• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Tossing my hat in the ring of electronics measurements and could use some advice.

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
Hello, all. First post here so if this doesn't belong in this sub-forum, please, don’t hammer me too hard…



I’ve been wanting to build an electronics measurement setup for some time now. I used to test loudspeaker drivers (some of you may have frequented my now-defunct website: medleysmusings.com). My foray in to electronics testing up until now had been stifled by hardware complexity, resolution and/or cost. And, to an extent, it still is. However, I recently decided the time is right and have put myself in the direction of going with an external USB soundcard based setup.

I have purchased the Pete Millet soundcard interface. Now, I’m down to the brass tacks of picking both the USB soundcard and the software to get me going.

As for what I’m wanting to measure, well, that’s anything from DSP, DACs, car audio headunits, to amplifiers. I’m looking to quantify what I can with at least the bare minimum set of data being frequency response, noise, and THD characteristics. However, I will not be measuring headphone outputs. (as an aside, at some point I would like to supplement the hardware testing with ABX testing)

Soundcard:
I have a Behringer UMC202HD for my in-room loudspeaker measurements but from what I’ve read here and elsewhere this thing isn’t necessarily ideal for electronics testing. In somewhat-recent tests here the RME ADI DAC looks like the best solution but I just can’t afford that. Is there an RME ADI-2 DAC alternative for measurement purposes only? Seems like I’d be paying for additional features (such as a fancy display and EQ) that I wouldn’t necessarily need. I was looking more in to the <$200 range (new or used). I could bump that up a little if there was a worthwhile option. I continually see Focusrite’s 2i2 mentioned as a good budget-friendly solution. Most of the modern soundcards would probably give me enough information to be dangerous but I’d also like to ‘future proof’ myself.

Software questions:
AudioTester and aRTA seem to be the “go to”. I’ll be using my laptop with Windows 10.
Some of the additional things I’d like to have the capability to do, if I can find software that permits it:
Stimulus applied via various sources such as the USB soundcard but also, namely for car audio (yes, yes, I know.. car audio is silly), CD to USB (external HDD and iPhone); can a stimulus be recorded and played back?
Overlaying of results (especially for power (voltage is fine) vs THD
Provide baseline (loopback) in same plate OR use the loopback to deduct any contributions from the measurement (I believe I read aRTA has this capability?)


Again, I’m trying to do my best to understand what I’m getting in to. I’ve read through this forum and diya a lot the past week but I can’t make up my mind on the soundcard and software so I was hoping for some suggestions here if you all don’t mind helping me out.

Thanks,
Erin
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,556
Location
Seattle Area
We have a long thread elsewhere on this topic. Net of it is that nothing out there is very satisfactory as far as software. ARTA is one of the better ones but still, you won't be able to do what I do. Async signal generator is cumbersome but can be done.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
REW works fine for me ... the developer even posts here and is very responsive.
Is free software.
Just like Audacity which can also be useful.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,687
Likes
4,068
I like my Focusrite 2i4 2nd Gen with the lastest drivers but seriously their stability is sh** on Windows 10. Still a good interface if like me you need recording or digitalizing vinyl.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
I second both REW and Audacity at least to start. Both free, both useful, REW is even automated somewhat and can save results to be viewed later if you wish.

REW can give you FR, distortion including graphed per harmonic, IMD both SMPTE and CCIR. Plus other things.

As for an interface well it depends upon what you hope to accomplish to what levels. I have a couple Focusrite units. It looks like they at least meet specs. I don't know there is a big difference in the $200 or so units. You might want to make sure you get one that has SPDIF in and outputs. That will come in handy for testing at some point. For the Focusrite Scarlett units that will be a 6i6 which is on sale for $229, but usually more like $269. Of course you could get one second hand off ebay or www.reverb.com to save a few bucks. Units like this will let you confirm whether something is pretty good or not very good, but you won't get state of the art results like Amir's AP unit.
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
Here are some results of a Scarlett 18i20 ADC being fed by a cleaner source. I think the lower priced ones are similar.

