What test equip do I need?

[Bowie]

I've been using spectral analyzers, signal generators and 'scopes' in my DAW for testing and calibrating gear for years but it's super clunky to use and not as thorough as I'd like. In fact, I'd really rather get actual hardware to run tests on my bench as dragging opened gear over to my studio PC is not ideal. The more I read on the subject the more I've grown confused at the options. Part of the problem being that there's 20 years worth of posts online and they seem to all change as the technology develops.

Is the general verdict still that FFTs on cheap scopes are garbage? Should I get a cheap ol analog scope and keep using my DAW for the spectral analysis? I'm wide open to suggestions as I really don't know what people are using, but it's got to be better than the cumbersome way I'm doing it.

The work involves testing completed builds, mods, some repairs.

Many thanks.

 

[pucho812]

You can get yourself a used AP machine and the like off ebay, that usually will cover all things. get a DMM, and an o-scope and you are golden.  good analog scopes and AP's can be had on ebay for cheap.

 

[Bo Deadly]

A halfway decent USB audio interface is about all you need. Honestly the ADCs on even the cheap ones are so good now, the noise and distortion is probably defined entirely by the analog circuits interfacing them. Some of them you could literally plug in a mic with no pre and the noise floor would still be below the ambient noise picked up by the mic. Meaning you could do all gain digitally and it wouldn't matter. BUT. What really makes a difference is the software. I still don't love any of the usual programs for looking at spectrums and such. That's why I use GNU Octave (aka poor-man's Matlab) to make my own plots from custom code that can sit there and average lots of samples together to make high-res plots. The DAW tools are pretty crude by comparison (although they are sometimes convenient and I use them accordingly).

A scope is still necessary when looking at DC voltages and higher level signals. For that, a cheap digital scope is fine. But yes, the FFT feature on many digital scopes is pretty much useless (completely useless on my Rigol).

 

[user 37518]

For audio measuring (THD, IMD, Freq. Response, etc...) there are basically 3 options, of which I've tried them all, and all are good options:

1st.- The best option IMHO, like pucho mentioned, get an old AP SYS-1, you'll want to buy a USB interface from Duke (Audio1Man here on GDIY) so it works on WIN10, unless you want to have an old windows 98 PC just to run it. With AP, the more features you want it to measure, the more it will cost, the basic units do freq. response and THD among other things like crosstalk, but if you want FFT and IMD you will need to pay more. Overall this option will cost you between $1K-$2.5K USD depending on which features you need. Be sure to be well informed on what you need to have an AP system running, there are a lot of things you need to take into account, again, Duke may offer some guidance. Also, the Sys-1 is bulky AF, so its not really meant to be moved around

2.- The intermediate option.- Get a QuantAsylum QA401, it does basically everything you'll ever need like IMD, THD, Sweep Freq Response, Power, FFT, Impedance, and more, it has similar specs to the AP SYS-1 and it costs around $500 USD, you can skip all the calibration since everything has been done for you just take it out of the box, plug it in and you are set to go. Its very portable.

3.- The cheap option, get a Scarlett 2i2 3rd Gen sound interface and download a copy of ARTA, REW, or whatever software you prefer, I've had no issues with this setup. Like squarewave said, sound cards these days are extremely good.

For electronic testing get a DMM or better yet, get 2, buy something from a reputable brand, doesn't have to be a Fluke but don't get the Chinese cheapy ones, specially if you are going to be working on high voltages, the recommendation I usually give is this https://www.amazon.com/dp/B07W1BL3RH/?coliid=I8WT321QHXHJ9&colid=1L6ZIX9IE7HWU&psc=1&ref_=lv_ov_lig_dp_it
its an Amazon multimeter, yes Amazon makes multimeters now, actually no, apparently its made by a company called CEM which is a reputable brand and rebranded it for Amazon. But the meter is a bargain if you ask me, 6000 counts, IP67, True RMS, CAT IV, and can measure almost everything you'll need. I've seen reviews on Youtube and they all say this is a wonderful meter.

You will also want an oscilloscope, you can either buy a digital one or a used analog. For the digital scopes the usual low end brands are good Rigol, Siglent, Owon, Instek, Hammeg, etc.. I actually prefer buying a low end Tektronix which doesn't have as many feautres as the other brands I mentioned but it has the Tektronix name to back it up. I agree with squarewave, most low end scopes have a crappy FFT display, my Tektronix sucks for FFT. The scopes that have good FFT or spectrum analyzer are usually considerably more expensive.

You might also want to look at the Analog Discovery or Red Pitaya stuff, some people really like it, however, for audio I still prefer something like the QA401 and standalone gear.

If you are testing/designing your own gear, you may also want to buy a dual variable power supply, they are quite affordable and unless you are designing stuff like power amplifiers you don't need too much current capability.

And last but not least, buy a function generator, it can be the most simple one, I know that you can generate waveforms with your computer but its not as convenient as having a dedicated unit to do this.

 

[gyraf]

..thanks, user 37518! Good overview on a complex landscape

 

[Bowie]

Thank you all for the help. I really appreciate the info. I'm looking into all recommendations and learning what will suit my needs best.
I'm a lot more productive with intuitive gear so if anything leans in that direction, let me know.

 

[ruffrecords]

Hey user 37518. Excellent, readable summary of the main options. Worth saving for regurgitation every time this question gets asked.

Cheers

Ian

 

[Winston OBoogie]

Another thumbs up to user 37518 from me too  8)

 

[john12ax7]

Fur starting out I would recommend getting a scope,  signal generator, adjustable power supply,  and 2 DMMs.  Then use a computer / soundcard combo for FFT. This would be a solid and economical setup.

