Hello. I want to measure opamps (noise vs R source, Distortion, THD+N), resistors (noise), PSU (noise) and Tube amplifier (Noise and distortion).
I have :
Am I good enough, what do I miss ? I need and uln amplifier, but what else ?
Right now I am looking to make measurement PCB in Kicad for opamp, measurements.
The journey is exciting !
I have :
- 12bit 2channel Oscilloscope (60MHz)
- 12bit 2channel Function generator (up to 25MHz - 200MSa/s sampling rate)
- 2 Bench PSU (I link them together for -+15VDC)
- Behringer UMC404 Soundcard (24bit/192Khz)
- EDIT : I already have 3 handheld multimeters 😉
Am I good enough, what do I miss ? I need and uln amplifier, but what else ?
Right now I am looking to make measurement PCB in Kicad for opamp, measurements.
The journey is exciting !
Last edited:
Not sure if you know who Sean Olive is? If not, here is some info on him: https://www.aes.org/aes/seanolive#:~:text=About,evaluation of new audio products.
Here is what he said about THD+N:
@Markw4: How is this possibly relevant to someone who wants to measure THD? I don't think @Frederico Acardi intended to start a subjectivist vs objectivist thread. Also, do you think it's possible that science (and the world in general) has moved on since that book was published?
Anyway... @Frederico Acardi I would add a voltmeter of some sort. If you are looking to measure noise voltages, a bench top DMM that can resolve AC voltages down into the uV range (or at least single digit mV with some precision) would be helpful. I use my HP 344401A for that. You can often find them for $400-500 used, calibrated. Make sure you get one that's calibrated.
Before you rush out and buy a bench top meter it would probably be prudent to build a low noise amplifier (LNA). That will allow you to amplify the noise to the point where your meter or oscilloscope can measure it.
The sound card is likely good enough to measure THD down around -100 maybe -110 dBc. If you want to improve on that, I'd look at getting a precision oscillator (Victor's for example) and a notch filter. A passive twin-T filter can be pretty handy.
Tom
Anyway... @Frederico Acardi I would add a voltmeter of some sort. If you are looking to measure noise voltages, a bench top DMM that can resolve AC voltages down into the uV range (or at least single digit mV with some precision) would be helpful. I use my HP 344401A for that. You can often find them for $400-500 used, calibrated. Make sure you get one that's calibrated.
Before you rush out and buy a bench top meter it would probably be prudent to build a low noise amplifier (LNA). That will allow you to amplify the noise to the point where your meter or oscilloscope can measure it.
The sound card is likely good enough to measure THD down around -100 maybe -110 dBc. If you want to improve on that, I'd look at getting a precision oscillator (Victor's for example) and a notch filter. A passive twin-T filter can be pretty handy.
Tom
@tomchr
I merely wonder if the gentleman wants to measure THD because he has heard that its particularly useful? If so, why let him suffer under a misconception? Is the right thing to do to encourage ignorance?
OTOH, if he knows THD is of questionable value and he wants to measure it anyway, then that's fine. I agree with your suggested items to acquire in that case. Or maybe one of IVX's boxes. Something like that. Also, might add that a DAC and LP filter as a test signal source may have less phase noise than a Victor oscillator.
I merely wonder if the gentleman wants to measure THD because he has heard that its particularly useful? If so, why let him suffer under a misconception? Is the right thing to do to encourage ignorance?
OTOH, if he knows THD is of questionable value and he wants to measure it anyway, then that's fine. I agree with your suggested items to acquire in that case. Or maybe one of IVX's boxes. Something like that. Also, might add that a DAC and LP filter as a test signal source may have less phase noise than a Victor oscillator.
With a decent sound card and a pc/laptop and a sw like REW you can measure THD, THD+N, harmonic spectra, frequency response, and noise down to the uV region using the microphone input. Furthermore the sound card can work as a low-noise low-THD sinewave generator inside the audio band.
I think the behringer he has already may be sufficient as bucks mentions. If I remember right, REW can do IM as well as harmonic distortion.
^^^ That is at least starting to suggest measurements which may be more useful than THD. However, human interpretation of harmonic spectra may be another issue to consider.
Also agree with mikeAtx, the Behringer may be good enough (if used with a notch filter and a clean makeup gain amp). With only 100dB or so of A-weighted SNR, questionable how useful it might be otherwise. A Focusrite Scarlett would likely be a step up.
EDIT Also, any thoughts from the group on time-domain distortion residuals as something the OP may wish to consider?
Last edited:
Link to recent post in Cosmos thread: https://www.diyaudio.com/community/...perform-apx555b-for-30000.386001/post-7492346
We are now able to measure reliably to -160 dBc levels for a very reasonable price. Here is measurement with APU and Cosmos ADCiso + thermostat board. Source is good signal generator with distortion < -150 dBc. APU was at default gain resulting with -10 dBFs for a 1.7 Vrms input signal.
