[BUILD] New FET/RACK Official Help Thread - Please read first post!

[alex_dempsey]

By zero, I assume you mean pinned left? -20?

That's been discussed in this thread in the last few pages. The rev A input is hot, twice as hot as the D or F. In addition, our TPad is hot so that makes it more obvious. The UREI TPad had a real weird hard to replicate taper. You can turn down the amount of signal you send to the unit, or you can do the 4 resistor mod we posted about a few posts ago. That lower the amount of gain hitting the sidechain.

If I understand you correctly, turning down your clip gain, or doing the mod, would make the unit behave like you expect.

Mike

Yes by left I mean -20.

In the mod you have posted, people said that they had to turn up the input to almost 6 to get a good amount of compression. I have the opposite issue. I have to turn my input way down to 48 just so the compressor isn't compressing. That doesn't sound like the same issue to me.

Alex

 

[Echo North]

Yes by left I mean -20.

In the mod you have posted, people said that they had to turn up the input to almost 6 to get a good amount of compression. I have the opposite issue. I have to turn my input way down to 48 just so the compressor isn't compressing. That doesn't sound like the same issue to me.

Alex

Ya it is. I'm not sure what they were referring to. Look back a few pages another guy did the mod with great success, as we have.

I know exactly what you are talking about. That's why I worked on the mod. So with high levels coming you get 5-7dB of compression at say 11 o'clock (modded) rather than 3 o'clock (un-modded).

Mike

 

[Echo North]

Like so:

https://groupdiy.com/index.php?topic=63600.msg873030#msg873030

Scroll back 1-3 pages to see the early discussion.

 

[cdeboda]

Were Q12-Q13 matched well?

I still have to fully test the mod, but it my brief test I didn't experience anything like that and those resistors shouldn't have an effect on that.

I'll try to get some testing in later this week.

Mike

Q12 and Q13 are are pretty well matched as far as I know. I’m going to poke around and see what I can find when I get some time. Going to try and solve it without pulling the trimmer.

 

[Echo North]

Q12 and Q13 are are pretty well matched as far as I know. I’m going to poke around and see what I can find when I get some time. Going to try and solve it without pulling the trimmer.

Ya it just doesn't make sense to me. Altering those R's just lowers the signal to the side chain, it shouldn't change the calibration, metering, or thresholds.

Are you sure you didn't accidentally change one of the other R's on the ratio board? That would effect the thresholds and FET biasing. 

Mike

 

[alex_dempsey]

Ya it is. I'm not sure what they were referring to. Look back a few pages another guy did the mod with great success, as we have.

I know exactly what you are talking about. That's why I worked on the mod. So with high levels coming you get 5-7dB of compression at say 11 o'clock (modded) rather than 3 o'clock (un-modded).

Mike

So I just hooked up the compressor to a test speaker to hear what was going on. On GR mode,  with the output nob set to unity, and the input nob set to 48, it is playing a 1k sign wave. I verified that at that point, it is outputting 1.23V. Obviously, the input nob should be close to 24 for unity instead of waaayyyy down like it is. I understand that a modification might fix it being really hot but I have another issue. Once I get past about 48, it spikes really really badly. It increases pretty linearly up until 24 (which should be unity). With the input nob around 24, it is reading about 8V. Im starting to think that maybe its a bad cap somewhere cause it shouldn't just suddenly unleash voltage like that. When you reach that breaking point with the input nob, it is also an incredible audible spike in the audio.

The modification you gave makes sense, but I'm not sure that this problem I'm having is what you are describing.

Alex 

 

[Echo North]

So I just hooked up the compressor to a test speaker to hear what was going on. On GR mode,  with the output nob set to unity, and the input nob set to 48, it is playing a 1k sign wave. I verified that at that point, it is outputting 1.23V. Obviously, the input nob should be close to 24 for unity instead of waaayyyy down like it is. I understand that a modification might fix it being really hot but I have another issue. Once I get past about 48, it spikes really really badly. It increases pretty linearly up until 24 (which should be unity). With the input nob around 24, it is reading about 8V. Im starting to think that maybe its a bad cap somewhere cause it shouldn't just suddenly unleash voltage like that. When you reach that breaking point with the input nob, it is also an incredible audible spike in the audio.

