0
\$\begingroup\$

SoundCraft Ghost Preamp Section

Basically, I began modding my Soundcraft Ghost sound board and wanted to put in a different op-amp that would sound more "colored" in the preamp stage. From everything I've read, the OPA2134 is the most suitable substitute, being the most stable and making a significant difference in the sound.

I was told to expect oscillations because the higher slew rates will increase the bandwidth and expose weaknesses in the surrounding circuit, and supposedly how to fix the oscillations. I'm not sure I would describe what happened from just plugging it in as oscillations, but rather as instability. Lots of pops and noises while adjusting the pot which is VR1 on the schematic, and when depressing the "Line" switch, fully opening VR1 caused intense feedback and extremely loud output with absolutely no signal connected. The phantom power also went absolutely nuts compared to the non-modified channels when it was switched on. When it is at a VR1 setting that it working properly and stably, there is less noise than with the TL072's, the same volume signal, and it sounds overall more musical to my ears (and yes I blind A-B tested this with several people).

First I tried connecting the voltage rails to ground via ceramic .1 uF caps, while keeping the leads as short as humanly possible and literally nothing happened. I was told for certain this would work. Then I was told I need to bypass the biasing voltage divider with .1 uF caps. That was equally effective. Another person told me that I needed to change the bias on the non-inverted input by adjusting the voltage divider so I hooked up a pot, and the most I could do was eliminate the gain pot scratchiness, while the phantom power and line input problem remained. I tried a .01 uF cap on the voltage rails of the op-amp and interestingly the line-input problem disappeared but gain pot scratchiness remained. Combining the changed voltage divider with the .01 uF cap made them both not work! The phantom power problem never went away no matter what I tried.

So I went to looking at why these things were supposedly working for others (it did cross my mind that maybe they were choosing to live with the instability and working around it) but not me. Something I noticed was that the 33 pf feedback caps were not on my channel revision, and neither was C6! The rest of the circuit is basically identical to the schematic.

My question is, why did people make these suggestions to me to begin with? What are they theoretically supposed to do? What is good practice in preventing instability and oscillations and how can I apply it here? Is it really a good idea to simply attempt to limit the top end of the bandwidth or is that just avoiding the root of the problem?

Could my board's revision lacking the 33 pf feedback loop caps (C67, C68) and the C6 electrolytic be the reason why the proposed solutions to my problem aren't working?

Lastly, how am I supposed to interpret a feedback loop that heavily involves a transistor and gain pot linking two separate op-amps? It seems more complicated than most literature I've found on the subject of op-amps.

\$\endgroup\$
7
  • 3
    \$\begingroup\$ ffffffff (sound of air being drawn between the teeth). 'Coloured', 'significant difference in sound', 'expose weaknesses', it sounds like you're suffering from golden-eared disease, or have been listening to people who do! It's a bit quieter and a bit faster than a TL072, but from the data sheets it looks like you are losing phase margin at a lower frequency, hence being more twitchy to get stable. \$\endgroup\$
    – Neil_UK
    Jan 15, 2017 at 20:02
  • \$\begingroup\$ "expose weaknesses" did not refer to anything that was audible... it referred to people telling me that this chip would pass higher frequencies than the TL072, which the circuit was not designed around handling stably... so once you put in a chip that has a high slew rate the ultrasonics oscillate \$\endgroup\$
    – Dave
    Jan 15, 2017 at 20:09
  • \$\begingroup\$ Regardless, I'm not here to debate you onn golden-eared-ness... Noise was lower when it WAS working at the same volume and an untrained individual (my girlfriend lol) correctly identified a distinction between A and B ten times out of ten. Don't think it's a stretch to say there's a difference between them. Preference is subjective beyond that. \$\endgroup\$
    – Dave
    Jan 15, 2017 at 20:14
  • 1
    \$\begingroup\$ Have you mistakenly come here thinking it was some audiofool (sorry audiphile) forum where subjective words and personal sensations are used to convey the pros and cons of this design versus that design. Bottom line - get hold of a simulator and look at the data sheets. They are the only (and I might say) objective answers you are likely to get here. How can you hear noise when working at the same volume? If you mean zero volume then that might be because the op-amps oscillating and blanking its own noise considerably. \$\endgroup\$
    – Andy aka
    Jan 15, 2017 at 20:28
  • \$\begingroup\$ Sure, I concede that I haven't been using the hard testing methods that pretty much anyone here prefers, nor am I an audiophile. I can hear "noise" by setting VR1 to the same level (this is testable) on an unmodified channel vs a modified channel putting my head directly next to a reference monitor and having my friend speak directly into the same microphone while switching channels. While this is not a hard test, it reveals that there is no difference in signal gain at the same pot setting (i've tested various), while noise is audibly lower to a point of obviousness. It's really not hard. \$\endgroup\$
    – Dave
    Jan 15, 2017 at 20:38

2 Answers 2

3
\$\begingroup\$

I know I'm responding to an old thread, but there were a lot of unanswered points that I can address, for any future readers.

