I would like to use audacity to measure the noise levels, RIAA response and any othe useful data I can source using this software and the tools I have available to me. Mostly I’ve built equipment from kits, but this is not one so I’d like to gather as much performance data as I can to find ways of optimizing.
I have the following:
UCA-202 interface from Behringer (one input one output)
Very low noise solid state Preamplifier with two preamp outputs
Inverse RIAA Filter
Phono preamp (obviously)
To avoid risk of directly plugging in the tube phono to the interface, I have put it through my preamplifier and linked the preamp outputs to the berhringer uca-202.
I recorded “silence” with no attenuation (preamp gain at full tilt) and went to Analyze Spectrum and that’s about as far as I’ve gotten.
This was my result: (see attachments)
I was advised I should find the “0db” level, but am a bit confused what is meant by this?
What is the best and most sensible way to go about accomplishing this? I would like to get repeatable results that I could share with others and compare but unsure as to how to create a baseline with differing measuring equipment in the chain.
Is there any other relevant data I should be looking for (as far as preamp performance) that I have not yet considered?
This is my first attempt at doing this, as well as using audacity so if you could please hear your response as if you were talking to a complete numbskull!
Also: I figured this would be the best location for this question, as it’s not really a problem I am experiencing with audacity or something compatibility related. Let me know if I was mistaken.
I think that’s a little ambitious. Every measurement is going to be a combination of the responses, noise and overload characteristics of every analog device in the channel.
I’m a little foggy what you’re going to share with others. “I use this preamp and my noise is this?” How many preamps do you have at hand to test? Good testing involves holding The Whole World stable and cycle one small part of it through unknowns. There is no sharing with other people using different A/D converters, for example.
For one small instance, your graphics postings have significant noise at 60Hz, 120Hz and 180 Hz typical of a power or shielding problem in the US. It is required you make that go away before you make any tests.
If you scroll over to the left of each scan, you see that the noise and trash goes all the way down to 2Hz. Real Sound only goes down to 20Hz. Everything to the left of 20Hz is analog or converter error. You have to resolve that, too. That trash will throw off noise measurements.
The latest AudioBook Mastering suite has a tool which cuts all that stuff off before processing, but it can affect the sound quality a bit. You wouldn’t want to do that to music.
I’m curious what you plugged into the preamp. Is there an actual cartridge plugged in? Did you connect the thin, black ground wire from the turntable? It’s there to suppress 60Hz hum…like the hum you have.
My presumption was based upon the general specification and performance characteristics I normally see published with consumer equipment. I am wondering what those would be in this use case.
While not an exceptionally precise and controlled experiment, there must be guidelines to ensure at least a general level of uniformity and point of comparison. I have friends in other geographic locations who are building the same device.
There are no grounding issues.
I believe the response curve is due to my measurements being at a nonideal decibel level or other factor I’m not considering.
Shielding? That 60/120/180 is coming from somewhere. It would be really good to sort that before you tried to pump music through it. Music systems don’t have problems like that.
Phono Preamps have RIAA compensation. Bass notes don’t fit on a record, so they’re reduced in volume before pressing. The Phono Preamp’s job is to push the bass notes back up where they were during the performance. Unfortunately, 60/120/180 are in the boost range.
I’ve been dancing around this, but intentionally putting a vacuum tube into a system may not be the best idea. Especially in sensitive preamp service, you have to do magic things like run the filaments on DC instead of AC (even on indirect-heated cathodes), and provide heat resistant metal shielding. That means you can’t use Mu-Metal magnetic shielding because Mu-Metal turns to trash in high heat.
None of that is easy or cheap.
Did any of this appear in the instructions for the preamp?
When you say you “recorded silence” exactly what did you do? I the preamp still connected to the turntable/cartridge? Are the inputs unconnected?
The correct way would be to install a 47k ohm resistor across the inputs because that is the nominal impedance that all turntable cartridges (or at least moving coil ones) are supposed to have. Leaving the inputs unconnected will usually lead to larger measured noise levels as the sensitive inputs pickup EM noise (like the obvious AC hum in your spectragrams). Shorting the inputs out will lead much lower (and equally wrong) noise measurements. Leaving the turntable connected will have two possible issues: First many turntables include a relay that shorts out the cartridge leads (mutes them) until the stylus is safely in the record groove (leading to the “too low” measurement). Second many cartridges are quite sensitive to EM radiation and will often pickup hum from nearby sources (leading you to improperly blame the preamp for the noise.).
As Koz points out you also will need to carefully characterize all of your measurement instrument (in particular the Behringer) before you can trust that what you are measuring is the device under test.
All that said, barring any issues with the USB bus or hum from your laptop’s power supply I do believe you could do meaningful measurements. I would not bother with either the very-low noise preamp you say you have, as your device under test’s output should be consumer line level. Or the “Inverse RIAA FIlter” unless you know it is an accurate passive design. Instead I would probably recommend using as simple as an attenuator between the Behringer’s output and the input of the Phono preamp.