The modification you gave makes sense, but I'm not sure that this problem I'm having is what you are describing.

Alex

Pretty sure it is.  Again, it's not a pot, it's a TPad.  You can't think of it as a pot. If you want one that's smoother at 48dB I can probably stock some, but they're going to be a lot of money. These T-pads are a middle ground between performance and cost.  You can always lower your input level into the comp.

The Rev A has significantly more gain than the later revisions.

If anything, you're tpad maybe a little scratchy at that 48 level when it starts to pass signal. You could email about it and we can try a new one. However, really if you either a) send less signal or b) do the mod, you'll never have your input down that low anyway so it shouldn't matter.

Mike

 

[alex_dempsey]

Pretty sure it is.  Again, it's not a pot, it's a TPad.  You can't think of it as a pot. If you want one that's smoother at 48dB I can probably stock some, but they're going to be a lot of money. These T-pads are a middle ground between performance and cost.  You can always lower your input level into the comp.

The Rev A has significantly more gain than the later revisions.

If anything, you're tpad maybe a little scratchy at that 48 level when it starts to pass signal. You could email about it and we can try a new one. However, really if you either a) send less signal or b) do the mod, you'll never have your input down that low anyway so it shouldn't matter.

Mike

Im not worried about having the input down from the design being really hot. What is concerning is that anytime I surpass that thresh hold where it spikes, the compressor is just killing the signal completely. Is there a way to share a video with you so that it might be more clear what I'm referring to?

Alex

 

[Echo North]

Im not worried about having the input down from the design being really hot. What is concerning is that anytime I surpass that thresh hold where it spikes, the compressor is just killing the signal completely. Is there a way to share a video with you so that it might be more clear what I'm referring to?

Alex

Hi Alex,

I know exactly what you are referring to, no need to share a video.

Think of it this way, the input is say roughly a 30dB pad.  When it's turned full CW it's like it's not there, there is no pad. As you rotate CCW back to 48 the pad starts applying itself.  Once the pad reaches it's maximum ability to pad the signal, at around 48 the signal dies. The 3 deck pot has just lost it's ability to apply anymore pad and pass signal. And yes it is pretty drastic when it happens.

With the Rev D and F it's not an issue because you never really need to be down that low in normal operation. However, with the A having a much hotter input circuit, you may find (if you send hot signals and want light compression) you need to lower your source material so you don't need to be applying that maximum pad. Or you can try the mod.

This is the limitation of the 3 deck t-pad. We can get them at a reasonable price, they generally don't get scratchy, but they do cut the signal once they've applied their maximum pad.

That spike/cutoff point location and the extremeness of it does vary from tpad to tpad. Depending on tolerance and or build. If you'd like to try a different tpad, your welcome to email us. However, it's never going to not have a spike at that early rotation point. That's why I suggest the mod or lowering your input level.

Mike

 

[Echo North]

Q12 and Q13 are are pretty well matched as far as I know. I’m going to poke around and see what I can find when I get some time. Going to try and solve it without pulling the trimmer.

Did you increase R22 as well? You might want to leave that one.

I hope to test the mod fully in a week.

 

[cdeboda]

Did you increase R22 as well? You might want to leave that one.

I hope to test the mod fully in a week.

R22 is still the original value. Checked all the values on the ratio board and everything is good. I also pulled and rechecked Q1/Q11 and Q12/Q13 and both pairs are very well matched.

I decided to poke around with the scope and tried to recalibrate with the probe on test point 17. All was good until the tracking adjustment - same as before. What I found however is that I’m getting some pretty heavy clipping. I can’t get more than about 6 dB of reduction from the input knob without the top of the waveform being clipped off, and by 10dB of reduction it’s pretty bad. I can post a picture if that helps, but I don’t think it’s the tracking adjust. Could it be a biasing issue in the front end?

