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

[paolomunarini]

Anyone having 4 main boards? 2 PSU boards would come in handy too.

I would like to make it electronically balanced with INA2134/DRV134 or THAT1240/1646... Due to the fact that the input trafo should be 2:1, do I need to setup the input stage with 6db atten (aka INA2137 / THAT1246)?

 

[lauritz]

Hi All zener builders

Finally finished my build today. Very nice sounding machine indeed! Here's a few notes that may be helpful to other builders.

Had to go through 4 different kinds of zeners before getting some quiet ones. Ended up using type 1N5231 since they were the quietest. I spend quite some time tuning the unit and i feel it's very much a delicate proces. Spend a long time wondering why no sound was coming through the unit at all, until I lowered FT2B a lot (set at 1k initially and ended up down at around 200 Ohms), then it worked and further tweeking could go on from there.

Also, I spend even longer trying to get the meters to work. I found that they only react in a very small interval of resistances governed by R88. I added a 1n60 germanium diode bridge to the vintage Sifam DC VU meters i used. I might test some more to see how they react to specific release settings and adjust further with a capacitor across the meter if needed. But all seems alright so far and maybe it does'nt matter too much. Anyway, i see the meters as a very rough indicator (especially with the scale my meters have!) and plan on using my ears for the most part. But now they move and look good

As for PSU i use the Chandler PSU which i own already and use with a Chandler preamp. So less concerns with EMI and noise getting to the audio path with an external PSU. Haufe transformers at the output and Carnhill at the inputs. Sounds very sweet. Ended up matching basically every component on the boards, so the 2 sides behave very alike (maybe that was overkill). Also i skipped the bypass switching so the unit is on all the time if the PSU is on. Just figured that it would be simpler and i have other possibilities in my studio if i just want the transformer warmth and nothing else.

Thanks a lot for a great build and a lot of good hints in this tread. I've added a couple of photos too.

all the best, Lauritz

 

[Shroomystic]

Hey !

I desperately want to build the TG1 in the future but as i dont have sufficient privilige to post in the black market section i wanted to ask here for pcbs, does by any chance somebody still have 2 Pcbs laying around and wants to sell them ?
That would be super nice !

All the best !

 

[SmokeyJoe]

Hello friends. I've just fired up my TG1 for the first time and ran some very quick tests.

Firstly, I have the hiss of doom, which means new zeners are likely required - there are so many suggestions on this thread as to what does and doesn't work that I'm not really sure which ones I should try.

These are the ones @fripholm says he used, so those will be my next to try https://www.mouser.co.uk/ProductDetail/Central-Semiconductor/1N5231B-TR-PBFREE?qs=u16ybLDytRY80fKcMFwc7w==

Can anyone recommend any other recent known good ones? I'm in the UK if that makes any difference. These are the ones I used which claim to be low noise https://www.mouser.co.uk/ProductDetail/512-1N5231CTR

I've also noticed that turning the release up to full causes the signal to drop out, although the meter continues to move.

I'll investigate further this evening, but in the meantime if anyone has encountered anything similar I'd be interested to hear how you resolved it.

I did have a brief go at calibrating the unit, however I couldn't see any effect from the CV adjustment - the waveform looked ok already. I'm not sure how much point there is further adjustment until I've sorted the noise problem.

 

[SmokeyJoe]

Update:
Channel 2 works correctly but channel 1 does not pass audio

Turns out I hadn't lowered FT2B enough on Channel 2 for it to trigger (it was right there in lauritz's post a few above mine ).

Both channels now appear to work mostly as expected other than a lot of hiss. With input and output gain both maxed, channel one sits at around -30dBFS on my interface meter and channel 2 sits around -22dBFS.

The channel 2 noise sounds exactly like the sample posted by octavez. The channel 1 noise is not quite as harsh, closer to white noise but still unacceptable I think. I'm going to try replacing the zeners unless anyone has any other suggestions.

I'm also still having an issue on both channels with the audio dropping out when the release control is turned up to full.

 

[Undos]

Hi, I have spare Fairchild 1N751 if you are interested. I am no longer allowed to sell items on groupdiy. (since the terms changed on the board) The diodes work fine in my DIY Racked TG. Sold some of them to another member and I think it solved the hiss problem in his TG unit (see older post in this thread).

 

[Undos]

Let me know and I can put it on ebay or so if you like.

 

[SmokeyJoe]

Let me know and I can put it on ebay or so if you like.

Much appreciated! I’ve just ordered three different types of zeners to try, so I may take you up if none of those work.

 

[SmokeyJoe]

I've replaced the zeners and the noise is now pretty good: around -58dB with input and output gain at max, and around -72dB at unity gain. Is this similar to what others have achieved?

The calibration is pretty good now. My main issues now are:

- the left channel input gain is about 3-4dB lower than the right, so the left input gain knob needs to be around three stops higher to achieve similar gain reduction.

- the right channel CV adjustment of FTB2 does not appear to entirely remove the CV from the signal, even when all the way to the right.

Could these be linked? Would installing a a trimpot at R22 help with either of these issues, or does it only affect the THD adjustment?

I'd like to understand a bit more about what the various trimpots are actually adjusting, is anyone able to explain please?

 

[SmokeyJoe]

I've made some measurements. 1k tone at -12dBFS, comp, attack CCW, rel CW, outputs set to return max level without clipping DAW input (-3 stops from max). I boosted the left channel in the DAW by ~4dB so that it tracked with the right.

