[BUILD] fripholm's TG1 Zener Limiter boards - support thread

[Potato Cakes]

Got a chance to circle back and work on this.

I pulled the diodes and matched 8 different pairs out of the new batch of 1N715As I used the two that measured the highest voltages, which I don't think actually matters. One of the channels is now clean but the other still has the same noise. However the channel that is clean still has several issues. I can't seem to get any of the fine tuning steps to actually do anything. Adjusting the trimmer does not change the waveform to what is supposed to look correct. I can adjust this properly on the noisy channel. I also cannot seem to do a proper adjustment on for fine tuning II. I do not have a distortion meter, but when I look at 1k on the scope I can see all the THD bumps on the sine wave, but adjusting R22 does not change any level of the distortion. I do not know if this is related to the fact that I cannot perform fine tuning I. The compression circuit is working, but it does have some weird things happening like when it comes to a quiet passage in the program material, the gain reduction circuit is being maxed out by several low frequency pulses about 2 seconds apart, then disappears. It does not do this with the side chain HPF. The noisy channel does not have this problem.

Another weird thing is that I get a pop on both outputs when I bypass either channel, but not when I engage the compression. I am using the square push button switches from Don Audio to control a pair of relays.

For now, I need to fix the noise that still persists on one of the boards. Then I can focus more on the other guy. But I am a bit surprised of how much trouble I am having with the quiet channel. I'll get back on it tomorrow. I hope.

Thanks!

Paul

 

[Potato Cakes]

I was about to go to bed. Then my OCD flared up....

SO... I went back to the "clean channel." After poking around some more I found that I had to make adjustments to R56 as well because when I turning the trimmer for fine tuning I, I would lose audio, and I remembered something being said about that one time at that party with R56, so I trimmed that guy to where the audio came back and then I could finish adjusting the other trimmer to figure out R7 and R8. I still don't know if I am doing Fine Tuning II correctly. The highest value trimmer I have is 20k and not 50k as instructed in the manual. I do have some 100k pots that I could use, but I don't know if I will be able to make an accurate enough adjustment to my liking. I'm more curious if will be able to see this change on the oscilloscope and I currently don't. I do know the compression and limiting and the THD all sound correct with no weird jumping in the gain reduction circuit or the distortion being so great it sounds broken.

There is actually some noise, but it is way, way quieter. And giving it another listening it is indeed a radio station and not oscillation. When I press on the Bypass 2 header it comes in loud and clear. But to hear the noise I have to really open up the volume on my studio monitors, so I doubt it will actually be heard in any track I use to process with this. I do have this problem with tube mics if the cover is left off, but I don't recall having this problem on any of the rack gear I have built unless there was an unshielded connection there should have been one. Maybe it will be even quieter when I close the top lid and put in a rack.

I have gray hill switches to replace the Lorlins that are currently being used for the output gain, which causes loud pops when I make an adjustment. However I am using Lorlins for the input gain and they are totally fine.

The relays that are controlled by the bypass push button switches still cause a loud pop when I engage/disengage the circuit. I have used the same relays before on many other builds and have never heard anything like this when they switch. The Link push button is fine, but it is not controlling a relay like the bypasses.

The fancy 2.4mA meters that I got from Don Audio are working very nicely. I reread the meter calibration and realized that when the unit is off, the meter should sit on the line below 0 and when it is on R88 should be adjusted so it moves to the 0 line. I haven't done a thorough accuracy test, but from what I hear from the compressor and what I see on the meter seem to match up pretty good.

We're very close now.


Thanks!

Paul

 

[Potato Cakes]

I need to change R86 and R87 to increase output so it is at unity gain on bypass. I found the formula for doing so, and if I decrease R1 and increase R2 I get the Vout value that I need, but I don't know if this is the correct approach or if I just change one value or if it even matters that much when dealing with low voltages.


Thanks!

Paul

 

[Potato Cakes]

I got one channel up an running pretty good. The other is still very noisy even though I changed the diodes. I sorted out getting unity gain in bypass for both channels and they both compress and limit properly, but I could not closely match the THD because of the excessive noise. I did do a full swap of control and transformer connections and the noise stayed with the board. I was hoping for something easy, but it seems that problems have been getting very good at hide and seek these days.