The T-Pad is really scratchy/jumpy along pretty much the whole rotation as well. Not 100% sure it’s the T-Pad but it seems like it is. This seems like a separate issue from the tracking adjust/distortion though.

 

[Echo North]

R22 is still the original value. Checked all the values on the ratio board and everything is good. I also pulled and rechecked Q1/Q11 and Q12/Q13 and both pairs are very well matched.

I decided to poke around with the scope and tried to recalibrate with the probe on test point 17. All was good until the tracking adjustment - same as before. What I found however is that I’m getting some pretty heavy clipping. I can’t get more than about 6 dB of reduction from the input knob without the top of the waveform being clipped off, and by 10dB of reduction it’s pretty bad. I can post a picture if that helps, but I don’t think it’s the tracking adjust. Could it be a biasing issue in the front end?

The T-Pad is really scratchy/jumpy along pretty much the whole rotation as well. Not 100% sure it’s the T-Pad but it seems like it is. This seems like a separate issue from the tracking adjust/distortion though.

The A does clip earlier than later revs. In the case of the GR, it's because there is no resistor on the Q1 source. After the Rev A they added a 10K to control that saturation.

Email us about that t-pad. We'll send you a new one.

Mike

 

[cdeboda]

The A does clip earlier than later revs. In the case of the GR, it's because there is no resistor on the Q1 source. After the Rev A they added a 10K to control that saturation.

Email us about that t-pad. We'll send you a new one.

Mike

Done, thanks.

I think I’m going to try those original values in there again and see what happens. I stuck sockets in those 4 spots for now.

 

[cdeboda]

The A does clip earlier than later revs. In the case of the GR, it's because there is no resistor on the Q1 source. After the Rev A they added a 10K to control that saturation.

Email us about that t-pad. We'll send you a new one.

Mike

So with original values the thresholds per ratio are back to normal (less compression at 20:1 than 4:1 at same settings) and the tracking adjust successfully calibrated around the center of the pot. With the scope hooked up I can see why dialing in gain was so tough with how jumpy the t-pad is but I’m not sure why my unit’s tracking adjust is acting wonky with the 4 resistor mod but is otherwise okay with the stock values.

Out of curiousity, which resistors affect the threshold and ratio?

 

[Echo North]

So with original values the thresholds per ratio are back to normal (less compression at 20:1 than 4:1 at same settings) and the tracking adjust successfully calibrated around the center of the pot. With the scope hooked up I can see why dialing in gain was so tough with how jumpy the t-pad is but I’m not sure why my unit’s tracking adjust is acting wonky with the 4 resistor mod but is otherwise okay with the stock values.

Out of curiousity, which resistors affect the threshold and ratio?

Ya that's weird. We (and a couple others) did the mod with no issues, but again, I didn't fully test it. I'll get to in in about a week.

The threshold/ratio resistors are R61-R64. They set the negative DC threshold at pad 21. Basically, the higher the AC to pad 22 and the higher the - DC at pad 21...the higher the compression ratio will be.

However, as you noted, at 20:1 the ratio is obviously higher, but so is the threshold, so you see less compression on the meter than you would 4:1 at the same settings.

Mike

 

[alex_dempsey]

Hi Alex,

I know exactly what you are referring to, no need to share a video.

Think of it this way, the input is say roughly a 30dB pad.  When it's turned full CW it's like it's not there, there is no pad. As you rotate CCW back to 48 the pad starts applying itself.  Once the pad reaches it's maximum ability to pad the signal, at around 48 the signal dies. The 3 deck pot has just lost it's ability to apply anymore pad and pass signal. And yes it is pretty drastic when it happens.

With the Rev D and F it's not an issue because you never really need to be down that low in normal operation. However, with the A having a much hotter input circuit, you may find (if you send hot signals and want light compression) you need to lower your source material so you don't need to be applying that maximum pad. Or you can try the mod.

This is the limitation of the 3 deck t-pad. We can get them at a reasonable price, they generally don't get scratchy, but they do cut the signal once they've applied their maximum pad.

