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

[fripholm]

It's mentioned in the post you quoted:

... the lams of a 300W power transformer from an old halogen string light ...

But it's really irrelevant what it was used for prior to tearing it apart - the lams of any huge toroidal transformer should work. They just happen to have to correct width and material.

 

[octavez]

I calibrated my stereo unit in THD mode as per the manual. Both my levels and distortion are identical in THD mode but then when switching to compression/limiter one channel is about 4db quieter. If I toggle on the LINK switch then it's the same level on both channel. Is that normal behavior?

 

[fripholm]

Only thing I'd imagine you could check is wether both sidechains are getting the same signal levels. If that's true then there's not much you can do about it.

It's almost impossible to achieve absolute symmetry between two channels. There's simply too much variables - even with carefully matched Zeners. If you want to have the same amount of compression on both channels (without link) you'd have to compromise the levels in THD mode and vice versa.

My stereo unit isn't symmetrical either and when using it un-linked (which I prefer to do) the left channel's output knob is always set a bit louder. Doesn't bother me...

 

[octavez]

Got it, that makes sense then. At least we have the link option! Thanks again!

 

[kazper]

Warning! Recent orders (Aug 2021) PSU board have major fault.

I’ve just noticed and sent a email to fripholm who hasn’t had a chance to respond yet. But the entire bottom pcb layout is one connected piece of copper.
Soldered In two resistors and noticed the error.

 

[fripholm]

Warning! Recent orders (Aug 2021) PSU board have major fault.

I’ve just noticed and sent a email to fripholm who hasn’t had a chance to respond yet. But the entire bottom pcb layout is one connected piece of copper.
Soldered In two resistors and noticed the error.
View attachment 84403

It seems to be a production error in a new batch of the PSU boards which I didn't realize until kazper made me aware of it. In the meantime I have responded to him and we sorted it out. He will get a new one without the error for free. It only affects two orders. I've already contacted the other one and he will also get a new, good one. A new batch of power supply boards is on the way as well.

Sorry for the inconvenience.

 

[kazper]

Glad is was a low impact issue as far as out in the wild is concerned. Thanks again for the quick email and the incoming replacement board.

 

[octavez]

Only thing I'd imagine you could check is wether both sidechains are getting the same signal levels. If that's true then there's not much you can do about it.

It's almost impossible to achieve absolute symmetry between two channels. There's simply too much variables - even with carefully matched Zeners. If you want to have the same amount of compression on both channels (without link) you'd have to compromise the levels in THD mode and vice versa.

My stereo unit isn't symmetrical either and when using it un-linked (which I prefer to do) the left channel's output knob is always set a bit louder. Doesn't bother me...

After using my unit for a while I noticed that after being on for at least 5 minutes, and feeding a 1khz sine then the channel that was 4db quieter, raises to match the second channel! But then it randomly fluctuates from 0 to -4db attenuation, especially right when I toggle the THD/Comp/Limit switch, it's fluctuating even more then slowly stabilize. The other channel is completely steady. Input levels are constant, no issue in THD mode either, so it looks like something is going on in the sidechain part. Any tips on where to start investigating? Maybe a bad THD/Comp/Limit switch?

 

[fripholm]

It's not necessarily the sidechain.

An oscillating output stage could also lead to level instability. If that's the case, you could try a small resistor value (47 ~ 220 ohms) in series with the output or a higher value for C22 (~10n max) to tame this.

 

[octavez]

It's not necessarily the sidechain.

An oscillating output stage could also lead to level instability. If that's the case, you could try a small resistor value (47 ~ 220 ohms) in series with the output or a higher value for C22 (~10n max) to tame this.

It was a bad THD/Comp/Limit switch in my case... Thanks again for all the support, I hope my comments and your replies will help others.

 

[Winston OBoogie]

Hi there. First time I've looked at this project.

Is there a typo or mixup in the schematic of the output stage? I see that, as drawn, the output transistor is biased at around a 3rd less current than the original, and also has a heavier load via the -tve feedback.

What was the reasoning behind this change?

