GSSL HELP THREAD!!!

[gyraf]

Where can I start to look for the error?

The best thing would be to get the basic GSSL working correctly, and then take the add-ons one at a time.
Error-tracing easily gets too complex. Temporarily install the two 47K's where you've connected sidechain pcb, and take it from there.
_____________
firfe,

The design is known-working, which means that there IS an error somewhere. Check, compare with schematic, and re-check.


Jakob E.

 

[Ericbazaar]

The best thing would be to get the basic GSSL working correctly, and then take the add-ons one at a time.
Error-tracing easily gets too complex. Temporarily install the two 47K's where you've connected sidechain pcb, and take it from there.

Hi again,

ok i deinstalled the sidechain board. solder all to basic gssl mode! The same problem, sound going in, sound came out and makeup gain works, but threshold, attack and release without function. i hold the pcb under the light and look for solder bridges, nothing! test all cable connection 3 times...all ok (i think). the vca's not seem to work? how can I test the motherboard?

many thanks

 

[submergent]

I think this thread isn't quite long enough...  so i'll add to the general confusion

bought a "non-working" build a month ago. redid the i/o and power connections and bam. worked like a charm.
added the Expat Audio Turbo board... even better... when i say "worked like a charm" - i mean i mixed 11 out of 13 songs of an album with this thing on the 2mix. it sounded just great.

then this problem surfaced:
the unit would "die" after a little while using it. LED would go out, meter would lock halfway through it's throw, etc.
i thought it was heat related, since when i had the top off to troubleshoot, it almost never happened. almost.

i benched it again and replaced the +/- 15v and +/- 12v regulators with new ones, and when testing... it dies within 30 seconds all the time now. (this is again, on the bench, top and bottom lids off). When i say dies, i mean the +12v rail is going down. -12, and +/- 15 are all still good.

i took the turbo board off the +/- 12v and gnd connections, and redid the sidechain connections (i.e. undid the mod)  and unit looks good for a while longer... but still goes down. I metered the +12V line and paid attention to the voltage:

with the threshold set all the way CW, i can watch the +12v line slowly, but steadily drop from 11.88v to 11.6ish in anywhere from 5 minutes to 30 seconds, depending on how long the unit has been on...and bam. out it goes (the +12v rail). with the threshold fully CCW,  voltage drop still happens, but way slower. removing various chips (VCAs, TL074, etc) i still see voltage drop, but at a slower rate.

fully CW on the threshold pulls down the +12 rail.

any ideas where to look ? anyone see this before?

thanks all!

-pete

 

[Harpo]

pete, this is heat/load related. Easiest way to fix it might be to remove the 10R resistor in front of the 78L12 from component side and refit same resistor from trace side, but now connecting between the 78L12 input (freed up component hole next to 22uF cap and 78L12) and the regulated +15V (rightmost pin of the 7815), so it sits diagonally in parallel to the 0V trace. Supplying the Power-On indicator LED from +15V instead of +12V with a piece of wire and a maybe 1K5 current limiting resistor in between (remove the 1K on control pcb) will free up another 20% of connected load to the 78L12.
Wild guessing you are powering the unit with dual 18VAC or 36VAC center tapped transformer secondaries instead of dual 15VAC or 30VAC center tapped transformer secondaries. This voltage regulator has to dissipate the differential between rectified raw DC voltage at input and +12VDC out, leaving less than 50mA for the connected load (complete sidechain section of GSSL, turbo board and power-on LED) before the chips thermal protection shuts the vreg down.

 

[submergent]

pete, this is heat/load related. Easiest way to fix it might be to remove the 10R resistor in front of the 78L12 from component side and refit same resistor from trace side, but now connecting between the 78L12 input (freed up component hole next to 22uF cap and 78L12) and the regulated +15V (rightmost pin of the 7815), so it sits diagonally in parallel to the 0V trace. Supplying the Power-On indicator LED from +15V instead of +12V with a piece of wire and a maybe 1K5 current limiting resistor in between (remove the 1K on control pcb) will free up another 20% of connected load to the 78L12.
Wild guessing you are powering the unit with dual 18VAC or 36VAC center tapped transformer secondaries instead of dual 15VAC or 30VAC center tapped transformer secondaries. This voltage regulator has to dissipate the differential between rectified raw DC voltage at input and +12VDC out, leaving less than 50mA for the connected load (complete sidechain section of GSSL, turbo board and power-on LED) before the chips thermal protection shuts the vreg down.

thanks Harpo - i'll try that.
the trafo i'm using is 15/30 - but first thing i'll try is getting that LED i added on the 12 out of there.
the 1K on the control board - when i take the LED out, that resistor gets removed completely? anything else need to be done in that part of the circuit? when i first opened this project up, the original builder had a diode in place of the LED.