Here is a spectrogram. The background goes to light gray at -100db. 1st is a 1 khz tone. THD+N for that is -98 db, but all the individual harmonics are below -100db. 2nd is a full scale 18+19 khz tone. You see some sidebands as labeled and all other sum and difference signals are below -100db. 3rd is a sweep from 2hz to 20 khz. If there were distortion harmonics above -100 db you'd see them. 4th and last is a twin tone sweep with 1 khz separation. You see faintly some - 99 db sideband only at the upper frequencies. So this is pretty clean, but not clean enough to tell what the true limits of an RME or Benchmark is. The DAC on the 18i20 is slightly cleaner than is the ADC, but not greatly different.

1563841679407.png


Here is the same signal, but I have the spectrogram extending to -130 db which uncovers the harmonics and other things. Like the oddity that this ADC has tones spaced every 1 khz no matter what else is going on at a level of about -127 db.

1563842078894.png


For comparison here is a Zen Tour from Antelope audio with the background going to -130 db. It is much cleaner. Though still far short of an AP.

1563842462205.png
 
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
I like my Focusrite 2i4 2nd Gen with the lastest drivers but seriously their stability is sh** on Windows 10. Still a good interface if like me you need recording or digitalizing vinyl.



Yea, that’s what I’m afraid of. I’ve read a few posts stating the same thing. That’s unfortunate.





I second both REW and Audacity at least to start. Both free, both useful, REW is even automated somewhat and can save results to be viewed later if you wish.

REW can give you FR, distortion including graphed per harmonic, IMD both SMPTE and CCIR. Plus other things.

As for an interface well it depends upon what you hope to accomplish to what levels. I have a couple Focusrite units. It looks like they at least meet specs. I don't know there is a big difference in the $200 or so units. You might want to make sure you get one that has SPDIF in and outputs. That will come in handy for testing at some point. For the Focusrite Scarlett units that will be a 6i6 which is on sale for $229, but usually more like $269. Of course you could get one second hand off ebay or www.reverb.com to save a few bucks. Units like this will let you confirm whether something is pretty good or not very good, but you won't get state of the art results like Amir's AP unit.

I’ve used REW pretty extensively for loudspeaker measurements. Love it. I’ll consider going that direction for my electronics testing. I wasn’t aware Audacity was useful for that kind of thing, though. I’ll have to look in to that.



As for quantifying the levels I want, I assume that’s really based on the expectation of the gear I plan to test and making sure I can measure to at least that level. Most of what I plan to measure will be car audio stuff; and of that there are some units that play DSD files. So the capability for high resolution/low noise is there with the DUTs. But I’m basically just shooting for the best I can get for my buck, which is rather limited. My main area of concern, though, is noise; the SNR specs. Using compression drivers, noise is something I’d like to reliably measure rather than “hook the unit up and hope it’s quiet” and the way some manufacturers rate noise (such as “>90dB @ rated power”) isn’t useful because I need to know how quiet the unit itself is during quiet moments and low levels. It seems most of the gear I’ve looked at specs for have a noise floor rated no better than 120dB so if I had something that could at least test down to that level I’d be happy. Just not sure if that’s feasible; I looked at your test results of the 18i20 and saw you measured SNR down to about 107dB. Heck, I don’t even know how reliable some of those specs are so I could be basing my needs on numbers that aren’t legitimate. Which is the whole reason I want to do this kind of stuff.



Is there a consensus on breakdown of performance for our needs? It seems to me it goes in order of highest to lowest RME ADI-2 DAC à Motu Ultralite MK4 à Focusrite Scarlett or E-MU 0404 USB



I also found another potential option in the MOTU microbook. Results here indicate good performance:

https://johnr.hifizine.com/2012/09/microbook-ii-measurements-i/
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
Yea, that’s what I’m afraid of. I’ve read a few posts stating the same thing. That’s unfortunate.







I’ve used REW pretty extensively for loudspeaker measurements. Love it. I’ll consider going that direction for my electronics testing. I wasn’t aware Audacity was useful for that kind of thing, though. I’ll have to look in to that.



