REW and Testing Electronics - what interface do you like?

dc2bluelight

Member
Thread Starter
Joined
Mar 17, 2018
Messages
68
Does anyone else use REW for purely electronic device testing? I use it for all sorts of testing of devices, everything from phono preamps to amplifiers, mixers, transformers, etc.

If you do the same kind of thing, what audio interface do you like to use with REW, and why? The requirements for testing electronics are a little different from making acoustic measurements.
 

DanDan

Moderator
Staff member
Joined
Mar 10, 2018
Messages
743
I have used the iMac Line I/O for testing. The smallest MOTU is highly rated.
 

dc2bluelight

Member
Thread Starter
Joined
Mar 17, 2018
Messages
68
Thanks.
My 2015 MacBook Pro only has a headphone jack, so not really useful. And, Mac line I/O on other models has basically consumer level, and unbalanced. So I use other interfaces. Just wondered what others use.

Yes, the current Motu M2 is pretty good, very clean. I have one. It suffers from limited output level, but so do most interfaces. It maxes at +16dBu, probably somewhat less for low distortion. For pro audio level testing you really need higher output. Otherwise the M2 is pretty impressive.

Anyone else found something good?
 

trobbins

Member
VIP Supporter
Joined
Nov 26, 2017
Messages
121
Location
Australia
Apart from signal output level, the configuration of input signal needed or preferred may be a significant reason for choosing a certain interface (ie. balanced/unbalanced and input impedance) so perhaps best to provide some insight in to your mix of testing experiences. For example do you use the M2's 1Megohm unbalanced input for any specific testing (as I would typically use that type of input for 95% of bench testing as it couples well with 10x and 100x scope probes), as not all interfaces provide that format and impedance.
 

dc2bluelight

Member
Thread Starter
Joined
Mar 17, 2018
Messages
68
For audio performance testing my personal "model" would be something like a Tektronix AA501, 100K input, full differential with high CMRR, 100kHz bandwidth or greater, capability to analyze signals as high as 30Vrms (The AA501 could deal with 200V!) and down to somewhere around -40dBv. I don't expect to pull that off with an audio interface without external attenuators of course. For output, be able to generate signals up to 30Vrms (I know that can't happen without an external amp), differential or single ended, with variable output that can be calibrated for 0dBFS at minimum of -40dBv, and up to 30Vrms. That would pretty much cover every application, and I know that's not going to happen in one box, but I'm trying to see how close I can come without breaking the bank. I might use the high Z inputs on the M2, but certainly not for everything. I'm not using REW purely as a scope, I have a real scope with 10X probe for that if needed. Its the signal analysis functions that I'd use it for.

One problem I'm seeing with lower cost interfaces is that calibrating input and output so that 0dBFS doesn't clip is not always possible. I have one low-cost interface that has some sort of gain ahead of the ADC that severely limits its maximum input level, and several interfaces that won't actually meet their maximum output spec without clipping. Not that it's a huge problem, but you have to work with your brain fully switched on. It's always surprising to see an interface that states it's "reference level" is +4dBu, but then t's maximum level is +13 or so.
 

trobbins

Member
VIP Supporter
Joined
Nov 26, 2017
Messages
121
Location
Australia
I use a USB soundcard interface (EMU0404) that is likely a close equivalent to the M2 in performance. I typically always have a 100:1 scope probe connected by metal adaptors (BNC to RCA, then RCA to 6.5mm mono) to the soundcard to use the unbalanced 1Megohm input. The dynamic range available is 'large' (eg. circa 120dB) so even a low level signal is still going to have a noise floor < -60-80dB, or I swap to a 10:1 probe or even direct connection if I am certain of the signal source. The voltage capability of the 100x probes is above 4-500V, although I typically don't connect to valve anodes and not above about 300Vdc as there is no practical need. I only rarely use the scope plot of REW, as all the detail is typically in the spectrum. I rarely break out my 100MHz DSO for common audio-related benchwork.

The M2 is likely to extend to 90kHz, I know the EMU has circa 90kHz -3dB bandwidth on ADC and DAC sides, and a compensation file provides flat measurement response out to 90kHz (even with 100x probe that can droop response above 60kHz for the probe I use). It's only very esoteric hi-fi situations where 90-95kHz bandwidth is not sufficient and I have to augment results with the DSO and other specialist instruments.

Any common soundcard is preferably used with a 'good' voltmeter - ie. with at least 100kHz bandwidth - for when absolute measurements are required - that's sort of a given.
 

jschwender

Member
Joined
Mar 16, 2021
Messages
140
Location
GERMANY
More  
Front Speakers
Nubert digital pro
Other Speakers or Equipment
Philips dss940
I do use a Steinberg UR22nk2 and do component testing, cables, coils, plugs… Although this device has high output level it is not the best choice as it has huge distortion both on the input and output if you go to levels above 5Vss. I use an amplifier in order to compensate for that and to be able to measure low impedance values. So for this use case, the UR22 is ok, but for distortion measurement it is unusable. I used a ASUS U7 sound card, which has excellent distortion values, but this one has serious issues with the USB interface, with the effect that REW sees channel swapping all the time, this makes it unusable. Also tried a soundblaster G6, which was advertized 32bit audio, but it is a complete desaster: resolution does effectively not exceed 14 bit distortion up to 2% -> returned immediately. At the end, if you really need high quality, you can't get around something like a focusrite clarett or audient id4.
 

