GSSL HELP THREAD!!!

[greyskull]

Sorted!!!! the G connection between main board and daughter board was dodgy. FIXED
we have compression AND meter!

Now..... How do I light this damn LED backlight on the Hairball 8027-B 1mA Meter??

HEEEEELP!

 

[pedroplanet]

Hi, I built a GSSL on january of 2010 and it was working great until yesterday.  I was mixing and suddently it stopped compressing  :'(

- It has the CRC and the Turbo mods

The unit passes audio and the makeup gain works but no compression.
I Tried to remove/bypass the turbo but the problem persists.

after some time I discovered that if I leave the turbo plugged, sum the 47k resistors as the original and remove just the ratio connector on turbo It starts compressing. wierd! So I guess the turbo is ok, and the problem is somewhere on the sidechain or control board.

I checked the diodes, changed all the electros and ceramics on sidechain section, swapped/changed all the ICs and nothing worked.

Does anyone have any idea on whats wrong!?

 

[Nanitlig]

Thanks Harpo.  Should I be able to plug unbalanced cables in and have the same level at both sides without floating anything?  Does connecting ring and shield (1 and 3), or using TS plugs not work with ne5534 inputs?  I know most other compressors I've done mods to use duals, ne5532, LF353 or TL072 chips at the inputs.  Would the tilt have anything to do with both L and R colds connecting with the shield 0V.  For whatever reason the only way I am able to get both ne5534 to output the same voltage I have to lift the left channel cold.  Tried swapping IC's around, and actually I replaced the ne5534's at the input stages with LT1357, same pinout, single channel op-amp.  Should I use transformers to balance the unbalanced inputs?  Like Jensen JT-11P-1?

Nanitlig

 

[jjaskuna]

Thanks for the direction Harpo!

Using the technique you described I found that I have ACV present at Pin 1 and Pin 7 of my TL074 when a signal is present and none when the signal is absent. I also found that I do not have any ACV present at pin 1 (input) or 8 (output) of all 3 2180s when using the wire link between the 1000F capacitors and the 7915. However, if I move the black probe to my right + input, I find the same ACV on all 3 2180's pin 1 and pin 8. Using this method, the ACV measured at pin 7 of the TL074 is affected by the threshold pot. 

So does this mean that I do have signal going into my sidechain and that I should starting looking for the lack compression or make up gain else where? If so anyone have a thought on a likely cause?

Thank you all for your help on this.


Jeff

 

[Harpo]

Should I be able to plug unbalanced cables in and have the same level at both sides without floating anything?

Sure. Imbalance between L/R will be caused by differences in parts values, a broken trace/wire/connection or differences in signal level.

Does connecting ring and shield (1 and 3), or using TS plugs not work with ne5534 inputs?

Sure will work. Connecting ring and shield at a TRS socket is the same as inserting a TS plug into this TRS socket (except slightly higher contact resistance). Your previous post description is differing from this, talking about lifting the connection to the ring of your TRS socket.

Would the tilt have anything to do with both L and R colds connecting with the shield 0V.

No.

For whatever reason the only way I am able to get both ne5534 to output the same voltage I have to lift the left channel cold.

So with same signal feed and comparing both L/R outputs, your mod is giving same conditions when lifting the left channel cold. Conclusion would be, a wrong part value between right channel cold in and NE5534-pin3, or more likely a broken trace or not working connection in this area.

Tried swapping IC's around, and actually I replaced the ne5534's at the input stages with LT1357, same pinout, single channel op-amp.

These don't have the same pinout. (pin 5).

Should I use transformers to balance the unbalanced inputs?  Like Jensen JT-11P-1?

For usual not needed.

 

[PeteC]

Hi everyone,

Ive been using my GSSL very successfully for quite a few months now and would like to add a HPF board

I've searched for Steffen's HPF docs but the links seem to be broken.

I etch my own stuff or build on veroboard so no pcb needed.

Can anyone share the HPF details with me ( if there is a pcb layout i can etch thats even better !

Many thanks
PeteC
 

 

[Nanitlig]

Can anyone share the HPF details with me ( if there is a pcb layout i can etch thats even better !

Here's a simple HPF setup that's quick and requires no new boards.  Feed the signal from the two 47k resistors feeding the sidechain to one side of a DPDT switch.  The off position returns to the 15k resister at the sidechain input.  Connect the on position to the off position with the capacitor across the two contacts, on the switch itself.  When switched on the signal will return through the capacitor and when off the signal will bypass the filter.  I used a .047uF Wima 63V on mine, which will give you a 6dB/oct. at 225 Hz filter.  No resistor is needed on the switch, because the 15k at the sidechain input going to ground completes the filter.  The other half of the switch can be used to switch an LED if desired.