I'll try pulling diodes again and see what happens. I find it very unlikely that diodes from two completely different matched pairs from different manufacturers made at different times would still be the issue.

Thanks!

Paul

 

[Potato Cakes]

I changed the diodes again and touched up some solder joints. One or the other sorted out the noise, but now I am have a very difficult time matching the THD between the channels. I had put the first one the scope and was able to get the distortion per harmonic to fall off nicely and still be able to have unity gain with input and output knobs in the middle. I can get the other to kinda do the same, but at much lower volume. I've been messing around with trimmers for the last several hours and the best I can do is still +2dB of the channel I am trying to emulate, but the distortion characteristics as per the scope is still way off, and this is with R56 at 0.4ohms and R22 being over 30k.

The "good" channel wound up having R22 being 13k3 and R56 being 240.

I've managed to break a trimmer leg and socket in the process. The really sucks. The one channel is exactly how I want it but I may have to compromise it so I can have them both be the same. It's getting late again and I'm starting to get angry as I have been working on the noise problem most of day and I should be done with this by now and moved on to my last build. 

Anybody else run into difficulty with matching channels' THD?

Thanks!

Paul

 

[fripholm]

Nice progress, Paul!!
Does that mean you have fixed the noise issue for good? What was it?

I wouldn't try to be too fastidious with the THD adjustment. As long as the onset of distortion is at about the same level for both channels it is good. Could be a zillion reasons why the Xth harmonic is higher on one channel or the other - component tolerances, temperature, phase of the moon...

One my stereo unit, channel two has more 2nd harmonic in THD mode, but the two together sound great on stereo signals, so why bother. Interestingly, in COMP mode with link on, there's almost no difference between the two channels with regards to the amount of harmonics.

 

[Potato Cakes]

Thanks!

I have no idea what was causing it, actually. I replaced the diodes and touched up some solder joints and then it was fine. There is still radio interference if I touch the cable going to the Bypass 2 header. I think the wires for this connection need to be shielded. I had a similar issue on some SSL preamps I built.

Swapped out the Grayhills for the Lorlins for the output switches but still have loud pops. I did some experimenting and found that a electrolytic cap with the + side on the pole and the - side on the 1st position that it solved that issue. I still have loud pops switching the circuit in and out. I might do something similar there. I still need to redo the resistor values for both input and output switches.

I think I have a pretty good THD procedure figured out, but I won't really know till I go back and redo the CV and THD adjustment on the channel that was good. After fighting with this for the last 15 hours, I thought I had it all good to go, but when I doubled checked my work, I saw that the "good" channel had a slightly sagging waveform when doing the drum loop test and it a little too clean in the THD mode. But that will wait till tomorrow. It's 6am here and I'd like to get a couple hours of sleep before I get up and get back at it. Once I confirm the THD setup procedure works, I'll share it here.

I will say if there is another batch of boards you wind up making, you might want to see if you can have the solder pads made out of a little more durable material. Or at least out of a metal/alloy that allows removing solder a little easier. For some reason, the pads are stubborn and hold on solder, requiring longer time that heat needs to be applied to fully clean it out. It's not a huge deal for me or should it be for anyone else who has the skills necessary to construct and troubleshoot this project, but it would make swapping out parts when tweaking a little faster. I already have another board that I'm going to cram into some sort of channel strip, so I'm set for this PCB. But maybe others might appreciate it.

The 2.4mA fancy meters do track very well with this circuit when using program material. The Stereo Link function also works very well.

This project has been quite the unexpected workout, but I'm glad I did and to have another high end piece of gear to add to my studio.

Thanks!

Paul

 

[fripholm]

I may have found a workaround for the loud pops when switching the bypass!

My desktop unit (it just has a toggle switch, no relays) also did this with the output gain control at certain positions. I just connected a 100k resistor from the middle pin of BYP2 to ground. A convenient spot for this might be the topmost pad of C34 (which is the middle pin of BYP2) connected to one of the ground pads of R94/R30. You could even remove C34 completely and use its pad for the resistor. This cap was a first and desperate attempt to minimize the pops, but as you can see, most of the time it didn't really work...

