA inserted. If I turn off the compressor the signal is not amplified but no LPF at least

Measuring the AC voltage on xlr pins the three are almost identical if not and the one is 2 times lower.

any ideas, this has to be something easy compared... damn the thing works but there is something in this simple audio signal path I cant find

 

[Script]

Good to hear 

Check wiring to your high pass filter. Wrong wire connection on the switch?

And I'd check all components in the faulty channel again. Correct cap values? Resistors read same in circuit between L &R ? You say one of your channels is working, right ?

 

[AmosMoses]

i have tried to remove the HPF switch and just short the connections and the same result. without the sidechain the result is the same too..

 

[AmosMoses]

Good to hear 

Check wiring to your high pass filter. Wrong wire connection on the switch?

And I'd check all components in the faulty channel again. Correct cap values? Resistors read same in circuit between L &R ? You say one of your channels is working, right ?

I will recheck and resolder the quite channel tomorrow, will see what that does, because values I think i covered, but still will double check even

What do you mean resistors read same in circuit between L&R? No, both channels are not working as should, because from somewhere they both have this LPF on them... and in the attachment i pointed which xlr output is low.

 

[Script]

You swapped out that resistor mentioned earlier -- the one sticking out near the output?

Also the 100pf caps in the output stage are all the correct values? Looks like you have at least two , if not three different makes there, one blue, the other amber. Correct values? Too high a cap in the negative loop of an opamp or a cap to ground after a resistor acts like a LPF.

Also check the resistor values in the outputs. Resistors while they are in the circuit interact with other components, so might not measure their nominal values. The L and R outputs are exactly the same circuit, so corresponding resistors should read the same (whatever that is!). Examples: the two 15Ks in the NFB loop of the 5532s on L and R both read 14.96k over here, while the 470Rs read 470 and 469, and the two 10Ks near the head of the opamps read 9K2 and slowly rise to 9K96 and 9K98. All close enough, so they don't need replacing.

You can't measure caps this way though.

 

[AmosMoses]

You swapped out that resistor mentioned earlier -- the one sticking out near the output?

Also the 100pf caps in the output stage are all the correct values? Looks like you have at least two , if not three different makes there, one blue, the other amber. Correct values? Too high a cap in the negative loop of an opamp or a cap to ground after a resistor acts like a LPF.

Also check the resistor values in the outputs. Resistors while they are in the circuit interact with other components, so might not measure their nominal values. The L and R outputs are exactly the same circuit, so corresponding resistors should read the same (whatever that is!). Examples: the two 15Ks in the NFB loop of the 5532s on L and R both read 14.96k over here, while the 470Rs read 470 and 469, and the two 10Ks near the head of the opamps read 9K2 and slowly rise to 9K96 and 9K98. All close enough, so they don't need replacing.

You can't measure caps this way though.

Thanks for the info, will try to look into it more, hopefully I find something, but I already checked 3 times

How do you measure these things when the resistors are soldered? With power on or off?

Yes, I have changed that ugly resistor to a new shiny one

 

[Script]

I measure when powered off. Just stick metre probes on resistor leads from top of board. Easy.. Checking all resistors in the output stage should take a few minutes only.

 

[AmosMoses]

You swapped out that resistor mentioned earlier -- the one sticking out near the output?

Also the 100pf caps in the output stage are all the correct values? Looks like you have at least two , if not three different makes there, one blue, the other amber. Correct values? Too high a cap in the negative loop of an opamp or a cap to ground after a resistor acts like a LPF.

Also check the resistor values in the outputs. Resistors while they are in the circuit interact with other components, so might not measure their nominal values. The L and R outputs are exactly the same circuit, so corresponding resistors should read the same (whatever that is!). Examples: the two 15Ks in the NFB loop of the 5532s on L and R both read 14.96k over here, while the 470Rs read 470 and 469, and the two 10Ks near the head of the opamps read 9K2 and slowly rise to 9K96 and 9K98. All close enough, so they don't need replacing.

You can't measure caps this way though.

pF 100 should be market as 101 as oposed to 104? which would make them 0.1 uF or 100000 pf? Just want to make sure if this is the silly thing that I did before resoldering all the capasitors?

Resistors are +- correct, besides the one 15k, which rises above 15k1

the 470 are 460 in mine. and the 10k are good.

 

[Harpo]

...I still get the low pass applied to the the signal.

As script already said, your cap values might be off. The only barely readable yellow cap value right side next to the TL072 looks like 104. If so, this 100nF (not 100pF) will set the bandwidth of this opamp stage down by factor 1000. Same applies for the other stages.

The signal is good until it reaches the 27k resistors to the VCA

These VCAs are current in/current out devices (not voltage). The following opamp for the current to voltage conversion will do its best to hold both of its +/- input terminals at same potential and as the positive input is tied to 0V reference voltage, the negative input node of this opamp hopefully will read the same (called virtual ground).

... Tried replacing opamps, tried bridging the pin 1 and 8 with sidechain disconected or connected it doesnt matter it still has a big low pass filter on.

confirming, this most likely is caused by forementioned cap values with maybe factor 1000 value off.

And this could be related maybe, my one output is lower, the one on the bottom left corner - output signal is much lower then the rest of them.

If this level difference is factor 0.5 (=-6dB), this most likely is caused by a short at the last inverting opamp stage, so check for an unwanted blob of solder. Its a little tight at some spots.
(the current setting resistor for the sidechain VCA in your build looks like 5K1. Should be 3K9)

Good luck.

 

[AmosMoses]

As script already said, your cap values might be off. The only barely readable yellow cap value right side next to the TL072 looks like 104. If so, this 100nF (not 100pF) will set the bandwidth of this opamp stage down by factor 1000. Same applies for the other stages. These VCAs are current in/current out devices (not voltage). The following opamp for the current to voltage conversion will do its best to hold both of its +/- input terminals at same potential and as the positive input is tied to 0V reference voltage, the negative input node of this opamp hopefully will read the same (called virtual ground).confirming, this most likely is caused by forementioned cap values with maybe factor 1000 value off.If this level difference is factor 0.5 (=-6dB), this most likely is caused by a short at the last inverting opamp stage, so check for an unwanted blob of solder. Its a little tight at some spots.
(the current setting resistor for the sidechain VCA in your build looks like 5K1. Should be 3K9)

Good luck.

Thanks for the detailed response, very informative and useful

The factor ir more like 0,3. But I will recheck anyway. So it's 104 across the board... Well that's a bit shameful, but I'm very glad that this is it most likely.

The resistor 3k9 sidechain even when using THAT2181? maybe I should change sidechain to 2180?

 

[Gustav]

The resistor 3k9 sidechain even when using THAT2181?

Sidechain is supplied by 12V, audio path VCAs by 15V, hence the difference.

Gustav

 

[AmosMoses]

Sidechain is supplied by 12V, audio path VCAs by 15V, hence the difference.

Gustav

Do I need to clip the 4 pin or what? Or just connect as is?

 

[AmosMoses]

These VCAs are current in/current out devices (not voltage). The following opamp for the current to voltage conversion will do its best to hold both of its +/- input terminals at same potential and as the positive input is tied to 0V reference voltage, the negative input node of this opamp hopefully will read the same (called virtual ground).confirming, this most likely is caused by forementioned cap values with maybe factor 1000 value off.

If its not too much to ask maybe you could explain this a bit more? For future and general knowledge, Im quite a noob regarding this but it's very interesting and extremely good to know for future builds.