 

[wthrelfall]

Another question which I can't find the answer to:  I am using a plastic case for the GSSL, how do I attach the ground or do I need to?

You can't, but to make your box safe in case a mains wire inside comes loose, you'd have to connect this safety ground from your IEC to any conductive spot to the outside world (pot housing when metal knobs attached to metal/conducting spindle, sleve of unisolated TRS sockets, ..). This will not protect your circuit inside your plastic case from induced garbage (radiating transformer from a unit below/above, RF, neons, ...)

Harpo - So is it a bad idea to use a plastic case because it won't shield from interference?  I was going to use plastic because A: It's cheap B: I don't care that it doesn't look authentic (it's the sound I care about) C: It will be easier to cut out the holes etc.  But now I'm wondering if I've made a mistake, if noise/grounding is going to be an issue..

 

[PeteC]

thanks Nanitlig - I will try that first as a quick way to reduce pumping on drum bus
Rgds

PeteC

 

[Nanitlig]

Thanks Harpo!  I found a resistor on the right channel cold path that I had only soldered in one side... that'll do it!  :  I'm up and running now.  Much appreciated!!  I installed another 78L12 regulator for the relays and it's working great.

You bet Pete.  It's been working well for me.  You end up 6dB down around 112Hz and 12dB around 56Hz.  Switching in and out I notice about 6dB less reduction with the kick drum and big toms which is roughly the same change as when the detector only sees one channel.  I also have switching set up for the Expat Turbo board, which splits the sidechain to two seperate detectors.  Just waiting on the board... 

Nanitlig 

 

[wthrelfall]

Nanitlig, this sounds like a great mod - do you have a photo you can post?  I'm afraid I can only follow insctructions if I can also see what i'm trying to achieve  :-[

Can anyone share the HPF details with me ( if there is a pcb layout i can etch thats even better !

Here's a simple HPF setup that's quick and requires no new boards.  Feed the signal from the two 47k resistors feeding the sidechain to one side of a DPDT switch.  The off position returns to the 15k resister at the sidechain input.  Connect the on position to the off position with the capacitor across the two contacts, on the switch itself.  When switched on the signal will return through the capacitor and when off the signal will bypass the filter.  I used a .047uF Wima 63V on mine, which will give you a 6dB/oct. at 225 Hz filter.  No resistor is needed on the switch, because the 15k at the sidechain input going to ground completes the filter.  The other half of the switch can be used to switch an LED if desired.

 

[Harpo]

So is it a bad idea to use a plastic case because it won't shield from interference?  I was going to use plastic because A: It's cheap B: I don't care that it doesn't look authentic (it's the sound I care about) C: It will be easier to cut out the holes etc.  But now I'm wondering if I've made a mistake, if noise/grounding is going to be an issue..

D: It could swim in milk ...
This highly depends on your cases neighbourhood that might behave different again when you move your case some centimeters.

For your HPF question, something like this schemo snipplet with cap values for some common frequencies should do. Lots of ways to shave a cat. Using one for each L/R channel, instead of a single cap in the summing node might make it easier, if you want to add a turbo one day.

 

[Nanitlig]

Here's how I set up my HPF.  A SPDT will work for mono sidechain or a DPDT if you want to switch an LED also on the other half of the switch.

Unsolder the two 47k resistors from the board.  Solder both 47k to a lead going to switch COM.  Solder lead from OFF position to the 15k resistor at the sidechain input.  Solder capcitor across the switch from ON position to OFF position.  Pretty basic.  Only need one capacitor with mono sidechain.

.047uF = 225Hz
.068uF = 156Hz
.1 uF = 106Hz

 

[wthrelfall]

Amazing, thanks man..
;D

Here's how I set up my HPF.  A SPDT will work for mono sidechain or a DPDT if you want to switch an LED also on the other half of the switch.

Unsolder the two 47k resistors from the board.  Solder both 47k to a lead going to switch COM.  Solder lead from OFF position to the 15k resistor at the sidechain input.  Solder capcitor across the switch from ON position to OFF position.  Pretty basic.  Only need one capacitor with mono sidechain.

.047uF = 225Hz
.068uF = 156Hz
.1 uF = 106Hz

 

[Harpo]

... I used a .047uF Wima 63V on mine, which will give you a 6dB/oct. at 225 Hz filter.  No resistor is needed on the switch, because the 15k at the sidechain input going to ground completes the filter.  The other half of the switch can be used to switch an LED if desired.

