Audio test gear - frequency sweeps and THD measurement

[AllTheGearNoIdea]

Hello all.
I have just started getting interested in repairing old audio equipment. I am completely tone deaf and would not be able to tell the difference between a super high end system and 80s Sanyo tape deck pulled out of a canal. So I would like to be able to do some characterisation using test equipment.  I have the usual collection of test gear a high-end scope plus various dubious vintage signal generator and an old Leader distortion analyser.     I would however like to automate the testing as much as possible to do frequency sweeps and THD etc.  Getting bored of manually adjusting the signal generator frequency, Manual levelling the Sig gen output then taking the amplifier output measurement, writing it all down spread sheet not sexy
I have had a look at some of the Vintage audio analysers from HP and I think Booton but these still demand high prices and to be honest I would be worried about  paying that kind of money for a knacker or something that’s faulty and cannot be repaired.
With all this in mind I have been looking at one of the 4000 Picoscope this is designed to have good signal integrity and very low distortion and is 12 bit sampling.  I am however a bit concerned about the signal generator as I understand is -+3dB across 20Khz.  Now is it me or is that awful ?
I am generally quite a fan of the Picscope stuff I have one of the 2000 series and its really useful and it’s great to have all the very powerful software features in a convenient little box. It’s very portable and really good for quick capturing screen shots. They also have a Dev kit  so I can tweak my own code.
I’m interested any advice here regarding audio test equipment and worthwhile testing I can do. I had a google round and could only really find people doing power measurement on YouTube for these ridiculous car systems
I also just want to buy some more Tech toys and have some fun playing with them as I  haven’t bought any new equipment in the last 3-4 hours.
Thanks Chris    - AllTheGearNoIdea

[MrW0lf]

I also just want to buy some more Tech toys and have some fun playing with them as I have haven’t bought any new equipment in the last 3-4 hours.

My last purchase was yesterday and already in great pain It all started when sold fast car, settled with specific girl and started consuming less beer
Currently I see as most interesting low-freq all-purpose PicoScope 4444, which is 4ch differential and 14bit. Might pose interest if want fight groundloops etc. However if youre into programming and want good signal gen I'd say go for Siglent SDG2042X. I use it almost every day and despite some little quirks excellent tool (16bit etc). Then you can program both, operate in sync and do some cool stuff. Note that Siglent does not have isolated outputs! Then there is Digilent AD2 which is 14bit 2ch and has 2x signal gen up to 10MHz but it's toy as all-purpose scope (effectively 10MHz max). But inside that 10MHz can do bode plot w phase.

[Alex Nikitin]

You can use a PC with a decent 96K or 192K sampling rate audio card and audioTester program. In my home lab I have a couple of RME Digi96 PST cards and uses these with audioTester regularly. You may need to make a proper front-end for the card, to accommodate the range of input voltages.

Cheers

Alex

[AllTheGearNoIdea]

Hi thanks

I think I may buy a SDG2042X at some point  as its up there on the sexy scale. But from what I can see its just a signal generators were as the picoscope has a signal genrator built in and I can use this along with its inputs to do frequency sweeps etc. The Pic scope also does have a lot of software tools built in for waveform analysis.   Good point regarding isolation  but  audio isolators are very cheap so I will probably build these into my test jig along with the dummy load anyway.  Thanks for the PC software idea that has legs. I would also like to be able to do wow and flutter.

Chris

[AllTheGearNoIdea]

Could I have some thoughts on this:

 I am however a bit concerned about the signal generator as I understand is -+3dB across 20Khz.  Now is it me or is that awful ?

 Chris

[MrW0lf]

Now looking at special 16bit audio PicoScope 4262:
https://www.picotech.com/oscilloscope/4262/picoscope-4262-specifications

Here are some FFT shots made with it:
https://www.eevblog.com/forum/testgear/fft-spectrum-analysis-reviewed/msg920786/#msg920786

This looks very good, much better than general purpose 14bit ones:
Crosstalk   > 50,000:1
Harmonic distortion   –95 dB typical @ 10 kHz, –1 dBfs input
SFDR   96 dB typical @ 10 kHz, –1 dBfs input
Noise   8.5 µV RMS (on most sensitive range)
Bandwidth flatness   [+0.25 dB, –3 dB] from DC to full bandwidth
Pulse response   < 1% overshoot on all ranges

Signal gen:
Frequency resolution   < 0.01 Hz
Voltage range   ±1 V (no load)
Amplitude and offset adjustment   100 µV steps (within overall ±1 V range)
Amplitude flatness   < 0.1 dB to 20 kHz, typical
DC accuracy   ±0.5% of full scale
SFDR   102 dB typical @ 10 kHz, -1 dBfs input

So signal gen is actually miles better than you suggested in first post (+-3dB). Scope is -3dB down full bw, which is 3 to 5MHz depending on range.