 

[Gold]

Excellent summary. I'd add that with DMM's you have to read the spec's to see the frequency range it covers. Most only are accurate to 1K or so. You pay a premium for one that is accurate over the audio band and further.

AP Portable One/ATS One's have been going for about $1500 on the bay. The testing isn't as comprehensive as a Sys1 but it's a stand alone unit that doesn't require a computer. Analog measurements. No FFT. 

I'd also recommend against the sound card/software route to start with. You have to calibrate it for accurate results. For a newbie this is not very easy and there's always that little bit of doubt it's not right.

 

[Brian Roth]

Excellent summary. I'd add that with DMM's you have to read the spec's to see the frequency range it covers. Most only are accurate to 1K or so. You pay a premium for one that is accurate over the audio band and further.

user 37518 made great comments, and Gold raises an important point.  I looked at the specs for that Amazon meter and the AC voltage ranges have a frequency response from 45 Hz to 1 kHz.  This very common for the less expensive meters.  "True RMS" in the specs does not mean the frequency response covers the entire audio range; that is an independent spec.

Bri

PS, that Amazon looks like a great value for doing many tasks.  I mihht even get one myself as an adjutant for my Fluke.

 

[user 37518]

Excellent summary. I'd add that with DMM's you have to read the spec's to see the frequency range it covers. Most only are accurate to 1K or so. You pay a premium for one that is accurate over the audio band and further.

AP Portable One/ATS One's have been going for about $1500 on the bay. The testing isn't as comprehensive as a Sys1 but it's a stand alone unit that doesn't require a computer. Analog measurements. No FFT. 

I'd also recommend against the sound card/software route to start with. You have to calibrate it for accurate results. For a newbie this is not very easy and there's always that little bit of doubt it's not right.

The AP Portable One is a good tip. Also, you are right about the bandwidth limiting up to 1KHz, another limitation is that the meter will only guarantee readings down to 100nF in the capacitance setting, can't have everything I guess, however, I still think the meter is a bang for the buck.

 

[Gold]

The reason I mentioned the bandwidth limitation is that it could be confusing for a newbie. Many use 1K as a test signal. That’s right on the edge of the measurement bandwidth. If you don’t know you will get dubious results on AC measurements.

 

[user 37518]

The reason I mentioned the bandwidth limitation is that it could be confusing for a newbie. Many use 1K as a test signal. That’s right on the edge of the measurement bandwidth. If you don’t know you will get dubious results on AC measurements.

Yes the most common test signal is a 1KHz sine, however if you would like to test a 1KHz square wave you are out of luck since it has many odd harmonics above 1KHz.

 

[dirtyhanfri]

Great post user 37518, it much resumes my findings in the last years (except the AP part, no experience with them)

My personal choice was QA401, but I'm thinking in get a used AP in the mid term, by the moment I'm more than happy with my setup:

Grundig MV5-0 Milivotlmeter for fast indication and signal tracing, got it dirt cheap in ebay and I love it, massive meter with db markings and a small scope CRT.

Escort EFG3210 Function generator, 2MHZ, nothing fancy, but works really well in conjunction with the Grundig.

A pair of cheap DMMs, one of them stays in my bench, the other one is in my portable toolbox.

Trio - Kenwood old analog scope, not much used actually, need some calibration and refurbish though.

Bitscope USB digital scope, used mainly for the logic analyzer, eventually will be fitted in a case with a Rpi and a 7" touch display, someday...

A crappy digital handheld scope from Aliexpress, helped me to assure signal integrity in some studio wiring jobs (the classic "wiring is not the problem").

And of course, my beloved probe connected to an old speaker and a LM386 amplifier.

Where I'm not able to find a nice solution is test leads, usual scope probes are too big sometimes, cheap test clips in different form doesn't seem to work very well and break to time to time, or maybe it's just my fat fingers...

 

[Michael Tibes]

I haven't read it all, but this looks like a very useful thread. Maybe it should be moved to the chamber? I'd never look for something like that in the brewery.

Michael

 

[ruffrecords]

Just had a thought. One problem using outboard sound interfaces like the Scarlett is their limited bandwidth even at high sample rates. However I believe the Quantasylum can go up to 100KHz. Add to this that it will now work with ASIO so you can use it with any ASIO compatible test software should mean together with REW it would make a very powerful test package.

Cheers

Ian

 

[fazer]

Yes excellent thread that should be in the chamber probably.  I have an old amber 4400 analog analyzer.  It needs some kind of repair every time I use it.  Seems like the QA 401 would be a great replacement. 

 

[john12ax7]

Just had a thought. One problem using outboard sound interfaces like the Scarlett is their limited bandwidth even at high sample rates. However I believe the Quantasylum can go up to 100KHz. Add to this that it will now work with ASIO so you can use it with any ASIO compatible test software should mean together with REW it would make a very powerful test package.

Cheers

Ian

There is a plot in the data sheet.  Seems to be -3dB at70kHz, and around -0.2dB at 20kHz.

Anyone compared it directly with other sound cards?

I find the choice of BNC connectors rather curious for something designed to measure audio.

 

[Gold]

I find the choice of BNC connectors rather curious for something designed to measure audio.

I generally like BNC connectors for their sturdiness but I don’t like them as used either. I use 5 pin XLRs on my builds. It would be difficult to fit four BNC cables in a 5 pin XLR. That means extra adaptor cables. I try to avoid extra cables and connectors in a test setup. One more thing to go wrong.

The first QA interface was unbalanced IIRC. Substituting a BNC for a phono connector is a good idea. I do that. Not so much for an XLR.