-pete

 

[Harpo]

the trafo i'm using is 15/30 - but first thing i'll try is getting that LED i added on the 12 out of there.

Good move. With your dual 15VAC transformer in front at nominal line voltage (whatever this might be, please update your profile with your location data), the 78L12 will see about 20.5V raw DC at input and this 8.5V differential times current load is transfered to heat at the vreg. The latest revision already has prementioned fix done on pcb. Whatever you additionally connect to this rail will challenge the 78L12.

the 1K on the control board - when i take the LED out, that resistor gets removed completely?

Leaving a one side connected resistor on pcb doesn't make much sense. Take it off, so you might reuse it elsewhere.

when i first opened this project up, the original builder had a diode in place of the LED.

at least he wasn't freezing when sitting in the dark ...

 

[submergent]

harpo -

ok - took the LED off the 12v line - put it on the -15v rail - tested.
looked ok. i could still see a slow drop in the +12v rail, but slow enough i thought it may be ok.
hooked the +/- 12v lines to the turbo s/c board. tested: same thing... with the threshold CW, the +12v drops continually to 11.6 (or so) V and boom. out it goes.

something else has got to be buggered here... my next thought was fly-wiring the 12v regs off the pcb and mount them to the case somehow to dissipate the heat. that can't be right though.

other ideas?

-pete

 

[Dissler89]

The -from pic- square right side pin above the 'N' pin will be the 'L' pin for the fused IEC socket.
Fuse parameters needed....

Thanks for the heads up. I decided to order a DPDT Switch to make sure both the life and neutral are switched simultaneously.
I would imagine the law here in Switzerland is most likely similar if not the same.
I edited the picture to include what my transformer say's on the side and how I wired it up  from the information given. I'm around 98% sure I've wired it correctly.
I wasn't aware that the paint/coating/anodizing could affect the conducting so I will review this on my case, thanks for that!

 

[Harpo]

other ideas?

Leaky 22uF cap located between 78L12 out and 20K resistor ?
Tantalum caps on control pcb connected the wrong way round ?
You already double checked parts values and orientation (especially the threshold pot and parts connecting to it) ?
Distant mounting of the vreg requires cap between Vin and 0V reference voltage close at the vreg.
Using the 7815 as preregulator for the 78L12 by moving the 10R as previously described still seems the safer bet to me. YMMV.
From your previous wording 'LED i added on the 12' you seem to have more than one LED connected to the +12V rail.
Good luck.

 

[Ericbazaar]

The best thing would be to get the basic GSSL working correctly, and then take the add-ons one at a time.
Error-tracing easily gets too complex. Temporarily install the two 47K's where you've connected sidechain pcb, and take it from there.

Hi again,

ok i deinstalled the sidechain board. solder all to basic gssl mode! The same problem, sound going in, sound came out and makeup gain works, but threshold, attack and release without function. i hold the pcb under the light and look for solder bridges, nothing! test all cable connection 3 times...all ok (i think). the vca's not seem to work? how can I test the motherboard?

many thanks

Hello,

No Help? Please... :-\

Ok, i know this problem came very often to, but i can't read 300 of sides in this Threat. My English is not very good. i read many sides and some solutions were discussed. many errors are: bad soldering, bad TL074, wrong input wiring, the voltage regulators 7915, 7815, 7812 ect.
I have all checked and i find a micro solder bridge on the 56k resistor next to TL072 opamp, but it was not :-[!
The Compressor is still not working, but in my Test Something has happened. i came with my fingers back on the pins from the Ratio switch and the compressor reduce the gain...this is very strange. There is a checklist for the compressor, with the correct voltages? i have checked all resistors, 4 1k of the vca board resistors do not show the correct values (ca 760ohm) but these are 1k resistor! I do not know what to do :'(

-eric​​

 

[audiophreak]

Ok MeT002,

I found it 

Your hint about the ratio switch connection helped!

Good.

Is it correct -  what I found out in this thread in regards of doing b) is:
I have to change both of the 15k resistors highlighted in the attached pic to 27k
and disconnect the (-) from the outputs?

Would be great to know before I order the parts - just in case there's something else I have to be aware of...

