GSSL HELP THREAD!!!

[Winetree]

Thought this would help but I Just looked at Lukas Color Road Map concerning
the balanced input and outputs.
Looking at the color codes for inputs and outs the colors change between
schmatic and board layout and don't even jive with the + and - codes on the board.
Compare the two and you'll see what I mean. I've got headers already solded and
can't read the board. What's  the correct way to hooks up the balanced ins and outs?

 

[mitsos]

positive input on the 5534 is pin 3. Trace back from the two pin 3's through the lytics and 22K resistors to your input header and those are your + inputs. The ones next to them are the - and the middle is ground (FWIW, I did not connect these to anything).

Or just check the ssl.pdf on gyraf's site. Lucas may have mis-labeled (or been inconsistent in his coloring) but the channels are correctly matched. That is to say, the innermost ones go together as do the outermost ones. But on the SSL pdf the right channels are the innermost two on both headers, the left are the outer two.

 

[Harpo]

positive input on the 5534 is pin 3. Trace back from the two pin 3's through the lytics and 22K resistors to your input header and those are your + inputs. The ones next to them are the - ...

The pcb silkscreen is correct.
In-phase signal (+/hot/XLR-pin2) connects to NE5534s non-inverting (+) input and out-of-phase signal (-/cold/XLR-pin3) connects to NE5534s inverting (-) input, in order to get the correct polarity out when your signal is polarity inverted by the audio-VCA and again by the I2V conversion stage, this also being the in-phase (+/hot/XLR-pin2) or unbalanced output.

 

[johnmellor]

Hi there I've just turbo'd by gssl and just purchased Lundahl LL5402's on the outputs to top it all off! Other than replacing the four 100 Ohm resistors with 22 Ohm, can I just follow the manual diagram?

I'm sure I read somewhere regularjohn recommending these?

 

[beatnik]

Hi everybody!

I need some help with my GSSL built from gyraf project. The clone isn't passing audio but the power section seems to be fine as the led lights up and the voltages from regulators seem to be ok.

Here are my voltage readings on the opamps, with no input signal. Please help me finding where some errors could be.

L&R inputs NE5534

pin1 -- 11,97
pin2 -- 0
pin3 -- 0
pin4 -- -15,09
pin5 -- -0,63
pin6 -- 0
pin7 -- 14,46
pin8 -- 11,96

L&R outputs NE5532

pin1 -- 13,76
pin2 -- -0,05
pin3 -- 0
pin4 -- -15,10
pin5 -- 8,10
pin6 -- 9,34
pin7 -- -13,86
pin8 -- 14,45

NE5534 on VCAs

pin1 -- 11,62
pin2 -- 0
pin3 -- 0
pin4 -- -15,12
pin5 -- -0,89
pin6 -- 0
pin7 -- 14,45
pin8 -- 11,65

TL072 & TL074 (some pins change voltage depending to the threshold pot)

pin1 -- -0,09 to 3,13
pin2 -- 0
pin3 -- 0
pin4 -- -12,18
pin5 -- 0
pin6 -- 0
pin7 -- 0
pin8 -- 11,89

pin1 -- -0,14 to 0,20
pin2 -- 0
pin3 -- 0
pin4 -- 11,89
pin5 -- 0
pin6 -- 0
pin7 -- -0,03 to 0
pin8 -- 0
pin9 -- 0
pin10 -- 0
pin11 -- -12,18
pin12 -- 0
pin13 -- 0
pin14 -- 0,22

i have made some more measurements.

I have a problem on the +12V rail for sure, because i have +22V on 78L12 input, but on the output of the regulator i have a voltage that oscillates from about -0,5 and +3V.

Could be a damaged regulator or something else?

i am still trying to debug my gssl, but i need your help.

i discovered a short between pin 7 & 8 of the sidechain VCA, and replaced the 78L12 too.

now +12V rail is fine too.

at this point the compressor is passing audio, so i made a step.

but the controls are not doing anything! only the threshold seem to affect the meter, but not the sound. and there is no difference in sound between compressor in / bypass. only the meter switches on / off.

the audio coming from the output seems distorted too... and sometimes i hear some clicks and pops on the speakers.

i double checked again for shorts and cold soldering joints, and wrong component values, but now i'm sure everything is ok.

any help is highly appreciated

 

[Harpo]

Beatnik, did you already try to fix it step by step ?
http://www.groupdiy.com/index.php?topic=47.msg483540#msg483540
Your +14.46 supply rail still looks suspicious. For usual these fixed voltage regulators aren't that far off their target voltage.
First fix it for a clean audio path. With no signal in, NE5532-pin7 should not throw out -13.86V DC (the 100uF electrolytic caps at output might already be blown from reversed polarity), so this chip might be broken, running open loop, the audio-VCA is broken or there are other shorts or wrong value parts in this area.
Good luck

 

[beatnik]

Hi Harpo, first of all thanks for giving help.

