GSSL HELP THREAD!!!

[Harpo]

Only the hum at about -50db is still there, maybe is the print tranformer. Do you think a Toroidal Tranformer has less hum?

Most often, yes (and assuming your build is a standard GSSL build without additional bells and whistles that might draw more current than your 4.5VA transformer can supply without getting stressed). You could unsolder your EI block transformer from pcb, put it on some piece of perfboard and mount it off board inside your 19" case. Distance is your friend. For transformer hookup have a look at the GSSL mainboard pcb traces. Look twice and build it safe. AC mains can be lethal.

 

[LesPaul]

Hello

A few years ago I built a GSSL with the turbo mod. It didn't work right away, and my daughter got sick, so I lost my motivation and I hid it away for a while.

After building a few preamps and some microphones, I thought I would go back to my old friend (or enemy) the GSSL.

I turned it on last night, and the meter immediately goes to max and then all the way down. After that it does nothing. The unit passes sound, but it is heavily distorted. The next thing is that the sound gets WAY higher through the unit, even if the makeup gain is set to the lowest setting.

What I have established so far, is that the PSU section is working.

Not trying to get a troubleshooting manual, but I was wondering if any of you could maybe point me in the right direction?

Pål N.
A little bit lost...

 

[gyraf]

Pål,

Check your front-panel pcb - that the wire link (shown as a line above the connecting wires) is correctly mounted, and that the connections from main- to-front boards are not reversed or wrong.

Check and recheck with PCB drawings AND schematic (pdf's on our website)

Jakob E.

 

[beatnik]

I've finally completed my gssl. It has CRC board, sidechain filter, led meter to display gain reduction, and a vu meter on the output.

It's working as it should, however it has an issue. Sometimes when the unit is powered on, something happens on the 12V rail, because the led meter displays 6dB gain reduction (three leds light up) and the illuminated switch for bypass is not illuminating.

Do you have an idea?

 

[gyraf]

Maybe a voltage regulator "hanging" at power up? We've had that issue before in this thread, I think (with specific brand 78L12/79L12/7815/7915's)

Jakob E.

 

[beatnik]

I remember that, actually I did the mod putting extra diode and capacitors on the back of the board.

But then I've installed the crc board... maybe the problem is back again?

 

[Harpo]

I remember that, actually I did the mod putting extra diode and capacitors on the back of the board.

The 7815/7915 anti-latchup fix or the fix for all 4 rail voltages ? (Your LED is powered by the +12V rail, so a 7815/7915 latchup fix won't help this 12V rail. Maybe read pages 268-269 again).

If any of your additional circuits (sidechain filter, led meter, illuminated switches, ...) is powered by the GSSLs + or - 12V rail, the 78L12 and/or 79L12 regulator might be in self protection mode due to overload.

 

[beatnik]

I did the 7815/7915 anti latch up, not for all the four rails

In fact I think the problem is happening on the +/-12V rail, because what don't work are the audio circuit, the relay+ illuminated switch for bypass, which are connected to the +12V rail

The LED meter and the VU meter light have their own 7812 regulator, and they work as they should.

But obviously since the sidechain has some kind of problem, the led meter displays compression.

Reading the previous pages I've found this solution.

http://www.hausverwaltung-heger.de/al_leck_trick/GSSL_anti-latchup.gif

Is that what I need? At the moment I don't have the diodes on the 7812/7912 so maybe is that the problem?

 

[gyraf]

First identify which power line you loose when it malfunctions. Then change that regulator to a different brand.

I have had no luck with latch-up prevention like the shown, with 1N400x diodes.

Jakob E.

 

[beatnik]

Which brands are proven to have not this problem?

 

[3nity]

Hi people.
Building a gssl using THAT 1246 (-6db) for inputs and 1646 (+6db) for outputs.
For that purpose i left the input and output sections blank on gssl motherboard.

Now is there 15K to change (unity gain) or do i left that resistor jumpered??

1 more question if i cant get unity gain i wont be able to calibrate meter isnt?

Thanks.

 

[Harpo]

Building a gssl using THAT 1246 (-6db) for inputs

for a little less circuit tweaking I'd chose the THAT1240 instead. With the 1246 you now need to adapt this by -6dB reduced level for the sidechain (47K resistors) as well.

and 1646 (+6db) for outputs.

why not, good choice.

For that purpose i left the input and output sections blank on gssl motherboard.

The VCA following opamp stage is still needed. The NE5532 on output most probably will not be that amused when you dremel half of this dual opamp away for the still needed current to voltage conversion, so assuming you will keep it as is. No need to connect a load to the last inverter, but might come handy for switched signal inversion as you don't want this opamp stage left unconnected.

Now is there 15K to change (unity gain) or do i left that resistor jumpered??