FYI Changing those 15K to 27K resistors are effectively adding an extra slug of ~+6dB make up gain to the +ve phase, to compensate for the lack of any -ve phase signal (-6dB).

You could consider using a 22K resistor + a 10K multi-turn trim pot in series so you can fine tune the 0dB gain point on both channels. Or you could just play about with a couple of fixed value resistors in series to make around 27-33K until it's correct. Or you could just put in the 27K and your make up gain calibration might be slightly off. Some people worry about being 1dB out, but others don't. it's all good, and won't effect the compressing character at all.

  Hi All,
          I am wanting to use one of my Dual GSSL's  in Inserts of a live console - I found this with search, and it was reported back that it works , what wasn't mentioned was the INPUT - do I just drop the signal (-)  on the Input as well ?? 

  Thanks ,
                Chip

 

[submergent]

other ideas?

Leaky 22uF cap located between 78L12 out and 20K resistor ?
Tantalum caps on control pcb connected the wrong way round ?
You already double checked parts values and orientation (especially the threshold pot and parts connecting to it) ?
Distant mounting of the vreg requires cap between Vin and 0V reference voltage close at the vreg.
Using the 7815 as preregulator for the 78L12 by moving the 10R as previously described still seems the safer bet to me. YMMV.
From your previous wording 'LED i added on the 12' you seem to have more than one LED connected to the +12V rail.
Good luck.

youre the man Harpo -
i think the move of the 10R to link the output of the 15v reg to the 12 did the trick.
+12v looks nice and steady now.
dare i re-assemble and rack it??

why not.

thanks again!

-pete

 

[Harpo]

...what wasn't mentioned was the INPUT - do I just drop the signal (-)  on the Input as well ??

Whatever the meaning of 'drop the signal' in your language might be,
you connect the 0V reference voltage of the sending device (shield connection of your insert send TS or TRS socket at your mixing desk) to this out-of-phase input terminal on GSSL pcb. The balanced line receiver in the GSSL circuit operates the differential between input+/-.

 

[Harpo]

just to clear .. is this the correct wiring ?

Neither input nor output look correct in my book. Maybe reread my previous reply.
There is no mention of connecting the out-of-phase inputs to the reference voltage center pin on pcb, except you want to build a ground loop on purpose. If you use shielded mic cables inside the gssl, connect shield at one side only.
The NE5532s on GSSL out will not be amused to see their output nearly shorted out to 0V reference voltage.

 

[audiophreak]

Thank you for your time Harpo , this isn't the first time you've helped me , however i am still unclear and its painfully obvious I cant seem to formulate a proper question , hopefully this drawing will help in what I am asking and dont understand

 

[Harpo]

Sleeve of TRS to input '-'. The gssl exclusively operates the differential between input '+' and input '-'. No connection at 'ø' needed. Same as for a balanced signal connection where XLR-2 connects to input '+', XLR-3 connects to input '-' and XLR-1 connects to case in the shortest possible way.
If you use shielded mic cable inside the gssl, 'ø' connects to shield of mic cable with the other end of the shield left unconnected, IE not connected to S.

 

[audiophreak]

.. and the black veil of uncertainty is lifted and I see the light !!    ;D    Thanks so much Harpo !!

 

[audiophreak]

... well ... the light has dimmed a little , so in the drawing I posted and your description of connection ... that would effectively connect Output 0v to Input (-) at the Sleeve of TRS 

maybe this, if you care to explicitly spell it out for me :

Tip goes to - ?
Ring goes to - ?
Sleeve goes to - ?

GSSL Input (+) connects to - ?
        Input (-) connects to - ?
        Input 0v connects to - ?

GSSL  Output (+) connects to - ?
          Output (-) connects to - ?
          Output 0v connects to - ?

  I so apologizes for my just not getting this  :-\

 

[Harpo]

... well ... the light has dimmed a little , so in the drawing I posted and your description of connection ... that would effectively connect Output 0v to Input (-) at the Sleeve of TRS

yepp

depending on your mixing desk,
some (FI Soundcraft) have tip-return and ring-send, others (FI Behringer) just the other way round,
description for the 1st.
Tip goes to - return
Ring goes to - send
Sleeve goes to - screen

GSSL Input (+) connects to - R
        Input (-) connects to - S
        Input 0v connects to - nada

GSSL  Output (+) connects to - T
          Output (-) connects to - nada
          Output 0v connects to - S

 

[audiophreak]

Thank you so very much Harpo !!