That spike/cutoff point location and the extremeness of it does vary from tpad to tpad. Depending on tolerance and or build. If you'd like to try a different tpad, your welcome to email us. However, it's never going to not have a spike at that early rotation point. That's why I suggest the mod or lowering your input level.

Mike

I understand there will be a bit of a spike because of the design and I understand its a hot circuit by nature. What doesn't seem right is that I'm reading almost 8V at the output when it should be at unity gain reading 1.23V (which is being sent in) or less (from the compression obviously). Do you have any suggestions on how to test where I'm getting all this voltage from?

 

[Echo North]

I understand there will be a bit of a spike because of the design and I understand its a hot circuit by nature. What doesn't seem right is that I'm reading almost 8V at the output when it should be at unity gain reading 1.23V (which is being sent in) or less (from the compression obviously). Do you have any suggestions on how to test where I'm getting all this voltage from?

Not sure where you're getting the unity gain info from. The Rev A by design had 45dB of gain. It's basically a mic pre at that point. So at mid rotation on the in/out and no GR you could easily be getting 20-25dB of gain.

Shoot me an email about a new TPad if you like.

Mike

 

[alex_dempsey]

Not sure where you're getting the unity gain info from. The Rev A by design had 45dB of gain. It's basically a mic pre at that point. So at mid rotation on the in/out and no GR you could easily be getting 20-25dB of gain.

Shoot me an email about a new TPad if you like.

Mike

Is there any test readings I can check on the input transformer? I want to make sure that it is functioning properly and I'm getting the correct amount of voltage there?

 

[Echo North]

Is there any test readings I can check on the input transformer? I want to make sure that it is functioning properly and I'm getting the correct amount of voltage there?

Absolutely. It's covered in the trouble shooting guide #2 "Amplification Stages".

http://www.hairballaudio.com/blog/resources/post/fetrack-troubleshooting-guide

Start from the beginning to get the set-up right. TP1 is directly after the input tx. Note that everything after TP15 is pretty dependent on the output pot tolerance so it can be +/- 20%.  I keep forgetting to add that.

Shoot me an email about that TPad.

Mike

 

[cdeboda]

Ya that's weird. We (and a couple others) did the mod with no issues, but again, I didn't fully test it. I'll get to in in about a week.

The threshold/ratio resistors are R61-R64. They set the negative DC threshold at pad 21. Basically, the higher the AC to pad 22 and the higher the - DC at pad 21...the higher the compression ratio will be.

However, as you noted, at 20:1 the ratio is obviously higher, but so is the threshold, so you see less compression on the meter than you would 4:1 at the same settings.

Mike

I'm back.

Swapped out the T-Pad and it's magnificent, thanks for that.

So I think I've figured out what the deal is for the most part. With the 1176 calibrated using the original resistors for the AC ratio section the meter shows the correct amount of gain reduction with the new resistance values (x5). The onset of clipping was making it impossible to calibrate the tracking with the new values because while compressed there was a clean signal at the output, but with GR off the voltage at the output isn't a good indicator of how much GR is actually taking place because a huge chunk off the waveform is chopped off.  Does that sound right?

Throughout all this testing I realized that the change in thresholds I was experiencing was also a change in the ratios.  I plotted the compression curves after I had everything put back together and calculated these ratios at the 4 settings:

4          4 : 1
8          5.5 : 1
12      5.8 : 1
20      6.34 : 1

Honestly, the tracking thing was something I could live with if it came down to it, but I feel like this is a deal breaker for me. At this point it's like having 2 ratio settings. The top three are so close together that it's like I'm losing  2 ratios.  This is besides the fact that the top ratios aren't even close to what they are supposed to be. Did I go wrong somewhere? It's also weird to me that with the x10 resistance values I couldn't achieve any compression at all, regardless of input level and position of the T-Pad so maybe that indicates a problem in my build. I'm not sure about that though since everything checks out with the stock values in those positions, but I certainly wouldn't write off the possibility.

I'd hate to call it on the mod and stuff those stock values back in there because I love what the mod solves, but when it's time it's time I guess.