DanDan

Moderator
Staff member
Joined
Mar 10, 2018
Messages
743
It is probably a matter of budget. I took a quick look at some potential 2 Ch interfaces.... Even the RME does only 19dBU.
 
Joined
Apr 25, 2017
Messages
95
Location
Omaha, NE
More  
Preamp, Processor or Receiver
StormAudio ISP.32 AVB
Main Amp
Digital Amp Company Cherry MEGASchino MONO's
Additional Amp
Digital Amp Company Cherry Maraschino's
Universal / Blu-ray / CD Player
HTPC
Front Speakers
JTR Speakers Noesis 215RM
Center Channel Speaker
JTR Speakers Noesis 215RM
Surround Speakers
JTR Speakers Slant 8
Surround Back Speakers
JTR Speakers Slant 8
Front Height Speakers
JTR Speakers Slant 8
Rear Height Speakers
JTR Speakers Slant 8
Subwoofers
4 ea JTR Speakers Captivator S2's
Other Speakers or Equipment
Infinite Baffle - 8 Acoustic Elegance IB15 drivers
Video Display Device
JVC DLA-NX7
Screen
Seymour AV Retractable CenterStage XD

dc2bluelight

Member
Thread Starter
Joined
Mar 17, 2018
Messages
68
Nice to know that if you have less of a funding limit you can hit +24dBu. But that also confirms that if I needed anything like +30dBu it's going to take an external buffer amp. At that point, +24dBu becomes unnecessary too.
 

JLM1948

Member
Joined
Jan 12, 2019
Messages
124
Since I haven't found on the market anything that comes close to an audio analyzer front-end, I have built an "interface interface" with 1Megohm input impedance balanced, -40,-20, 0 and +20dB gain control and a balanced output stage based on a THAT1646. I have successfully used it with Behringer, Native, Tascam interfaces, and now a Cosmos ADC and Topping DAC.
I think looking for THE perfect soundcard that does it all in one box is a lost cause.
 

JLM1948

Member
Joined
Jan 12, 2019
Messages
124
I use this sound card https://www.behringer.com/product.html?modelCode=P0484 and this probe https://www.warwickts.com/1122/TT-LF-212-Passive-Probe for inside the amplifier, but I think I damaged the probe recently.....oops.
Are you aware that the probe is supposed to be used in conjunction with a unit that has an impedance of 1Megohm? It is the normalized value that all oscilloscope manufacturers abide.
In the direct (X1) position, it's not an issue, but in the X10 position, the actual attenuation is 370. So you may not have damaged the probe, but you mus tuse a scaling factor of 370 instead of 10. I'm afraid it results in very low readings.
 

tonescout

Registered
Joined
May 30, 2021
Messages
19
ok - that's interesting, it worked and then it stopped working (very low or non existent signal) I was not aware it was able to operate in 2 settings and as an obvious numpty I have no idea how to make the adjustment back to 1X ??
 

jschwender

Member
Joined
Mar 16, 2021
Messages
140
Location
GERMANY
More  
Front Speakers
Nubert digital pro
Other Speakers or Equipment
Philips dss940
There is a sliding switch at the handle on the probe which has two positions. It is also labeeled 1x and 10x.
 

xmag

New Member
Joined
Nov 11, 2021
Messages
38
I'm not very good at writing English but I'm trying!
I usually use RMAA and a USB sound card when I measure electronics and measure speakers with REW. Recently I tested measuring distortion for electronics with REW which I first measured with RMAA. In RMAA, for example, the noise measurement goes completely horizontally from 20hz to 20khz, but in REW, the noise measurement starts to go uphill from about 500hz. The curve looks like a small hill up to 20khz. Why do you get worse upward measurements in frequency in REW than RMAA when everything else is exactly the same?
 

JLM1948

Member
Joined
Jan 12, 2019
Messages
124
I usually use RMAA and a USB sound card when I measure electronics and measure speakers with REW. Recently I tested measuring distortion for electronics with REW which I first measured with RMAA. In RMAA, for example, the noise measurement goes completely horizontally from 20hz to 20khz, but in REW, the noise measurement starts to go uphill from about 500hz. The curve looks like a small hill up to 20khz. Why do you get worse upward measurements in frequency in REW than RMAA when everything else is exactly the same?
how do you do these measurements? Do you measure the complete chain (ADC+DAC) or just teh DAC? With or witout a DUT (Device Under Test) connected?
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
Sweep measurements have a rising noise floor. To measure electronics it is better to use the stepped sine option in the RTA.
 

xmag

New Member
Joined
Nov 11, 2021
Messages
38
how do you do these measurements? Do you measure the complete chain (ADC+DAC) or just teh DAC? With or witout a DUT (Device Under Test) connected?
Thanks for the quick answers! I measured the sound card with a loop cable. Nothing else was connected.
 

xmag

New Member
Joined
Nov 11, 2021
Messages
38
Really need to read the manual now.
Interesting reading in the REW link .
Thanks!
 
Top Bottom