Skip to main content

Weird preamp behavior

hello guys,

I have kind of a weird situation here. Let me explain.
We are working with that guy who developed a great machine that uses lasers and such to read sound out of different supports.
The technology is great and everything but there is one thing I don’t understand.
The guy came and plugged the audio outputs of his thing into a RME quadmic preamp. So I went “oh no, I have better preamps you can plug that into” and he says “actually no, this only works with that RMEpre”. Since I couldn’t understand why, I still plugged the audio output of his rack into our new grace preamp. All I got was a really distorted signal, not usable. Went to the RME. Perfectly fine. Tried with a PreSonus preamp, weird distortion again...
So he’s right. The audio outputs of his machine only work with that RME quadmic preamp. But I do not understand why! I am thinking some kind of impedance thing but still, I find it really weird.

Could it be that his audio outputs are not really proper line audio outputs and that for some reason, only the RME is, by luck, designed to be ok with his signals!
Anyway, if some tech guy has any input here, i would be glad to hear it.

Thank you and good day!

N.

Comments

niclaus Sat, 05/18/2019 - 18:44
Yeah that is what I thought.
But I forgot to mention something.
So let’s say that we are using the RME, when gain is turned all the way down, the levels are ok, not too high but ok. When you start adding some gain everything is fine for a few dbs, but then you cross a point when the pre starts indicating some clipping and starts distorting like hell, even though the output levels are not that high. And it is on and off. One minute you are good and then you add just a little gain, and boom, clip led and high distortion!
I am puzzled.
Does that also sound like what you are mentioning?

N.

PS : thanks for your answer by the way.

Kurt Foster Sat, 05/18/2019 - 20:52
niclaus, post: 461094, member: 33719 wrote: I guess you mean line output.

I am sorry but i have to disagree with that.
A lot of preamps can totally handle line level. And the ones I have do. That is not the issue here. Nobody’s clipping or anything.

i've been recording and doing audio for over 40 years. i think i know what i'm talking about.


A line-level signal is approximately one volt, or about 1,000 times greater than a mic-level signal. ... Connecting a line-level source (such as mixer output) to a mic-level input will cause the sound to be loud and distorted because the line signal is much stronger than what the mic input will accept.

Line level is in the region of 0 dBV (1.000 volt). A line-level signal is approximately one volt, or about 1,000 times greater than a mic-level signal. Connecting a microphone to a line-level input will result in almost no sound at all because the mic signal is so faint that the line input cannot hear it."


https://www.thebroadcastbridge.com/content/entry/7578/the-difference-between-line-and-mic-level-audio

Kurt Foster Sat, 05/18/2019 - 21:32
i think you may be confusing things a bit. many "mic pres" have both mic and line level inputs. if you are putting a line level signal into a mic input you will get distortion, even if you aren't "clipping" the signal in your DAW. it's the same thing as a distortion box in front of a guitar amp. you are overdriving the signal.

niclaus Sat, 05/18/2019 - 21:43
Ok then, from the grace manual for example :


AUDIO CONNECTIONS
Microphone Inputs: Female XLR, pin 2 positive, pin 3 negative and pin 1 ground. 48V phantom power is supplied on pins 2 and 3. With a gain range from -6dB to 69dB these inputs can be used for any type of microphone or as line level inputs.

niclaus Sat, 05/18/2019 - 22:58
You know what, i got carried away here, that is not even the problem.
When I take this signal that is supposed to be line and feed it to the RME (which is supposed to be mic level), it works. When I feed it to a line level input of some sort I have distortion.
If I follow what you are saying (and which I understand perfectly) that would be the other way around, right?
I mean, if I had distortion in a mic input, I would understand, the signal is too hot, impedance is wrong, I should get distortion, but I don’t!
I get distortion in a line level input!
Which is what I don’t understand.

Boswell Sun, 05/19/2019 - 00:23
My guess is that the output from his gear has a d.c. problem. This could take the form of different d.c. levels (unbalance) on the positive and negative outputs, or else an output overload or instability when connected to a low-resistance d.c. path such as a transformer input of a high-end pre-amp.

