PCBs for 1176-flavoured FET compressor w/ opamps - (almost) no wiring

[fripholm]

I don't know what exactly you want to achieve with this circuit on breadboard. Don't you have my boards?

You shouldn't have to attenuate anything. This is taken care of by the input control. If calibrated correctly, the unit should start compressing at a certain level which means that the signal within the sidechain gets large enough to pass the diodes D1 and D2 which increases the DC voltage on the gate of the FET which in turn decreases the signal level on its drain.

FWIW, my mono unit starts compressing when the signal level at the drain is about 90mV pk-pk.

 

[mgb54]

TillM provided me with some more silkscreen options for front panels (for Hammond boxes) and I made a few slight measurement corrections to his 1RU version. Now it should fit Frank's NRG cases (frontpanels.de) perfectly and/or Don Audio's V-Cases.

The FPD files are now included in the guide archive. See the link in the first post.

HI Fripholm! If using 1U rack case based on this FPD file you provided, do you know what size standoffs are needed for the PCBs?
Cheers

 

[fripholm]

It depends on the dimensions of your particular case and how far the bottom panel is from the bottom of the boards and how thick it is. My design for the Hammond case was using 8mm stand offs. Till might know if he had to change that for the 1U file when he designed it.

You might not even need stand offs. If the potis and the ratio switch are screwed to the front panel it could be sufficiently stable.

 

[tedsorvino]

Hi Fripholm. Since I'm collecting the components before I start building I would like to ask you if I can use the 2sk30A Fets (that I already have) instead of 2n5457? I've read that they can be used as substitutes (depending on the use in the circuit).

 

[fripholm]

I guess they could work but I have no experience with them. Try them and let us know

 

[tedsorvino]

Hi Thomas and everyone. I 've finished with all boards assembly and now after an initial successful simple test, there is no smoke . The
red led lighs up properly and the slam switch led works. Also, after switching on all the green leds stay on all the time (but I think this is a known issue mentioned in the manual and there is a solution)
So I 've started using signal for test. 1.5k sine wave. I 've tried with low signal (mono) about 200mV. The problem is the signal becomes even smaller after the input stage, so I can't follow it with my old oscilloscope.
Then I used a really loud signal of about 4-5V. Everything clear up to the THAT 1646 IC. After that no clear signal.
After trying with just a sound clip from my phone in the compressor's in and the output of the device connected on an old mixer, without measuring, the signal is going through the circuit, in and out levels work, but it's way too distorted (by listening from the headphones). In this case there are a lot of suspicious parties. Adaptors, mobile phone, old mixer channels (ok the headphones are working perfectly well).

So I have some questions:

1) Can a loud signal of 4v sine wave damage any op amp IC on the compressor?

2) What is the recommended test signal voltage for the device? 100mV or 1V I usually use?

3) Is it ok to test it following 2 signal traces- first cold, then hot- because I only have a mono signal generator? Any alternative idea for managing both hot and cold at the same time?

4) Could you please provide good working voltages around the ICs pins in order to make sure that the ICs work well? Especially with Ali Express sourcing it would be really helpful. Of course if it's only +- 17V and 0V no need for that.


Thanks in advance

 

[fripholm]

1) Can a loud signal of 4v sine wave damage any op amp IC on the compressor?

No. The circuit is powered by +/-17V rails. It should also take about those voltages on the input without damage.

2) What is the recommended test signal voltage for the device? 100mV or 1V I usually use?

There are no specific voltages. Use whatever you can comfortably measure with the devices you have. Use the high impedance probe (eg. 10x or more) on your oscilloscope. Remember that in a FET compressor the signal needs to be attenuated (a lot) before the FET and amplified (a lot) after that. It's normal that around the FET the signal is in the range of around 100mV. Otherwise there would be a lot of distortion.

3) Is it ok to test it following 2 signal traces- first cold, then hot- because I only have a mono signal generator? Any alternative idea for managing both hot and cold at the same time?

Connect the unit to balanced ins and outs on a mixing board or audio interface. HOT/COLD is used by gear with balanced I/O and has nothing to do with mono/stereo. Make sure your connections are reliable. There's nothing more frustrating than trying to measure something with the alligator clips falling off all the time...