I've redone the voltage measurements, with no input signal. Now they look like this:

L&R inputs NE5534

pin1 -- 12,44
pin2 -- 0
pin3 -- 0
pin4 -- -15,06
pin5 -- -0,62
pin6 -- 0
pin7 -- 14,92
pin8 -- 12,49

L&R outputs NE5532

pin1 -- 0
pin2 -- 0
pin3 -- 0
pin4 -- -15,06
pin5 -- 0
pin6 -- 0
pin7 -- 0
pin8 -- 14,92

NE5534 on VCAs

pin1 -- 12,38
pin2 -- -0,10
pin3 -- -0,10
pin4 -- -15,08
pin5 -- -0,72
pin6 -- -0,10
pin7 -- 14,92
pin8 -- 12,40

TL072 & TL074

pin1 -- -0,10 to 3,13
pin2 -- 0
pin3 -- 0
pin4 -- -12,14
pin5 -- 0
pin6 -- 0
pin7 -- -0,30
pin8 -- 12,05

pin1 -- 0,18
pin2 -- 0
pin3 -- 0
pin4 -- 12,05
pin5 -- 0
pin6 -- 0
pin7 -- 0
pin8 -- 0
pin9 -- 0
pin10 -- 0
pin11 -- -12,17
pin12 -- 0
pin13 -- 0
pin14 -- 0,37

SIDECHAIN 2180

pin1 -- 0
pin2 -- 0
pin3 -- 0,23
pin4 -- 0
pin5 -- -2,93
pin6 -- 0
pin7 -- 12,04
pin8 -- 0

L&R VCAs

pin1 -- 0
pin2 -- 0
pin3 -- -0,02
pin4 -- 0
pin5 -- -2,92
pin6 -- 0
pin7 -- 14,93
pin8 -- 0

I tried to follow your procedure, but my problem is i don't have an audio generator device. Maybe can i use Logic Test Oscillator plugin?
Please let me know if you can spot other strange voltages.
Thanks

 

[Harpo]

I tried to follow your procedure, but my problem is i don't have an audio generator device. Maybe can i use Logic Test Oscillator plugin?
Please let me know if you can spot other strange voltages.

Rail voltages look OK and the strange readings at the NE5532 pins are gone. Hope, your output caps have survived, if previous numbers were for real. Nothing more to look right now without signal present. Some opamp pins are for compensation or are current, not voltage. Your oscillator plugin should do the job just fine. Use a frequency, your multimeter can handle. When measuring with removed and jumpered sidechain VCA, remember to drop your signal voltage by at least 20dB, as this procedure is breaking the sidechain feedback loop.

 

[beatnik]

ok, so i have to set the test oscillator on two outputs of my audio interface, then going to gssl inputs.

then i should set my multimeter to VAC reading, but i don't know how to measure AC voltage... is like DC voltage so i have to measure between ground and ic pin?

sorry for this noob question...

 

[Harpo]

Most often, yes.

 

[beatnik]

perfect, tomorrow i'll try to follow your instructions and let you know what happens.
thanks again

 

[beatnik]

Ok, i'm following your procedure, but i don't understand how to set the level of the Test Oscillator plugin, since we are talking about dbV but i suppose the plugin scale is dbFS. anyway here is what i've done until now.

- if i set the level to 0dB (400Hz sine wave) both NE5534 pin 6 measure 0,66VAC

- then if i pull out audio VCAs both pin 7 on NE5532 measure 0,32VAC. pin 1 too and xlr positive pin.