The circuit is an instrumentation amp as seen in the Art of Electronics (figure 5.88 C). The NPN transistors boost the signal to reduce the op amp noise, as well converting the voltage input to a current on the inverting input, which helps to reduce distortion at low gain / higher voltages.

The reverse diodes are pretty standard for mic inputs to shunt off any voltages beyond the supply rails. R13/R12 are a voltage divider for the bias voltage to the non-inverting inputs, and C6 is to provide a low impedance source at signal frequencies. The actual current needed is nearly zero for a JFET op amp like the TL 072, which is probably why it was not included.

For measuring or comparing noise on a mic input, you do not want to leave the input open. Use a resistor between 47 and 200 ohms connected to pins 2 and 3, to give you a fixed input noise level to simulate a microphone.

Finally, yes C67 & C68 may be needed when switching to an op amp with higher bandwidth and less or no internal compensation. The original TL072 may be stable enough to not need them, but it's not going to hurt to add them.

\$\endgroup\$
3
  • \$\begingroup\$ Hi... Thanks so much for the answer... definitely useful to have in these old threads. I figured out much of what you were saying after the fact (and updated my philosophy on "color" lol). Ended up using the TL0xx amp and simply replacing the cheap cost-cutting coupling caps for the biggest difference after many distortion tests and subjective hearing tests showing almost no difference in the properly implemented op amp swap. \$\endgroup\$
    – Dave
    Dec 25, 2020 at 19:24
  • \$\begingroup\$ I recently read a couple of great articles by a guy who did proper distortion measurements of capacitors, and yes they can be a major source of distortion. As for color / distortion in op-amps, the less distortion the more they sound alike - just the input signal at a different level or impedance. It's the non-linear distortion that creates 'color'. \$\endgroup\$ Dec 27, 2020 at 16:37
  • \$\begingroup\$ linearaudio.nl/cyril-batemans-capacitor-sound-articles \$\endgroup\$ Dec 27, 2020 at 16:49
3
\$\begingroup\$

wanted to put in a different op-amp that would sound more "colored" in the preamp stage.

I think that replacing the TL072 with a "better" opamp like the OPA2134 would actually make the sound less coloured. If there was any "colouring".

You seem to ignore the fact that this circuit is designed around the TL072.

Assuming a circuit is well designed in an electronic sense, the added distortion should be extremely low.

The idea behind using an opamp is that the components around the opamp determine the circuit's behavior and not the opamp. If the opamp does determine the behavior, you're asking to much of that opamp and should be using a "better" one.

Now this circuit looks a bit weird to me. I do not like the many 1N4148 diodes as they're in reverse and could add noise. No idea how bad this is as it also depends on the nominal signal levels. The circuit around the two NPNs is also a bit weird with a DC feedback path through the NPN.

The bandwidth of the circuit is a combination of the TL072's own BW and the 2 feedback caps around it (33pF and 47 pF). If you only replace the TL072 with anything else you'd have to re-design the frequency compensation.

Lots of pops and noises while adjusting the pot which is VR1

That has nothing to do with any opamp. My guess is there's dust in the pot causing bad contact and thus crackling and/or a DC voltage across it while there should not be.

\$\endgroup\$
3
  • \$\begingroup\$ Thank you! Lots of great clues and information here. Good to know the feedback caps are pretty important. As for the pot scratchiness, if I put the TL072 back in it ceases completely. I've switched them out (there is a socket) to check this. This tells me that if there is an unwanted DC voltage going across it or something of the like it's the result of replacing the op amp. I'm not really sure how that's possible. \$\endgroup\$
    – Dave
    Jan 15, 2017 at 21:34
  • \$\begingroup\$ One more thing: In response to the coloured comment... yea that was just what I was told about this op-amp in comparison to the TL072 and I'm taking it with a grain of salt. If I don't like the end result once I know for certain that everything is working properly (will be a long road), I'll turn around and start from scratch. \$\endgroup\$
    – Dave
    Jan 15, 2017 at 22:10
  • \$\begingroup\$ Just wanted to make you aware, adding .01 uf caps from the voltage rails of the OPA2134 to ground and adding the 33 pf feedback caps indeed got rid of the pops and noises on four seperate channels that prior occurred at the exact same position in the gain pot. The pops also disappeared on one of the channels when I restored with the original TL072. I don't think that's coincidence. The only problem I have left now is with the phantom power which causes much louder pops of its own in quick succession for a short time upon activation. \$\endgroup\$
    – Dave
    Jan 18, 2017 at 8:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.