I would make the following test: use an XLR splitter cable (not an insert cable) to connect the two pieces of gear, and then take measurements with an oscilloscope or a multimeter at the open split end. What you need to read is the d.c. voltage between pins 2 and 3 (multimeter only), or else the two readings between pins 2 and 1 and between pins 3 and 1. Try this when the RME is the receiving device (it has a true differential input) and then with the Grace or another pre-amp that causes the distortion problem. Transformer-coupled inputs require true balanced input signals to avoid core saturation due to excess d.c. flowing in the windings.

This is an illustration of just one of the differences between audio and instrumentation setups.

niclaus Sun, 05/19/2019 - 00:50
Thank you so much Boswell. That makes total sense.

I will try to make those measurements as soon as I can.
So if that is a DC problem, is there a remedy that you can think of? You would probably have to see the unit itself I guess, but anyways, if you ever think of something on the top of your head, let me know.

Thank you so much.

N.

Boswell Sun, 05/19/2019 - 09:30
I would avoid transformer-input pre-amps for this job. If you need the variable gain that a pre-amp gives you, stick with the RME Quadmic. Otherwise, if the setup can cope with a fixed 0dBFS level of (say) +18dBu, then try something like the Audient ASP880, which has balanced insert returns direct to its ADCs.

I would class the use of a suitable pre-amp as a "remedy". If you are obliged for other reasons to use one of the problem pre-amps, try a large-ish capacitor in series with one pole of the input. A 100uF 10V electrolytic should be sufficient, but be sure to connect it with the correct polarity to match any unbalanced voltage. What's the frequency range of interest in the signals coming out of this rig?

BTW, a voltage gain of x1000 is 60dB.

niclaus Sun, 05/19/2019 - 09:54
Really interesting. Yeah, I need the variable gain as sometime the output level of that thing is quite low, depending on what it is reading (old cylinders for example).
But even with the quadmic, we do have some weird things happening. Like I said earlier, sometime you just want to add a little gain, but for some reason the preamp starts acting up. When you pass a certain point, a lot of noise starts appearing. And it is not the background noise beeping brought up, that is something else. You add one dB, and boom you have 10db more noise. As if it was reacting to something when you were turning the pot.

The frequency range of interest? Well, it is supposed to be audio so i’d Say up to 20000hz, but the reality of things is that the bandwidth of the material we are reading is not that wide.
I am not obliged to use another preamp, I can stick with the quadmic if that is the best option here.

Anyways, I am going to make the measurement today and let you know what I read.

Thanks a lot for your help.

N.

Oh, and by the way :

Boswell, post: 461102, member: 29034 wrote:
BTW, a voltage gain of x1000 is 60dB.

Yes, you are correct (as always), sorry, I got mixed up with power (Saturday night!). But that is still making my point.

Boswell Sun, 05/19/2019 - 14:50
niclaus, post: 461103, member: 33719 wrote: But even with the quadmic, we do have some weird things happening. Like I said earlier, sometime you just want to add a little gain, but for some reason the preamp starts acting up. When you pass a certain point, a lot of noise starts appearing. And it is not the background noise beeping brought up, that is something else. You add one dB, and boom you have 10db more noise. As if it was reacting to something when you were turning the pot.
I've met this type of effect before, and at that time it turned out to be due to h.f. oscillation being rectified and turning into audio noise. It may be nothing of the sort in this case, but it's worth putting an oscilloscope on the equipment output lines and checking for oscillation.
niclaus, post: 461106, member: 33719 wrote: So when plugged into the rme the d.c. measurement between pins 2 and 3 shows 0.15v.
With the grace 0.40v. What do you make of that? there is less D.C. current with the RME but there is still something? Which would justify why it behaves better but still erratically?
Yes, interesting, but both those values are uncomfortably high for a transformer-input pre-amp. The d.c. input resistance of the Grace would be only a few tens of Ohms. Did you happen to note the voltage at the equipment output when it was open-circuit?

niclaus Mon, 05/20/2019 - 07:23
Hey Boswell,

So yesterday I copied pasted your first message to this inventor guy. Here is his reply :

“Well - if only he knew how right he is :-)
only, its not a problem, it is the very concept.