Also can do bode plot with phase:
https://bitbucket.org/hexamer/fra4picoscope/wiki/Home

So if your whole goal is audio then why not.

[Alex Nikitin]

I would also like to be able to do wow and flutter.

Chris

Hi Chris,

Actually, W&F measurements is one area where the audioTester is not good at all (as it tries to use a long FFT to analyse the signal, resulting in inaccurate and inconsistent results). However, there is a nice free and accurate utility - the WFGUI by Alex Freed, available in the Library section of my website. This utility also does logging and allows to record the demodulated signal into a WAV file, so you can do a spectrum analysis and find the components of W&F (very useful for finding the offending mechanical bits from the frequency readings). I have a hardware W&F meter, however now I always use the WFGUI, as the results are exactly the same and additional information is very useful (i.e. you can measure the speed drift and log W&F values for a period of time).

Cheers

Alex

[nctnico]

I'd go for a setup which can also measure the phase since that is also important for audio work. In general a good soundcard + PC software will offer very good value for money.

[b_force]

Use ARTA; http://www.artalabs.hr/
and a very decent soundcard will do.
If you also want to check loudspeaker systems, a soundcard with XLR and phantom power + a measuring microphone; dayton emm-6 (for example)

[dimkasta]

What are the lowest THD and noise limits that you want to measure?

I like my toys, but going with a sound card and ARTA is probably the best and fastest solution if you are on a budget and do not have time to spare playing/goofing around. Going below ~60dB THD is going to be an issue for most budget commercial generators/synthesizers, so you'll have a problem having a clean baseline.
Check the Nelson Pass area on diyaudio forum. You can find many examples of its use, and I think someone even posted a guide for ARTA.

I am currently using a DIY generator with my Rigol DS1052e and applying its bandpass filters to isolate harmonics and check their amplitude and phase while I fiddle with stages and biasing. FFT is not really usable for detail work on the Rigol. Just enough as a way to check for problematic areas that you can focus on. And I cannot really get a direct total THD and noise figure, but I do not really care about it. I am usually mostly interested in minimizing 2nd and 3rd harmonics and checking the sonic effect of different types, amounts, and combinations of harmonics.

[AllTheGearNoIdea]

It’s funny what actually kicked all this of is I spoke to the Technical guy at Picoscope this morning just to confirm the flatness of the signal generator output and he messaged me back saying it was +-3dB and hence I posted on here to ask what everyone thought. Ok huge face palm.
If I use two channels on the Picoscope I can also do the phase analyse.
I think I will probably give the Picoscope a try as it may prove more of an allrounder than the PC sound card option. But then again, I can just buy both and add to my growing test gear collection.
I like the other options the Picoscope scope has for rolling my own and perhaps writing some automated test scripts etc.
Thanks, everyone. No doubts I will be posting another video of me floundering around with some test gear I don’t know how to operate properly sometime very soon.
Cheers

Thanks dimkasta you posted as I did.   Actually I dont have a handle on the actual numbers but wanted to get something that would just work out the box. If i had the choice it would be some big box from HP but dont have the money for that.

[MrW0lf]

It’s funny what actually kicked all this of is I spoke to the Technical guy at Picoscope this morning just to confirm the flatness of the signal generator output and he messaged me back saying it was +-3dB and hence I posted on here to ask what everyone thought. Ok huge face palm.

Tell me about it... they are selling 2Mpts FFT scopes as 1Mpts, speccing 250MSa/s for 1GSa/s scope at Farnell and I had argument with some support person who tried to prove me that I cannot measure timings down to sampling jitter and therefore do not need sub-nanosecond timing skew setting
We here in Estonia have a say that translates roughly as "fast eat slow". Now Pico has to be real fast with getting stuff right or chinese will have them (along with Tek) for breakfast sooner or later...