 

[fripholm]

It was a bad THD/Comp/Limit switch in my case... Thanks again for all the support, I hope my comments and your replies will help others.

I'm glad you got this sorted.

What was the reasoning behind this change?

Thanks for asking. I may have mentioned this elsewhere in this thread but I started this project for myself as a 1:1 remake of the original circuit - long before I even thought about selling the boards.

When testing I noticed that the output transistors were getting extremely hot. I don't remember If I thought it was an error in the schematic or so but I've tried to find a way for them to run cooler and what you're looking at is the result. In today's studio environments they don't have to drive heavy loads, do they? Frequency response is good, distortion is negligible, drive capability is adequate, so why waste an excessive amount of current in the output transistors?

As mentioned multiple times in this thread and in the guide, I have made numerous small changes to the original schematic - not only limited to the output stage.

 

[Winston OBoogie]

The issue is that, compared to the standard today, this unit was designed to operate to a lower operating level (.447v ref. 200 ohms.)
In order to approach anything close to current headroom standards, we therefore need some free gain with an output transformer, and a 1:2 would give almost 6dB.
Any secondary load on that transformer will then be reflected down to about a quarter that value as loading on the output transistor.

Secondary load of 600 might not be the norm, but if it were to happen, that's a 150 ohm load on the transistor.

EMI transformers were slightly less ratio than 1:2 but, same principle.

To stay close to original, the output transformer would be a Mu-Metal large core and the transistors would have heatsinks for the higher bias.

Edit: when using line output transformers, a lot of times they work better with a load on the secondary for reducing resonant peaks too so that's a consideration as far as reflected load.

 

[Winston OBoogie]

P.S. EMI TG units that were sent abroad and that would need to operate to .775V ref. 600 ohms standard were supplied with extra booster amps to increase headroom.

P.P.S. I'm not knocking your decision, if it's working for you guys, then OK. I was just curious

 

[fripholm]

Nice to know, thanks. I'm using 1:1 output transformers and this is what I recommend in the guide as well. As I said, I never claimed this was 100% true to the original.

According to LTSpice, my version of the output stage is more than capable of driving 600 ohms to >20dBu without clipping. Haven't tested this on a real unit as I don't have anything in my studio that comes even close to 600 ohms standard...

 

[Winston OBoogie]

Yep, as I said, if it's working then, no worries.
Headroom relative to today's standard was intended to be somewhat higher than 20dBu and, for the most part, the output amp was sonically invisible and distortion free.
But then we are talking about a compressor/limiter so, one can expect high peaks are contained.

Cheers

 

[fripholm]

Great news: the new PSU boards for this (and other) projects are here and available for order. They now have actual traces again

According to the tracking info, both shipments of leftover boards without errors have already arrived at their destinations as well. Thanks for your patience, guys.

 

[warpie]

Just wondering if anyone has used the Carnhill VTB 9046 in the input and if so what are the correct values for the load resistor (in the secondaries) and the zobel network (if any)? thanks

 

[fripholm]

Just wondering if anyone has used the Carnhill VTB 9046 in the input and if so what are the correct values for the load resistor (in the secondaries) and the zobel network (if any)? thanks

If you're using this transformer in 2:1 mode (-6 dB) you don't have to change the values from the schematic. 2:1 will be in either series-to-series or parallel-to-parallel configurations. I don't have experience with this transformer in the TG but I'd think it could be a nice fit.

 

[warpie]

If you're using this transformer in 2:1 mode (-6 dB) you don't have to change the values from the schematic. 2:1 will be in either series-to-series or parallel-to-parallel configurations. I don't have experience with this transformer in the TG but I'd think it could be a nice fit.

Hi Fripholm, I'm not sure what values from the schematic you're referring to. My question is about the resistor that has to be connected on the secondaries of the input TX in order to set the input impedance of the unit. I'm probably missing something but I can't find it anywhere in the schematic...

I'm thinking of wiring the 9046 as series:series (10k:2k4). I don't think there will be any benefit in using the parallelarallel (2k4:600) configuration.