Also some distortion residual examples:
https://www.diyaudio.com/community/...st-possible-thd-n-really-the-best-way.367692/
We are now able to measure reliably to -160 dBc levels for a very reasonable price. Here is measurement with APU and Cosmos ADCiso + thermostat board. Source is good signal generator with distortion < -150 dBc. APU was at default gain resulting with -10 dBFs for a 1.7 Vrms input signal.
Also some distortion residual examples:
https://www.diyaudio.com/community/...st-possible-thd-n-really-the-best-way.367692/
Last edited:
You're basing this entirely on the opinion of one person whom you've quoted out of context. And now you seem to be saying that those of us who measure THD are ignorant. Really, dude?!
A measurement of harmonic distortion is certainly indicative of how close to an ideal amplifier the amp under test is. "Ideal" in this context meaning OUT = IN * A. But as I've pointed out many times, it's one measurement out of many.
Tom
Ah! Good point.
I'd still add a digital multimeter if OP doesn't already have one.
Tom
No. If you know what the opinions of Olive and Geddes (and D. Massa) are, then how would that make you ignorant? Just the opposite: You are aware of their opinions.
Besides, you may have good and practical reasons for wanting to measure THD. For example, some of your customers might expect to see it. Says nothing useful about "the degree of distastefulness to the listener" of your products however (quotation from D. Massa as seen in #2 above).
Last edited:
The degree of distastefulness wasn't one of the parameters OP wanted to measure. At least I interpreted his question as "I would like to measure A, B, and C. Which equipment should I get? I already have these..." You seem to want to answer the question, "which of these measurements are indicative of high perceived sound quality according to Sean Olive and/or Massa?" I agree that the measurements vs subjective experience discussion is interesting and relevant, but wouldn't it be better served in a thread of its own?
Tom
Tom
Tom,
Reading the OP made me suspect the gentleman is new to this stuff. If I am correct, I don't think its wrong to start a discussion with him about his interest in measurements. My first post, #2, was intended to act as a probe to see if he already was aware that THD+N is controversial; it was to start a conversation. Depending on his response the follow up could be to provide him with more information, or to simply give him a list of additional equipment to acquire (as you did). IMHO the difference between your response and mine was that I wanted to make sure the OP knew what he was doing well enough to ask the right questions for his underlying purpose. You seem to have assumed he already knew enough to ask the right questions for whatever his underlying purpose might be. Maybe if he was given a chance to respond to me before making assumptions about where it might go, and or making assumptions about my intentions, then it might have turned out to be very satisfactory to the OP. I don't have a problem recommending the requested list of kit, if he is sure he knows enough to ask informed questions, or even if he doesn't wish to be informed. However, IME some people turn out to be grateful for getting a heads up before they charge into something without first understanding much about it.
Reading the OP made me suspect the gentleman is new to this stuff. If I am correct, I don't think its wrong to start a discussion with him about his interest in measurements. My first post, #2, was intended to act as a probe to see if he already was aware that THD+N is controversial; it was to start a conversation. Depending on his response the follow up could be to provide him with more information, or to simply give him a list of additional equipment to acquire (as you did). IMHO the difference between your response and mine was that I wanted to make sure the OP knew what he was doing well enough to ask the right questions for his underlying purpose. You seem to have assumed he already knew enough to ask the right questions for whatever his underlying purpose might be. Maybe if he was given a chance to respond to me before making assumptions about where it might go, and or making assumptions about my intentions, then it might have turned out to be very satisfactory to the OP. I don't have a problem recommending the requested list of kit, if he is sure he knows enough to ask informed questions, or even if he doesn't wish to be informed. However, IME some people turn out to be grateful for getting a heads up before they charge into something without first understanding much about it.
I'd be careful with choosing a voltmeter: while the vanilla DMM's are remarkably accurate for the price of a couple of burgers (OK, with sides), most of them are only good to about 1KHz: not for audio! A viable alternative is an old school needle meter, like my HP3400, if you can find it.
See my Leak-Enstein thread for a reason why you need a 1MHz meter, even when you're messing with a 59 y.o. tube amp.
See my Leak-Enstein thread for a reason why you need a 1MHz meter, even when you're messing with a 59 y.o. tube amp.
Alternatively you can
use a good DVM
set sine output frequency of sound card to 400Hz
and then calibratite its output using REW.
This procedure makes your sound card a precision, calibrated ac voltage meter covering the hole audio frequency range.
Furthermore you can generate a calibration file that equalizes the frequency response of your sound card.
use a good DVM
set sine output frequency of sound card to 400Hz
and then calibratite its output using REW.