4) Could you please provide good working voltages around the ICs pins in order to make sure that the ICs work well? Especially with Ali Express sourcing it would be really helpful. Of course if it's only +- 17V and 0V no need for that.

Look up the datasheet for the 1646 and check its power pins for +/-17V. All the other pins should be around zero anyway. If you see the proper voltages on pins 5 and 6 and there's signal on the 1646's input (pin 4), it should produce an output signal. If not, Ali Express probably sent you a fake one.

The THAT 1646 won't drive headphones. Connect it to a line level device, preferably one with balanced inputs - but it also works unbalanced.

 

[tedsorvino]

Thank you Thomas.
Even if I 've already taken most of the correct steps mentioned on your reply, I will procceed to all of them (along with the calibration manual guideliness) more thoroughly.
Just one more question:

should my signal be clear before starting the calibration procedure or will the calibration of the trimpot reduce any distortion?

 

[fripholm]

If you physically remove the compression FET (Q3) from its socket, the unit should pass a clean signal and there should be no audible distortion. With the FET in the circuit, the distortion depends on the voltage on its gate. This (negative) voltage is what you will set when calibrating QBIAS. But first make sure the unit works without the FET inserted.

 

[tedsorvino]

All good for the first part of the job. Clean and good signal. The 2SKA30 fet (actually similar, older and for some people better "sounding' than 2n5457) back into place and then my calibration problems started.

FIrstly all of my trimmers seem like they have infinite turns (or at least many many turns) and some (gr scale and q bias) are horizontial - because it was impossible to source verticals, so it's harder to set them especially on the bottom pcb ,the one i'm currently setting up.

I faced a serious software problem with loop latency and I couldn't start working on the q bias calibration (my daw and my card want a specific way to collaborate in a loop kind of way), I sorted it out and I managed to set the q bias correctly.

Then I moved to the meter adjustment.

The good part is that by turning the gr null trim clockwise I managed to sort out the all led on problem.
Now when the switch is in the off position there are no leds and when the ratio is up to 20:1 the leds are blinking. And the gr null adjustment was easy. Then the first take on the grscale adj was easy up to the point of the last led....I mean it takes a lot of turns from the one before the last to the last led (in that awkward horizontial position) and still haven't seen the last led barely lit (the led is working well).
I somehow decided to move to the to repeat the adjustment of the two GR pots. Gr null smooth and easy but gr scale now became even more difficult, I've stuck to the second led.
Should I start turning the grscale trimmer the other way round like going back to where I started, or because the threshold of the grnull has become too high I 'm facing this problem. Should I just find a working balance betwen the 2 trimmers and move to the distortion trimmer calibration?

Just to mention that I'm using a 5M release pot with all the appropriate mods.

 

[fripholm]

I haven't found anything about a JFET with that specific part number you mentioned. Only a MOSFET with a similar designation that definitely won't work!

The circuit around the QBIAS adjustment won't allow for the full range of negative voltages between 0 (GND) and negative 17 volts. It's actually much less! If your JFET already has a Vgs of -9 volts (voltage measurement on page 2 of the guide) you won't be able to set this with the trimmer. The voltage at the FET's gate should end up roughly at this Vgs voltage reading. The actual QBIAS voltage at the trimmer is less than that (more negative). Use a JFET part with a Vgs of about -2 to -4 volts.

Don't adjust any of this with measurement equipment connected to the circuit around the FET as their impedances skew your results. Ignore the meter calibration until QBIAS is set correctly.

 

[tedsorvino]

Thanks Thomas.
Please re-read my edited post.

Apparently you were replying when I was editing. . Now much better since q bias is set right.
The problem was particular daw software "loop" issue with particular interface - all "pro' grage, but they have some tricks, i wouldn't like to be more specific and technical since it's not the subject of the post.
As far as 2sk30A is concerned here is some more info (excellent fet for guitar pedals builds)
https://datasheetspdf.com/pdf/643313/Xiaosheng/2SK30A/1
Please reply to my edited post if you can. Thanks again for the support.