This got rid of the pops when switching in and out. Now there's merely a small click sound like my stereo rack unit does. I wonder why the relays in that unit don't cause popping sounds...

I will add this to the guide if someone can confirm it's working for them, too. Sorry for the inconvenience  :-[

 

[Potato Cakes]

Thanks for info. I was going to do something similar, but this info will save me time. And don't worry, this is DIY! It's all part of the love/hate emotional roller coaster that is inherent in this process. Hopefully I'll have this wrapped up today minus the power transformer that I've been waiting to arrive.

Thanks!

Paul

 

[Potato Cakes]

I tried the 100k resistor thingy and it's still popping. What's interesting is that it will pop on the respective channel when engaging and then pop again on both channels when disengaging. For a quick test, I grabbed a random resistor and connected it to the middle position and ground and that solved the problem when engaging but still pops when disengaging. Right now I'm going to work on redoing the input and output gain resistors resistors and replace the connection to Bypass 2 with balanced line.  After that I'll mess around with the popping issue.

Thanks!

Paul

 

[Potato Cakes]

I've got the output gain resistors replaced with the new values and it's approximately 2.5-3dB steps over 11 positions. Now I need to redo the input ones as now I want them to be more consistent as well. I was initially lazy at first and just did the 10k pot as per the BOM and divided into 11 steps. But what i discovered with output pot is that it worked best with my transformers using a 20k. So I attached a variable pot and turned till I had a 2.5dB difference, recorded the resistance, and repeated. I really should find the formula for just doing the math.

I replaced the wire for the Bypass 2 switch for shielded cable, but it didn't change anything as when I touch those wires, the same AM radio station comes through clearly. The output gain wires do the same thing as well, but at a lower volume.

When I redid the fine tuning I on one of the boards, I couldn't get it to track right till I got it to 2ohms, which I feel means something else it wrong, however the image is correct. The only side affect that I noticed is that with just using tone, when the other channel is at -4dB of reduction, this one is at 0, but it matches the other channel very closely after that. I feel this may have something to do with the abnormally low values for R7 and R8. This is board that I thought I had figured out the first time then when I double checked after calculating for the parallel resistors I saw that the drum loop wave form was askew.  I guess I can stand to have to desolder it one... more.. time....

For the other board I have, thing I may just mount the trimmers permanently. For future revisions, it might be a good idea to have a third solder pad to mount trimmers that way tweaking this will be much easier to do.

After all of this then fix the switch pops and we're good to go. This definitely will not be used in instances where exact calibration and finesse are the goal. I was really hoping that this would be done today (technically yesterday) as I really didn't have that much left to do, or at least I thought. Kinda bummed it will require another day of poking around.

Thanks!

Paul

 

[Potato Cakes]

I've gone back to the initial setup and I think I have something wrong with the board that has unreasonably low values for R7 and R8. I've put R22 and R56 back to stock values and the board that seems problematic is 18dB lower than the other channel, which is the same when I first started testing. During the the fine tuning I, the gain reduction goes from nothing to maxing the meters out. This and the gain reduction is what causes me to think there is something else wrong. I thought I was about done but this might take one more day of poking around. I know it's not the diodes as I've swapped those out a couple of times and still get the same results.

I still get unity gain when channels are bypassed.

Thanks!

Paul

 

[Potato Cakes]

Found some more things that will hopefully narrow down where this problem with the huge gain difference.

Swapped boards with the control connections one more time to be certain that it is the board and not a wiring issue. Problem moves with the board. Both boards still have unity gain when bypassed.

With both in compress and same setting, if I short R56 on the bad channel temporarily, the gain is then about 1dB off, which is more of what I would be expecting with two boards built with the exact same components. When I bypass then engage the compressor the gain issue returns. In THD mode, shorting R56 does not move the output levels up to match the working channel. It seems similar to when a power regulator has a latching issue if there are too many capacitors for the PSU to support.

When running the drum loop, the problem channel goes from no compression to maxed out compression (needle pegging on the beats), and the output level goes from lower output signal to much higher output, almost a 25dB swing. The HPF Sidechain affects the dB reduction (as far as the meter is concerned) as one would expect.