This procedure makes your sound card a precision, calibrated ac voltage meter covering the hole audio frequency range.
Furthermore you can generate a calibration file that equalizes the frequency response of your sound card.
THD+N is not controversial. Its impact on sound quality can be debated, but measuring THD+N is done all the time with no controversy. It seems to me that you're trying to create controversy where none needs to exist.
The fact that THD isn't strongly correlated with a positive listening experience is well documented. Belcher (1978) springs to mind for example. Belcher further argues that a multi-tone signal should be used as multi-tone IMD correlates more strongly with the perceived sound quality. But even that is not a perfect correlation because no correlation is perfect.
You see that anecdotally here as well. Some have listened to various sound samples with different levels of THD to find their personal threshold of detection. You can see the results here: https://www.diyaudio.com/community/threads/take-the-thd-challenge.395242/ As you can see, some need 1% THD before the distortion becomes objectionable whereas others say they can hear 0.001%. So THD is not a perfect correlate of the perceived quality.
That does not mean that THD measurements are irrelevant or useless. I think we can all agree that an amp that adds 1% harmonic distortion is not as precise or as close to an ideal OUT = IN*A amplifier as an amp that only adds 0.001% harmonic distortion. Whether you prefer one over the other is a completely different question. I think we can also agree that if an amp adds 0.0001% harmonic distortion then the distortion is very likely inaudible, and if inaudible distortion is your goal then that amp could be a potential candidate for purchase or build for you.
So basically THD and THD+N measurements are as relevant as ever. They're not shrouded in controversy. But there are better correlates of sound quality than THD and THD+N. Multi-tone IMD would be a better option. If OP is interested in multi-tone testing, I'd say he's pretty well set up. All he needs is a multi-tone test signal to play through the sound card.
The HP3478A is another good option. They can be had in tested and calibrated condition for $150-200.
It looks like the handheld meters max out at 20ish kHz even for the good ones (such as Fluke 87V). The lower cost ones seem to max out at 1 kHz.
Assuming the sound card does not have any significant droop in the frequency response towards the ends of the audio band. Great suggestion, though.
Tom
Not among some people. Some people like THD+N, because more elaborate measurements would show what the problems with their audio products are. Also, maybe some people use it because they are designing measurement instrumentation, not music reproduction equipment. There may also be other people who are simply ignorant.
As far as using THD+N for music reproduction goes, according to Earl Geddes "THD and IMD have no correlation to the perception of the distortion that they are intended to represent." http://www.gedlee.com/Papers/The Perception of Distortion.pdf
Also, Geddes states: "In our next paper we will show that .01% THD of one type of nonlinear system can be perceived as unacceptable while 10% THD in another example is perceived as inaudible. Even one of these simple examples is sufficient to invalidate THD as a viable metric for discussion of the perception of distortion."
http://www.gedlee.com/Papers/Distortion_AES_I.pdf
He goes on to say: "It is well known that an MP3 sound transmission can have a measured THD of upwards of 50%, and yet be perceived by listeners as an acceptable quality reproduction."
http://www.gedlee.com/Papers/Distortion_AES_II.pdf
Geddes summarized in one of the documents on his website as follows:
...people are satisfied with THD and IMD. It’s like the story of the cop who asks a drunk under a street light what he is doing on his hands and knee’s. The drunk replies “I’m looking for my car keys.” The officer asks “Where did you loose them?” and the drunk replies “Over there by my car.” Baffled, the officer asks “Then why are you looking for them here?” to which the drunk replies, “Because the light is better.” Everyone knows that THD is meaningless, but it’s easy to do and “the light is better.”
http://www.gedlee.com/Papers/Comments on howard.pdf
Therefore, IMHO the use of THD measurements in the specifications of audio gear sold for music listening is quite disingenuous on the part of manufacturers, unless they are perhaps plainly ignorant. If not ignorant, and if they were honest then they should tell the truth and say that don't specify THD because it has no correlation to the perception of distortion by human listeners.
Last edited:
The Earth is shaped like a sphere. Some consider this statement to be controversial. I don't. I consider it to be a fact.
A correlation of 0.0 is a statistical impossibility or at the very least improbability. I'm going to assume that you mean a weak correlation rather than no correlation. I don't know what R value you use as the threshold for a low or weak correlation. I'm partial to Cohen (1998) who uses R = 0.1 for weak; R = 0.3 for moderate, and R = 0.5 for strong correlation.
That's why I keep hammering on the importance of showing multiple measurements, not just a single one.
I use THD measurements (among many other measurements) to show the performance of my products. Am I then disingenuous or ignorant? You stated above that I wasn't ignorant, so I guess I'm in the disingenuous category. And you base this exclusively on the fact that I make THD measurements available. Fascinating!
Anyway. Safe travels on your crusade.
Tom