As we can see from the graphs, they track very evenly up until the threshold point, when the left channel starts to compress more heavily in both linked and unlinked modes. The channels are out by ~1dB at max input setting.

Can anyone shed any light on what I need to adjust in order to to:

a) adjust the ratios to match
b) increase the left channel input gain

Thanks!

UPDATE: the channels now track perfectly when unlinked. With the left channel increased by 2.8dB in the DAW, the left output turned down by two stops from max, and the right output turned down by three stops from max, it’s spot on.

However, when linked the response is still the same as the graph whichever way I adjust the inputs and outputs. It seems like the summed side chain signal is being applied too heavily to the left channel.

 

[thelivingroom]

Finally finished mine. Worked perfect right out of the gate, thank goodness. No hum, no buzz, no hiss, no oscillation, no switching pops, and no noisy Zeners. I did plenty of homework in this thread before starting the assembly and wiring, and I thank all of you for the tips and/or problems solved.

I bought 200 NOS Fairchild 1N751A’s when I purchased these boards around 5 years ago, and spent hours until I had all 8 that I needed matched within .001v at 1k and 10k. I took most of today to calibrate it, listening to a mono drum OH track multed to both channels, program material (also mono and multed to both channels), or a 1k tone, resetting the trimmers for R22 and R56 several times and starting over and trying different approaches before I felt warm and fuzzy. I ended up being happiest leaving R22 at 12k and only adjusting R56 until the compression acted symmetrical on both channels across many different settings and the THD still sounded nice when I cranked the input pot up. Ch1 R56 ended up around 560R and Ch2 around 600R. Probably the most annoying thing I encountered was the expensive meters having two very different color temperature LED’s and different resistors on them. Luckily I had some 3W resistors and clear LED’s to be able to make the lighting match. Auto release, link, and bypass are all switched via relays.

Thank you for the boards and this project @fripholm; this thing sounds incredible!

 

[TwentyTrees]

What a lovely clean build! It absolutely looks the business. Fantastic job.

 

[thelivingroom]

What a lovely clean build! It absolutely looks the business. Fantastic job.

Thank you! I may pop the hood one more time to try log pots on the input, as it’s very sensitive with a hotter track/program material, but other than that I’m very happy with it.

 

[warpie]

Thank you! I may pop the hood one more time to try log pots on the input, as it’s very sensitive with a hotter track/program material, but other than that I’m very happy with it.

Very nice built indeed!

Correct me if I'm wrong but from what I can see, you connect the input XLR Pin1 (shield?) to pin 6 and pin 11 on the 9046M?

 

[thelivingroom]

Very nice built indeed!

Correct me if I'm wrong but from what I can see, you connect the input XLR Pin1 (shield?) to pin 6 and pin 11 on the 9046M?

Thank you! I have Pin 1 XLR going only to Pin 11, which goes to star ground. Pin 6 is going to signal ground that I routed from the output of each pcb to one of the screw connecters on the relay perfboard. I kept Channels 1 and 2 signal grounds separated so no loops.

 

[warpie]

Thank you! I have Pin 1 XLR going only to Pin 11, which goes to star ground. Pin 6 is going to signal ground that I routed from the output of each pcb to one of the screw connecters on the relay perfboard. I kept Channels 1 and 2 signal grounds separated so no loops.

Thanks for clarifying. So, the inner shield goes to the signal ground and the mu-metal can to the chassis. Makes sense.

Do you mind me also asking you where did you find these nice push-button switches? Are they NKK kb series? They seem to be obsolete.

 

[SmokeyJoe]

Hey folks, I've spent hours experimenting with adjusting these circuits but I'm still struggling with the calibration on this. The compression curves refuse to match between channels.

Is anyone able to explain exactly what each of the trimmers does? I've not added a trimmer for FT2A as the general consensus is that it's not necessary, but I'm starting to wonder if it will help.

I've attached graphs of my attempts at matching the response curves between channels using a 1kHz sine wave (note colours are not consistent between graphs). So far I've only tried making adjustments in compression mode:

- Graph 1 shows the response when both channels are set with every trimmer at the same resistance and unlinked.

- Graph 2 shows the effects of adjusting FT2B 1 turn and 1.5 turns clockwise, and adjusting the input of R to compensate for gain changes (-1 means one step back from full).

- Graph 3 shows the effects of adjusting FT1 and R input to compensate.

@thelivingroom your build looks great. Hoping I can get there soon... Any chance you could explain your calibration process in more detail?

 

[thelivingroom]

Thanks for clarifying. So, the inner shield goes to the signal ground and the mu-metal can to the chassis. Makes sense.

Do you mind me also asking you where did you find these nice push-button switches? Are they NKK kb series? They seem to be obsolete.

They are indeed NKK KB series, and unfortunately obsolete now, which I realized recently when I looked up the datasheet on Mouser. I ordered them a few years ago while stockpiling parts for this. I also used them on my buss compressor and really liked them. Oh well.

 

[TwentyTrees]

Hey folks, I've spent hours experimenting with adjusting these circuits but I'm still struggling with the calibration on this. The compression curves refuse to match between channels.

This may be a stupid question, but have you matched the diodes between channels?

 

[SmokeyJoe]

This may be a stupid question, but have you matched the diodes between channels?

Not a stupid question! Yes the corresponding pairs on each channel are matched to 0.01V.