By nature the output signal is asymmetric. It is a wonder that it works so well, but it does.

Greetings to your friend.”

So I guess I just have to accept that for some reason this crazy machine works as it does with the RME...

But thank you so much for your help Boswell. You were spot on!

Boswell Mon, 05/20/2019 - 07:49
niclaus, post: 461112, member: 33719 wrote: By nature the output signal is asymmetric. It is a wonder that it works so well, but it does.
OK, that goes some way to explaining why it doesn't work correctly with transformer-input pre-amps. From your colleague's response, it seems that, as I suspected, the equipment was designed assuming it would feed a fully differential input rather than simply a balanced input. Many non-transformer pre-amps have fully differential inputs, so it should not be too difficult to find others that work in this application, maybe even over a wider range of gains. Audio quality would be a separate consideration, and may narrow the field to a very few candidates.

As far as the RME Quadmic goes, it doesn't feel right that the output noise level should make a significant jump between two gain settings. In the quest to find the source of this, a quick test would be to try it again when connected to the rig, but with the rig powered down. That would give a crude indication whether the noise jump was coming from the rig when powered (probably in the form of an h.f. oscillation), or whether the Quadmic was exhibiting instability at medium gain settings with that input cable. The cause may be more subtle than that, but at least you could get an indication of whether or not your Quadmic itself had a problem.

niclaus Mon, 05/20/2019 - 08:08
Ahahah!! will try that then.

The hf modulation thing is what this guy seems to think it is as the RME quadmic MK II supposedly has a frequency range of 5hz to 200khz!
He used to have it working with the quadmic mk I and as he remembers he did not have that problem. We will try and and get a MK I and see if that is better.

Boswell Mon, 05/20/2019 - 11:21
OK, that indicates it's likely to be due to one or both of two things:

1. In addition to the audio signals, the laser rig is outputting inaudible frequencies (above or below the human audible range), which the Quadmic pre-amp is having trouble dealing with, starting at a certain gain setting. This gain setting could well be the point at which rogue high frequencies in the waveform cause excursions that reach the amplifier's slew rate limits, or that low frequencies cause the signal to hit the rails.

2. When the laser rig is powered, its output impedance has a complex component that interacts with the microphone cable inductance to cause a high frequency instability in the pre-amp. This might sound far-fetched, but I've had an instrumentation system misbehave in exactly this way, and it was something I found hard to believe at first.

If you have the time and inclination to persue this further, the diagnostic tool for both of these conditions is an oscilloscope.

niclaus Thu, 05/23/2019 - 07:40
Hey Boswell,

Sorry for the late reply, I spent the last few days with my head in that thing alongside the inventor. And what I have learned is that for some reason, it works and I should accept the fact that it does and embrace it. I mean when you know how to set it up, it sounds really good. We tried a bunch of things and we have to accept that even though there are problems, if you know how walk around them, it does what we want it to do, and it does it well.
I wonder what this guy could have achieved working with a knowledgeable guy like you though.

But, the high frequencies are probably the cause of that behavior since the input signal is almost (if not completely) a square signal.

Anyway, thanks a lot for your help. I really appreciate your time and effort and I will let you know if at some point we are able to make it better.

Boswell Thu, 05/23/2019 - 10:07
Most audio pre-amps are not guaranteed to behave impeccably when fed with large-amplitude square waves. That said, it is one of the standard set of tests I do on the pre-amps I have designed.

If you are able to, I would try feeding the input of the RME Quadmic through a balanced low-pass filter set to (say) 25KHz, just to see if the noise levels then track linearly with gain.

Tags

x

Register

Your recently read content