 

[fripholm]

Should I start turning the grscale trimmer the other way round like going back to where I started, or because the threshold of the grnull has become too high I 'm facing this problem. Should I just find a working balance betwen the 2 trimmers [...]

Yes, going back might be a good idea. Finding the right spot for both trimmers is not an exact science. You'll find a good one, eventually. As mentioned in the guide, the meter is not meant for precision readouts - just for a general tendency.

 

[tedsorvino]

Yes, going back might be a good idea. Finding the right spot for both trimmers is not an exact science. You'll find a good one, eventually. As mentioned in the guide, the meter is not meant for precision readouts - just for a general tendency.

Thanks Thomas. Managed to find the right spot on the pot. As I said these trimmers never stop. There is a slight click (like they are broken- but they are not) and they start again. Actually the useful "points" for each job, are within few turns.
Tried the dist. set. It seems that the 2nd harmonic doesn't really have big difference difference between highest and lowest - around quarter and half db.
The strange thing is that when I moved the trimmer far away I had some kind of "crash" and reset inside the 1176 (like shutting it down and re switching it on just on the compressor pcb). When I was moving within limits everything went down.
I will try re calibrating it a bit later. It seems like it's a really hot unit so i guess it needs really good calibration and very attenuated signal on the input, or something makes my unit distorted.
I'll get back soon with update.

 

[fripholm]

The clicks are normal for most trimmers. They indicate that you are at one of the ends of the range and there's usually around 20 or 25 turns between them. Just stop turning in that direction if you hear the clicks.

Have you made the modification from the errata section in the guide regarding the dist trimmer? This simplifies the process but there are FETs that simply do not allow to fully cancel 2nd order harmonics. It also depends on the amount of compression (or no compression at all) while doing the adjustment. For example, if you calibrate DIST without any gain reduction happening, you'll still have a certain amount of 2nd with 10dB of compression and vice versa...

It seems like it's a really hot unit ...

Hot?! Are you talking about temperature? Haven't experienced or heard anything about that before...

 

[tedsorvino]

Thanks Thomas.
By HOT I mean on the very sensitive input-gain side. I decided to make some modifications with my horizontial trimmers just to make my life easier - getting them much higher. If it doesn't work I will try to source some vertical ones.
The good thing is that by now I have a pretty clear idea what needs to be corrected and how calibrations works.
Even if I prefer to work with cheap hardware test equipment instead of software and interfaces - much safer and easier - I think eventually I will fix it. I will keep you updated as soon as i reach a final stage.

 

[swinglab]

Hello,

I am very interested in your project, I have a few questions.

Do you have a bom mouser? or an approximate total cost with the components? (I'm from France so probably the same cost as you)

And I have not all understood in the manual when you talk about the stereo link that can be installed (the switch) what this one could really link as parameters or not?

Great job anyway! I would like to make a mono or stereo to take it live.

 

[fripholm]

As far as I remember, there should be a Mouser cart done by a fellow member somewhere in this thread. Just do a search.

I don't buy parts just for a specific project, so I have no idea about the total cost. Maybe someone who actually does this and built this project can help you out. The cart mentioned above should give you a hint as well...

The stereo link just links the control signal but you have to manually set timings and input/output levels on both channels. Just as you would on two real linked 1176 compressors. It's there 'just in case' - I haven't actually used it.

 

[rmaier]

Hi Everyone,

I'm in the process of building a pair of these compressors to put in a 1U rack box. My boards are pretty much stuffed and I'd like to order a front panel. The fpd file for the 1U version silkscreen is in black, but I'm thinking of using a black front panel. Unfortunately, I can't seem to figure out a way to change the color of print in the original file from black to white. I'm sure there's an easy way to do it, but I can't figure it out. Could some kind soul please point me in the right direction?

Ralph

 

[fripholm]

I don't have Front Panel Designer in front of me at the moment but IIRC you change the color by changing the material of the panel somewhere in its properties. If it's a bitmap that gets printed on the panel, you just replace that with one that has the colors you like.