A ruler that is not calibrated is called a “stick”, and you can’t expect accurate or meaningful measurements with a stick, you need a calibrated ruler.
To make accurate measurements you need your measuring equipment to be calibrated, so your first question needs to be, “how do I calibrate my test rig?”
The usual and most convenient way to do that is to use test equipment that is already calibrated, but if you have the equipment to do that, then you would do better to use that calibrated equipment directly to measure the Tube Phono Preamp.
Without calibrated equipment, it’s a case of Jack lifting himself up by bootstraps. It may still possible to make meaningful measurements, but it’s not easy and you have to be clever about it. For example, you may be able to determine the SNR because it is a ratio rather than an absolute measurement, but only if the SNR of the equipment that you are testing is substantially worse than the SNR of the test equipment (which is probably not the case), otherwise the noise that you are trying to measure will be swamped by the noise of the test equipment.
What I meant by no major grounding issues is that there’s no audible hum for me during normal playback. I’ve used this in normal listening before attempting to make measurements. Obviously there are some power supply issues that need to be sorted out.
Yes I agree the first question should be how to I calibrate / create a baseline or reference point.
I’d like to remove the vinyl from the equation and use an inverse RIAA so that results are more uniform.
I don’t require laboratory grade readings I’m just trying to do the best with what I have available to me.
I thought my initial question was how to do just that?
Flynwill’s response came closest to providing an actual course of action.
MIne is in a box somewhere (and it’s not CBS, but some other version… HiFi magazine I think)
Steve is right I did neglect to address the calibration issue. I would probably start by calibrating the output of the Behringer, and then use it’s output to calibrate the input. I’d probably just use my relatively new Oscilloscope as for this sort of work 5% accuracy is probably ok. A good DVM can probably be used to get a calibration starting point as well. But as Steve point out you have to be a bit clever, very meticulous, and willing to accept that your measurements will probably not have laboratory accuracy.
CBS Laboratories. They were heroically obsessive. They claimed a calibrated, constant-velocity lathe and they published what the rainbow looked like when you held the vinyl disk up to the light. Some engineer died happy.
60Hz is rough to hear, that’s an organ pedal, and 120 and 180 are very low level. I’m not shocked you missed the sound. That’s why you apply instruments and tools to find under-the-radar problems like that.
Can you explain this statement: I would probably start by calibrating the output of the Behringer, and then use it’s output to calibrate the input.
Or, can you provide a link or example of this being done?
I appreciate all of your willingness to help, but I still do not understand in the slightest how to apply anything of what you are saying in a practical use case. I’m not asking for a 10-page walkthrough I understand this is a forum but if I am to learn I need to know where to look, and for what, etc. Any links to useful information would be appreciated.
How does one “calibrate” the input and output?
Using a CBS record, what recorded volume should I be using?
What amount of attenuation is correct? Or, if there is no “correct” what should I be looking for?
If it is variable, how would I know when I have reached the right amount of attenuation in my use case?
I understand there is no basic on-size-fits-all solution, but also without any general reference points / approximations of things to look for I am adrift at sea.
“The correct way would be to install a 47k ohm resistor across the inputs because that is the nominal impedance that all turntable cartridges (or at least moving coil ones) are supposed to have. Leaving the inputs unconnected will usually lead to larger measured noise levels as the sensitive inputs pickup EM noise (like the obvious AC hum in your spectragrams). Shorting the inputs out will lead much lower (and equally wrong) noise measurements. Leaving the turntable connected will have two possible issues: First many turntables include a relay that shorts out the cartridge leads (mutes them) until the stylus is safely in the record groove (leading to the “too low” measurement). Second many cartridges are quite sensitive to EM radiation and will often pickup hum from nearby sources (leading you to improperly blame the preamp for the noise.).”
Yes what you need is to establish the relationship between actual signal levels in volts and the digital values that you will likely use for the actual measurements.
I would use Audacitys generate function to create a sine-wave of known amplitude play it out the UCA-202 and measure the actual output with whatever instrument is available. You might measure the P-P amplitude with an oscilloscope (hopefully a reasonably calibrated one) or the RMS (or pseudo-RMS) amplitude with an DVM. If the former I’d probably use a 1 kHz sinewave, for the latter I would use a 60Hz sinewave unless I knew for sure the DVM was rated to accurately measure a higher frequency. For this test I would set the software playback gain controls their maximum setting so you have an easily repeatable setup there.
I would probably do the measurement twice, once driving only the measuring instrument and again with 1K resistive load in order to also measure the output impedance of the UCA.
Next step is to connect the output of the UCA-202 back to it’s input and to play the same sinewave while recording in Audacity. Again set the gain controls to the maximum, and this will give you the relationship between volts on the input and the digital values.
Once you have a calibrated input and output you have a basis to actually measure the amplifier. Did your reverse-RIAA filter come with a specification as to it’s response? My inclination would be to build a passive attenuator (or possibly attenuators) from 1% metal-film resistors to get the level out of UCA-202 down to Phono cartridge levels.
You sound like you know some EE, how good are you at impedance calculations?