With both channels set to compress, bypassing the circuit on the good channel creates pops on the output of both channels, but only the meter of the bad channel pegs with this pop. When circuit is bypassed on the the bad channel, it still pops on both channels but the meter of the good channel does not move. I pulled the header with the link connection on the bad channel and the meter does not move when doing the above procedure, but there still is popping on both channels. When engaging the circuit, there is only popping on the respective channel. What may be needed is putting DC blocking caps to the output transformer as I'm sure the popping can't be good for it.

I'm going to start testing voltages around R56 and see if anything will show up. My focus right now is the gain issue. What ever is causing the output level problem is affecting the fine tuning I procedure and how the compression circuit is behaving.

The one thing that is really confusing is the pops on both channels when only one side is disengaged. The only connection between the two other than tying to a star ground is the link control, which is bypassed and is not making that connection in the first place. Plus, this still happens then link header is disconnected from the board. But for now, this and the radio station thing will have to wait.

Comments and suggestions are very welcome indeed.

Thanks!

Paul

 

[Potato Cakes]

Correction, I do see the DC blocking cap before the output transformer. So I really don't know what's going on with the popping.

 

[fripholm]

Here are some DC voltages, I've just measured from my stereo unit, set to COMP with no input signal. If they differed from the two channels, I've included the other channel in blue. Let me know if you need more.

Regarding the relays: I'm using Fujitsu NA24WK and there are no popping sounds when they switch, just a tiny click.

Hope this helps.

 

[Potato Cakes]

This does help, thanks. I started comparing DC voltages and they seem the same for the most part. I started checking AC with signal applied and I am only able to get a reading on some of the transistors in the gain reduction section. When I try to get a reading in the input section, the meter shows VAC for a second then falls off rapidly, even though signal is still present on both channels. I can measure audio signal up to Q1 and that's where I can't seem to get a steady voltage for my DMM to read. This applies to both channels.

I think I have a bad transistor in the input section of the trouble channel that is acting like an expander that opens up after I hit it with enough signal. It's going to be mostly guessing as I get a little lost trying to follow the audio path amongst all the transistors. This will just have to be a painful trial and error session.

I'm using OMRON G5V-2-DC24 relays. I have used them for pretty much every build that as needed relays, so this popping is a first for me. Plus, I am mostly baffled this happens on both channels when either one is disengaged.

Thanks!

Paul

 

[Potato Cakes]

I didn't have any BC559 left so I used some 560Cs for the suspect input PNP transistors but to no avail. I've gone over solder joint and checked voltages, and triple checked the control connections by swapping headers but still nothing. All the voltages in the input section matches pretty good with each other. The diodes and transistors in the compression section are different because they are not getting enough signal to do anything. When I crank the input signal to were it opens up whatever is choking the signal and the circuit is compressing, those voltages then match the other channel when equal compression is shown on the meters. Once the signal is passing, I can then turn down the input signal to where I read 4dB of reduction on the meter then past that, the signal falls off and I have to repeat the above procedure. I also found is when testing voltages, on certain points it will make the meter spike and then the compression is happening until the signal falls below that certain magical level. At this point I'm just guessing. I don't know if it's a faulty diode or transistor, and I'm not aware of diodes having this particular characteristic. My options are to replace all diodes and transistors on this board or build up the other board I I've been saving for another project and throw the bad one away.

I'll start with the diodes. Then we'll try with the transistors. This is becoming very frustrating....

Thanks!

Paul

 

[Potato Cakes]

Seems I don't have the diodes or the BC559s to make the swap. I'll have to make an order then wait. So hilarious.

Thanks!

Paul

 

[fripholm]

Indeed, sounds like a zener diode slipped into the place of an ordinary diode or something like that. Sometimes they are hard to tell apart.

FWIW, what does your grounding scheme look like?

 

[Potato Cakes]

I've been following Ian's Grounding 101 for all my builds and it works really great. So no shared grounding between boards. So PSU, both boards, and any transformers that need to all tie to the star ground separately. If there is a ground plane connection and a 0V on a board those are connected separately as well.

Thanks!

Paul