As for quantifying the levels I want, I assume that’s really based on the expectation of the gear I plan to test and making sure I can measure to at least that level. Most of what I plan to measure will be car audio stuff; and of that there are some units that play DSD files. So the capability for high resolution/low noise is there with the DUTs. But I’m basically just shooting for the best I can get for my buck, which is rather limited. My main area of concern, though, is noise; the SNR specs. Using compression drivers, noise is something I’d like to reliably measure rather than “hook the unit up and hope it’s quiet” and the way some manufacturers rate noise (such as “>90dB @ rated power”) isn’t useful because I need to know how quiet the unit itself is during quiet moments and low levels. It seems most of the gear I’ve looked at specs for have a noise floor rated no better than 120dB so if I had something that could at least test down to that level I’d be happy. Just not sure if that’s feasible; I looked at your test results of the 18i20 and saw you measured SNR down to about 107dB. Heck, I don’t even know how reliable some of those specs are so I could be basing my needs on numbers that aren’t legitimate. Which is the whole reason I want to do this kind of stuff.



Is there a consensus on breakdown of performance for our needs? It seems to me it goes in order of highest to lowest RME ADI-2 DAC à Motu Ultralite MK4 à Focusrite Scarlett or E-MU 0404 USB



I also found another potential option in the MOTU microbook. Results here indicate good performance:

https://johnr.hifizine.com/2012/09/microbook-ii-measurements-i/

Plenty of gear will have SNR of 100 to 105 db. You don't get very much better than that without spending more money. Also important to notice the difference between SNR and dynamic range. They are often very similar, but aren't the same measurement. SNR is the noise level with a max level fundamental notched out. Dynamic range is a -60 db signal notched out vs max signal level. That dynamic range will usually be a few db better plus most quote A-weighted which will add 4-10 db to the number.

Also, I assume you know the difference in noise over a given bandwidth vs what noise levels in FFT's look like. You can see FFTs with a floor of noise at -140 db or lower, but that is from FFT gain and doesn't mean you have a -140 db noise floor for the device.

You mention compression drivers, so I assume you want to measure amps to make sure they don't have noise with those drivers. I would think even a 2i4 is quiet enough for that. You will need to attenuate the signal on amps to measure them with a recording interface. At a minimum you'll need some resistor attenuators so the voltage is something that won't damage the interface.

There are some tricks for some measurements. As an example, I will feed a test signal to a DAC where nothing is over -60 dbFS. Feed it into microphone preamps. Okay as long as you are careful and don't send a larger signal. I can then apply gain to push the noise in the DUT well above the low noise floor of the microphone preamps. Then I can take the recorded result, adjust for the gain I used and see some really low levels of noise. You could do something similar with amps. Feed the output with a low level signal into the line inputs with gain in the 2i4(I wouldn't use the mic pres for this). Then correct the level in the digital file and you'll know the low level noise performance of the amp. You have to be careful not to accidentally feed it a high level signal or you might smoke the 2i4. You also might need to be careful of switching amps that have an ultrasonic switched voltage at high voltages. Using attenuating resistors is safer.
 
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
You will need to attenuate the signal on amps to measure them with a recording interface. At a minimum you'll need some resistor attenuators so the voltage is something that won't damage the interface.

Thanks for the information. I'm going to have to read it a couple times to digest it. But I wanted to note that I will be using the Pete Millet Soundcard Interface which knocks down voltages of up to 200Vrms down to 1Vrms. I've learned this lesson the hard way before. :(
 

SIY

Grand Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
10,386
Likes
24,752
Location
Alfred, NY
Thanks for the information. I'm going to have to read it a couple times to digest it. But I wanted to note that I will be using the Pete Millet Soundcard Interface which knocks down voltages of up to 200Vrms down to 1Vrms. I've learned this lesson the hard way before. :(

I have the Pete Millett interface and it works very well for its low cost; I made hundreds of measurements with it without killing it or running into significant issues. If you can pop for the extra money, Jan Didden's "Autoranger" is much better. I had to get a B series Audio Precision to measure it.
 
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
I have the Pete Millett interface and it works very well for its low cost; I made hundreds of measurements with it without killing it or running into significant issues. If you can pop for the extra money, Jan Didden's "Autoranger" is much better. I had to get a B series Audio Precision to measure it.

Funny enough, I contacted Jan early last year about buying an Autoranger and he was out. I contacted him a few weeks ago and he said he's got another batch coming through soon so he put me down on the list. I went ahead and purchased the Millet interface (used) because I don't know how long it will be before I get the Autoranger (or, frankly, if it will actually happen). I definitely appreciate the vote of confidence for it, though, because I was not even sure if it was necessary now that I have the Millet interface. So thanks for that.
 
Top Bottom