The 15K (in series with the 100pF) between ground and virtual ground has nothing to do with your HPF response. You could leave these 106kHz LPF parts out when using THAT218x VCAs.
BTW, your 47nF between the 47K summing resistors and virtual ground is setting this HPF to 1/(2*PI()*1/(1/47000+1/47000)*47E-9)=144Hz, not 225Hz.

 

[Nanitlig]

Sorry, I used an online calculator.  At any rate, I like the .047uF, it is working well.  Does that not create a band-pass at the 15k?  Do you suggest removing the 100pF and placing a jumper, using 2180?  With the 15k there I will still have the HPF, correct?  When I set -6dB reduction at 1k and sweep down, reduction decreases starting around 300Hz, at 225Hz reads -5dB, 160Hz reads -3dB and 85Hz reads 0dB.  Then bypassing the filter returns the meter to -6dB at 85Hz??

Nanitlig

 

[Harpo]

Sorry, I used an online calculator.  At any rate, I like the .047uF, it is working well. 

If you like it, keep it. It's all about music, not numbers.

Does that not create a band-pass at the 15k?

Every gain stage in audio (not DC) is a bandpass to a more or lesser degree. There is no such thing as infinite bandwidth. A DBX2150 from the originally gssl circuit has 14MHz less bandwidth at unity gain than a THAT218x, so this 15K/100pF LPF @ 106kHz might be useful with this ancient part.

Do you suggest removing the 100pF and placing a jumper, using 2180?

IMHO you could leave the 15K resistor and the 100pF cap out when using a THAT218x.

With the 15k there I will still have the HPF, correct?

As already said, this 15K resistor is not part of your HPF. The resistive element of your 1st.order HPF are the 47K summing resistors and assume both R/L inputs of your stereo compressor are driven from a correlated signal source.

When I set -6dB reduction at 1k and sweep down, reduction decreases starting around 300Hz, at 225Hz reads -5dB, 160Hz reads -3dB and 85Hz reads 0dB.  Then bypassing the filter returns the meter to -6dB at 85Hz??

The HPF cutoff frequency is defined where your amplitude has dropped by -3dB (-45° phase angle). Your dB readings will change with a different ratio setting. There is no standard 1:1 ratio setting in the GSSL sidechain section.

 

[wthrelfall]

Harpo, would you be able to post a picture similiar to Nanitlig's one showing your method?
thanks!

Sorry, I used an online calculator.  At any rate, I like the .047uF, it is working well. 

If you like it, keep it. It's all about music, not numbers.

Does that not create a band-pass at the 15k?

Every gain stage in audio (not DC) is a bandpass to a more or lesser degree. There is no such thing as infinite bandwidth. A DBX2150 from the originally gssl circuit has 14MHz less bandwidth at unity gain than a THAT218x, so this 15K/100pF LPF @ 106kHz might be useful with this ancient part.

Do you suggest removing the 100pF and placing a jumper, using 2180?

IMHO you could leave the 15K resistor and the 100pF cap out when using a THAT218x.

With the 15k there I will still have the HPF, correct?

As already said, this 15K resistor is not part of your HPF. The resistive element of your 1st.order HPF are the 47K summing resistors and assume both R/L inputs of your stereo compressor are driven from a correlated signal source.

When I set -6dB reduction at 1k and sweep down, reduction decreases starting around 300Hz, at 225Hz reads -5dB, 160Hz reads -3dB and 85Hz reads 0dB.  Then bypassing the filter returns the meter to -6dB at 85Hz??

The HPF cutoff frequency is defined where your amplitude has dropped by -3dB (-45° phase angle). Your dB readings will change with a different ratio setting. There is no standard 1:1 ratio setting in the GSSL sidechain section.

 

[Harpo]

would you be able to post a picture similiar to Nanitlig's one showing your method?

Something like this.
Now compare this with the schematic snipplet from some post above and how this matches your pcb when following the parts and traces, so you can do this yourself in the future.

 

[jjaskuna]

Thanks to the direction I got on the forum, I was able to eliminate a variety of causes for my lack of compression and trace the problem to my bypass switch. It was not wired correctly. I figured this out by checking the pots I had connected with a continuity meter, which I had not yet purchased when I originally soldered the switch. After some rewiring and resoldering I now have compression!!! Very Happy. Unfortunately, I now have some gain problems. I was only able to mess with it for a few minutes this morning before work and could not get an accurate observation of what was happening. It seems like the signal is coming out EXTREMELY hot whether the sidechain is in or not.  I will get back to it after the holidays.

Cheers,

Jeff