With your -6dB in, +6dB out parts, you want unity gain for the audio-VCA/I2V-conversion, so the feedback resistor (15K) will need to change to same value as the resistor in front of the audio-VCA (27K). Depending on type of VCA you might decrease these now 4x 27K resistors for lower resistive noise because you dropped level in front of the VCA by factor 2.

1 more question if i cant get unity gain i wont be able to calibrate meter isnt?

Meter reflects VCA control voltage, generated by the sidechain. Nothing to do with audio path unity gain.

 

[3nity]

Harpo thanks for chiming in!
Yes i got THAT1240 at 0db...less tweaking!
I will use only 1 NE5532 on the vca follower using both halves on both channels. (nice??)

VCAs are 2181LB with suggested 120K instead of 47K.
Any opinions now?

 

[Harpo]

Yes i got THAT1240 at 0db...less tweaking!
I will use only 1 NE5532 on the vca follower using both halves on both channels. (nice??)

nice, cannes, st.tropez - should work.
With the now 0dB THAT1240 in front, Rin (former 27K in front of the audio-VCAs) might decrease to ((Vcc-2V)/SQRT(2))/0.001A=~9K2, pick a 10K (for 8.6dB less resistive noise, at cost of THD).
As the THAT1646 will increase signal by 6dB, the audio-VCA following I2V converter needs to drop signal by this same amount for overall unity gain in VCAs bypass condition (Ec=0V), so former 15K feedback resistor will be 10K/2=5K. Lead compensation cap across this 5K might increase to 220pF.

Sidechain-VCA Rins (former 2x 47K) might decrease to 2x 15K and Rfb in the following I2V stage (former 33K) might decrease to 15K/SQRT(2)=~10K. The THAT2181 could be driven harder, but the following TL074 will not drive a much higher load with Rins of following fullwave rectifier in parallel to this 10K. Lead compensation cap across this 10K might increase to 220-270pF. You might as well keep this sidechain section as is.

VCAs are 2181LB with suggested 120K instead of 47K.

no idea what the reason behind this suggestion might be or what you're talking about...

 

[3nity]

Matt fischer website was suggesting 120K here:  http://diy.fischerworks.com/gssl_vca.shtml

 

[Harpo]

Jorge, some typos seem to have an infinite lifetime...
How would a single 120K resistor be a substitution for the two 47K summing resistors and still keep the L/R channels separate or fighting each other ? (further ignoring the doubtfully wanted sidechain level drop by 11dB, caused by this increase)

 

[synnys]

Ok, so i just finished my build and one thing i immediately noticed before connecting any audio cables is the following:

There's a -5V at my meter, and thus the needle goes way below zero. Also my power led is not working, only about 0,1V there...

I checked for bad soldering with an extremely bright flashlight, almost like a röntgen-picture there were none.

Any tips on where i'd best  start troubleshooting? 

Edit: when i flip the bypass switch, the meter goes to zero 0, flip it again and there's my -5v again. powerled stays off.

Edit 2: I had the 10pin multi cable to the front panel reversed... problem solved, hope i didn't damage anything

Edit 3: all is fine now just writing about my problems seemed to help a lot

 

[3nity]

thank you very much harpo.
YOu have helped me lots of times!
GSSL is working and sounding better than before!
Thanks

 

[WNStudios]

I'm going insane...

I have checked every resistor on the board, i've swapped all the IC's and even changed from 2180 to 2181 THAT's. I've made sure the electrolytics are the right value and direction. Diodes are right. Even lent it to my old man so he could check for solderbridges under a good magnifying glas and he found 2 which he fixed. But still nothing!!!

It is working, i can hear it compressing a tad, the compression charecteristics does change when i change ratio, release and attack but it's so minor and absolutley not what it shows on the meter. Any ideas?

 

[Harpo]

I have checked every resistor on the board, i've swapped all the IC's and even changed from 2180 to 2181 THAT's. I've made sure the electrolytics are the right value and direction. Diodes are right. Even lent it to my old man so he could check for solderbridges under a good magnifying glas and he found 2 which he fixed. But still nothing!!!

It is working, i can hear it compressing a tad, the compression charecteristics does change when i change ratio, release and attack but it's so minor and absolutley not what it shows on the meter. Any ideas?

and from a previous post "... so now i got makeup gain, works fine on every setting ...", the sidechain and your audio-VCAs are working, so
-> double check the '127K*' resistors real world parts value (located between sidechain-VCA and TL074) and check, there are no broken traces connecting to this resistor or cold solder junctions. For increasing compression (or adjusting ratio) by decreasing this resistor value, you might substitute this '127K*' with a maybe 82K-91K in series with a 50K trimmer/rheostat, initial setting about centered.