- pull out sidechain VCA. pin 1 on TL074 is reading 0,08VAC and pin 7 0,06VAC. at the 20k/10k junction i measure 0,02VAC

- at varying ratio settings, pin 14 measures respectively 0,15 0,77 and 1 VAC

- then other measurements varying attack, release, threshold and makeup don't produce any change in voltage, so i guess now it's the big problem

Please let me know if i'm doing the procedure correctly, and other suggestions for debugging the compressor. thanks

 

[Harpo]

- at varying ratio settings, pin 14 measures respectively 0,15 0,77 and 1 VAC

- then other measurements varying attack, release, threshold and makeup don't produce any change in voltage, so i guess now it's the big problem

Signal flow obviously gets interrupted.
Continue following TL074-pin14.
Diode orientation between pin14 and 'C' OK?
Check bypass wiring / switch connection / lorlin end-stop washer, if your bypass switch is a lorlin type switch ('D' will have varying voltage, depending on ratio setting)
TL072-pins 1 and 7 should have varying voltage when bypass disabled and dialing threshold or makeup pot, else check the +12V and -12V connections at control pcb.

 

[beatnik]

diode orientations are ok

bypass wiring is ok, used lorlin rotary wired as gyraf layout and checked connections with multimeter

ratio setting doesn't influence the voltage on D point of connector (pin10 on TL074). it is always 0

TL072 pin1 varies with threshold pot, from -0,08 to 3,12 VDC
TL072 pin7 varies with makeup pot, from 0 to -1,92 VDC

i made some other tests, feeding the compressor with an audio track. the sound seems ok, but still no difference between compressor engaged / bypassed. and it's the same if i pull out all the VCAs and jumpering pins 1 & 8. maybe i have broken VCAs?

 

[beatnik]

according to the "voltage test points" pdf posted by mac on the previous page, i should read -8.7V on pin 5 of the sidechain VCA, but i have -2.9V.
i have -2.9V too on pin 5 of the audio VCAs, where i should read -15V instead. i have -15V before 3K9 resistor, but after the resistor the voltage drops to -2.9V.
other problems....

 

[Harpo]

according to the "voltage test points" pdf posted by mac on the previous page, i should read -8.7V on pin 5 of the sidechain VCA, but i have -2.9V.
i have -2.9V too on pin 5 of the audio VCAs, where i should read -15V instead. i have -15V before 3K9 resistor, but after the resistor the voltage drops to -2.9V.
other problems....

The readings from mac are with chips pulled out of socket.

Your control signal is still interrupted between 'C' and 'D'.
Follow the pcb traces until you lose your signal and fix the missing connection, broken part or trace just before this spot.
Is your control pcb already mounted to a frontpanel ?
(These lorlin type switches have a plastic nose supposed to go into a hole in a frontpanel to prevent a turning of the switch body. This is not needed here, but when you fix the nut too hard, this off center nose might internally bend the switch so that some pins could lose contact. For usual I cut this nose off or file it down. May not be your concern.)

 

[beatnik]

i followed the pcb traces between C and D and everything seems fine, and checked the lorlins where making the right contacts.

bypass switch makes connection between C and attack switch.

attacks switch works ok, switching on the six resistors.

the other side of the attack resistors (all connected together) goes to D and so to pin10 in the TL074, and to the release switch.

the release switch is ok.

maybe should be the tantalum caps?

 

[beatnik]

i had broken VCAs... replaced and now everything works fine!!

thanks everybody in the community especially Gyraf for developing this free project and Harpo for his support in the debugging process.

Long live DIY!

 

[ilfungo]

sorry I have a strange problem ....
I measured all the voltages and they seem ok except on the side chain VCA pin 5  7.7v instead of 8.7v.
I made measurements without IC and VCA
another question
lorlin switches have a orientation?
thanks

 

[Matthew Jacobs]

lorlin switches have a orientation?
thanks

Well yes and no.

The don't have a orientation as in it doesn't matter which way round they go but on the GSSL you have to make sure that the pins go in the hole that has a trace connected to it. That's all.

So in a 2x6 lorlin switch, the PCB has 4 holes but only two have traces. make sure the switch pins go into those holes.

Sorry can't help you with the voltages. Although it's normal to have a voltage on pin 5 of the VCA. Check the schematic, you'll see that pin 5 is connected to the -15V supply via a resistor.

All the best with your build, just check and double check everything. I had a issue with my GSSL not compressing and it turned out that one of my Lorlin center pins wasn't soldered. So maybe the problems you are getting are there, if you got their orientation wrong. Just double check.

J