[Messtechniker]

You can use a PC with a decent 96K or 192K sampling rate audio card and audioTester program. In my home lab I have a couple of RME Digi96 PST cards and uses these with audioTester regularly. You may need to make a proper front-end for the card, to accommodate the range of input voltages.

... and a proper output end for the card with a bit of uuuumpf -- say +20 dBu into 50 Ohm or so.
Inputs and outputs balanced if you require such. +26 dB output can be useful when measuring up audio transformers. The lesser ones produce beautifully bent curves 

I use instead of the RME Digi96 PST a no longer available M-Audio Audiophile 192 since the RME has only WXP 32 bit drivers which may be used with W7 32 bit as well. Don't know if they work with W10 32 bit. They definitely do not work with any W 64 bit software. The M-Audio 192 is more flexible here.

[AllTheGearNoIdea]

Hi Thanks for all the replies. I don’t want to go too far down this rabbit hole of technical specifications. I just want to be able to compare one amplifier to another and I’m not trying to get back to a traceable standard or anything. On the other and I don’t want something that cannot discriminate between 0.1 and 1% TDH.   I can afford a few hundred but not a few thousand etc.   My understand of audio equipment is not that great or what test and measurements are appropriate but you have to start somewhere.

Thanks to the website links lots of interresting stuff there

Many Thanks

[AllTheGearNoIdea]

https://www.picotech.com/oscilloscope/4262/picoscope-4262-specifications

This is what I look at

Thanks

[dimkasta]

If i had the choice it would be some big box from HP but dont have the money for that.

Stay away from those. They simply do not have the capacity to measure modern levels of THD

[macboy]

I have an HP 8903B. This can measure THD, S/N ratio, amplitude (flatness measurements with logarithmically stepped frequency points), DC offset, AC voltage down to sub-mV levels, and even power (estimated by voltage). The residual THD+N when looping its generator output to its input is < 0.0020% with 30 kHz BW, and about double that at 80 kHz. Floating differential inputs that can handle 300 V are a big bonus for testing amplifier output - you can skip the extra circuitry needed to feed such signals to a soundcard. There is software floating around out there to plot gain or THD vs frequency at various power levels, etc., but you need a GPIB interface for your computer to use that.  It has a "monitor" output, which is the signal after the notch filter, basically just noise+distortion with fundamental removed. You can run a FFT on this signal with a scope or soundcard to see relative levels of individual harmonics. In general I'd say the 8903B is good enough to measure the THD+N of just about any amplifier, but definitely not of line level stuff. You can pick one up for a few hundred $.  Panasonic makes similar sets that have nearly about an order of magnitude lower residual distortion, and cost $1k+.

I also have an M-Audio Audiophile 192 sound card which I plan to use for computer based measurements, but I haven't gotten around to setting that up yet. It should give very good results, as it has a very low THD+N floor and differential inputs.

[DaJMasta]

The mentioned siglent function gen has one of the lower THDs for bench signal generators, but there is really no comparison to the performance of a decent sound card.


If you absolutely need every parameter characterized and lots of input protection, you can spend tons of money and get an audio analyzer from AP or Prism Sound or something, but that's several grand and it doesn't sound like you even are considering that level of performance.

Going with a purpose-built oscilloscope also means you lose out on most of the wide dynamic range that a modern sound system has to offer - even high resolution 16 bit converters are notably worse than the industry standard 24 bit ADCs used for audio work, and the extra bandwidth means nothing if you're talking strictly audio levels.  Faster sampling rates can help responsiveness for a given FFT, but if your software is smart, you can actually do that without extra samples.

Knowing that, I think a PC based sound card will give you the best performance.  Something like the QA401 will give you high performance converters and a good software suite, though it lacks standard sound driver outputs so using regular sound software may be difficult, plus if you need phantom power or audio-specific connectors you are out of luck.  The route I went was to get a high quality recording interface that is just standard audio gear, but then you can use almost any PC based software you can find, and there are several good free ones.  I like Spectrum Lab and Arta for free tools, but there are many pay options and if you have Lab View or similar mathematical manipulation tools, you can code up whatever analysis and graphing tools you can desire.

Bottom line is that aside from the very highest end audio analyzers, you're going to get the lowest noise and THD from DACs and ADCs designed specifically for audio use in well designed interfaces, and having the flexibility that computer software affords you, you can do a lot with the cleaner signals coming in (FFT size limits be damned when you have several gigabytes of main memory to deal with).  It's easy to get a THD+N floor of lower than -100dB for only a couple hundred bucks in a sound card, and getting to even -95dB or so can be quite cheap.  On the high end, you can get to -110dB or better for probably only a little over $500.... none of the cheap(ish) 90s era test equipment can really touch that level of noise performance and it's well beyond the dynamic range of human hearing.  With older dedicated test equipment you get the built in functionality, input protection, and extra characterization, but you sacrifice the digital hardware limitations (processing power and memory size), the noise floor of modern audio equipment, and the flexibility of using whatever software you want.

[Loboscope]

For simple function tests and repair of my audio gear I take my scope (Rigol DS1104Z), but for more special and precise audio testing (Noise / THD / IM / ...) I use the professional software "hpw-works" together with a decent Sound Card/Interface (RME).
I already recommended hpw-works in this forum, but you will better have a look for yourself at the homepage: http://www.hpw-works.com/
It is also possible to download a test version of the software there and do a checkout, if it will fulfill your requirements.

[rstofer]

Could I have some thoughts on this:

 I am however a bit concerned about the signal generator as I understand is -+3dB across 20Khz.  Now is it me or is that awful ?

 Chris

If you are talking about the Siglent SDG 2042, you have slipped a decimal point.  The signal is flat within 0.3 dB from 0 to 100 MHz and 0.4 db from 100 to 120 MHz (depending, I suppose, on which model).

I have the SDG 2082 and it works quite well.

If you want automatic testing of Vout vs Vin over a range of frequencies, you might want to look at the Digilent Analog Discovery II and the 'Network' feature.

http://store.digilentinc.com/analog-discovery-2-100msps-usb-oscilloscope-logic-analyzer-and-variable-power-supply/

[ebclr]

Nice instrument, and ease to use

[AllTheGearNoIdea]

Well I have had a bit of a rethink I still like the picoscope but when I get the version that includes the signal generator and pay the VAT it going to cost more than £1,000 and I guess I just cannot justify that on a whim. Sorry apologies for wasting your time.  So I think the USB sound card could be a better option with a home brew dummy load and front end.  So with that in mind I have been looking at this. Would this be a good choice or any other suggestion. My thought are don't want to pay to much as no doubt I will blow up the input.

Creative Sound Blaster X-Fi HD USB Audio System withwww.amazon.co.uk

Link: http://amzn.eu/35BOWOP

I would probably be buying from amazon as I'm basically too lazy to shop around

Thanks again for all your good advice

[ruairi]

I would recommend getting something better than the Creative.  Even a cheaper focusrite unit for example.

Why do you expect to blow up the input?  I've owned and used 4 different USB audio interfaces over the last 10 years for measurement of gear and acoustics (from $200 to $2500) and I've never come close to damaging and input or output.

[DaJMasta]

I'd agree that damaging the input is probably tough even without the protection.  Extra is important if you're working in an industrial environment with high powered gear around or if you're trying to directly interface with the output of high powered amps, but that doesn't come up in most home use situations.

Also, just to be clear, when we say Sound Card, there's no reason it has to be a physical card.  I was sort of confused when I saw people using it, but any usb or firewire interface would qualify as one.  While there are some good units by Creative, they often cater towards gaming sound acceleration or other things, so they're rarely as good bang for buck performance as ones designed for audio recording or studio usage.

There are some decent interfaces available on amazon for around that price... but if I were you, I'd check out your local music shop that has used gear or check around local deal sites (like craigslist in the US).  You can get a nicer interface for the same price or maybe just save yourself some money; generation old studio grade products by RME, Motu, or others are probably within your price range if you figure used gear.  Since they are fairly durable devices and there are a number of people buying and upgrading, buying used can be a nice alternative.  Regardless, check up on the specs for the unit you're considering, and look around for user reviews or measurements - there are a lot of companies that don't give more than basic specifications, and if you're using it for measurement, you probably want a clear idea of where your numbers can be trusted to.

[AllTheGearNoIdea]

If you had not figured it out - I can be quite stupid.

Well i assumed a mistake in setting up the home brew input attenuators could be terminal if testing a 100 watt amp etc.