Author Topic: Audio distortion measurement techniques  (Read 14826 times)

0 Members and 1 Guest are viewing this topic.

Offline ruairi

  • Frequent Contributor
  • **
  • Posts: 302
  • Country: us
Re: Audio distortion measurement techniques
« Reply #25 on: September 24, 2016, 06:41:29 pm »
I'm not sure feeding sonic waves into a computer for analysis is such a good idea. I have a RIGOL dg4102 digital oscillator claiming 0.1% distortion in its audio range, when I measure it I do get 0.1% distorton but the signal after the notch is full of spikes in the time domain.

I can't tell if you are joking here?  Needless to say that almost every record you've heard in the last 20 years has been created by feeding "sonic waves into a computer".  A reasonable audio interface will have low enough distortion to do meaningful work.  Yes it will add it's own distortion, and has frequency response limits but it is way more useful than a cheap old hardware unit.

Quote
Nobody was claiming thd figures to be the holy grail however it can be used as a figure of merit, also when checking the spec. of any commercial amp, thd figures are always given but no reference to individual harmonics.

The specs provided on many audio products are either optimistic, or manipulated in ways to make them look better, for my purposes as a working audio professional for 20+ years they are effectively meaningless.  Sure, they can be used for very broad stroke analysis but that's about all.

We don't need to know the levels at each harmonic but trends can be very telling - more odd than even, only low order harmonics, lots of higher order harmonics.  These can be correlated to actual listening experiences in a way that THD+N often cannot.

Quote
What do the individual harmonics tell you?

See above.

The DIYaudio.com forum is a really useful resource for affordable distortion measurement techniques and information.
 

Offline commieTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: gb
Re: Audio distortion measurement techniques
« Reply #26 on: September 24, 2016, 08:20:39 pm »
Hi

I hear this a lot, "the only good stuff is not where I am". Oddly enough, when I go shopping, that's *exactly* what I find. All the good stuff at low prices is being sold in the UK and it will cost me a fortune to get it..... time to spend some more time shopping. It's not a function of location as much as *time*. If you don't have the time, you will pay somebody else to have it for you right now.

Let's go through the numbers:

8903B for $150 at auction.
VAT at 20% (or whatever) is $30
Shipping is $70 to $120 depending on where it's from
Customs processing is $30
Duty at 20% is $30

Price delivered is $260 to $310. Still a *long* way away from the $1200 you pay somebody to have it right now.

Bob

Hi Bob,

Yep, I agree with what you have said, if I wait something viable should come my way.

Cheers and thanks for talking to me.
Commie
 
 

Offline deadlylover

  • Frequent Contributor
  • **
  • Posts: 314
  • Country: au
Re: Audio distortion measurement techniques
« Reply #27 on: September 25, 2016, 03:50:34 am »
I recommend just making do with a cheapish sound card or audio interface until you find a good deal on used test equipment (or building your own).

The 8903B is a bit...obsolete, so unless you just want one for fun (who doesn't? =P), I would steer clear unless you find one very cheap. Its residual of ~-95dB THD+N @ 1kHz is in the range of a sound card, so unless that's good enough for your needs then you'll quickly run into its limitations in modern audio.

If you're a bit daring and willing to purchase overseas, look at Yahoo Japan auctions for an old Panasonic or ShibaSoku analyser. IMO they are the only 'old dogs' that hold up today in terms of residual distortion and noise performance. You can sometimes find them for around JPY30,000 before shipping, maybe add 20k yen on top for that, they're usually a bit under 20kg or so.

My Panasonic VP-7722A was around $500 delivered and its residual is ~0.00023% (-112dB) THD+N @ 1kHz, 30kHz BW at a couple of volts or so.
And if that's not quite good enough for you, the 2 decade old ShibaSoku 725B distortion analyser (no built in osc.) can do -116dB THD+N stock and about -120dB with a few easy mods, mainly op amp swaps and a ground rerouting. That's near flagship Audio Precision performance by the way.
 

Offline Pjotr

  • Frequent Contributor
  • **
  • Posts: 461
  • Country: nl
Re: Audio distortion measurement techniques
« Reply #28 on: September 25, 2016, 10:15:52 am »
It is not clear OP is up to: Designing and building audio gear/amplifiers, marketing amplifiers or just toying around?

For designing amplifiers you actually don't need a distortion meter at all. Although it helps to see if you make progress in areas you are interested :) Nowadays THD levels between 1 ppm and 10 ppm for audio power amps are not a miracle and a sensitive spectrum analyser set-up is more of help --> AP etc. Although with a good PC sound card you can do also a lot.

If OP is in the vacuum tube area where distortion levels between 0.1% and 10% are common place, then an ancient distortion meter from steel, alu and moving coil meters can do a good job. If he is in the solid state league of 1 - 10 ppm THD, better look for something more modern. It is amazing what a modern good soundcard can do then.
« Last Edit: September 25, 2016, 10:38:07 am by Pjotr »
 

Offline calexanian

  • Super Contributor
  • ***
  • Posts: 1881
  • Country: us
    • Alex-Tronix
Re: Audio distortion measurement techniques
« Reply #29 on: September 25, 2016, 11:30:27 pm »
I got my HP 334A for next to nothing at a ham radio swap meet. Audio fools sell them off cheap because they generally don't know how to use them. The principal is you start with a known distortion free sine wave. The analyzer has a very very good filter in it you tube to remove the direct fundamental sine wave. Whatever signal is left therefore must be distortion and that remainder is measured. Pretty simple stuff, just very hard to implement in a DIY fashion. You can pick them up for very little because they are rather tedious to tune in to work and most people just think they are broken when in reality they are pretty much bullet proof.
Charles Alexanian
Alex-Tronix Control Systems
 
The following users thanked this post: Electro Detective

Offline Conrad Hoffman

  • Super Contributor
  • ***
  • Posts: 1928
  • Country: us
    • The Messy Basement
Re: Audio distortion measurement techniques
« Reply #30 on: September 26, 2016, 12:12:05 am »
First, read the Cordell article on building your own THD meter here- http://www.cordellaudio.com/papers/build_a_thd_analyzer.shtml
I built mine many years ago and it has a mid-band floor of about 0.00015%. Next, get a copy of Doug Self's power amplifier design book. He shows how to routinely build amplifiers at the 0.001% level. You might also like this thread on the 8903- http://www.diyaudio.com/forums/equipment-tools/295895-hp-8903a-audio-analyzer-mods-upgrades.html I use one because it's good enough for most things, but especially because it can be run over GPIB, which speeds measurements dramatically. I usually discover problems or areas for improvement just because it's easier to run routine tests you might not otherwise take the time to do.
 
The following users thanked this post: 3db

Offline Gary.M

  • Regular Contributor
  • *
  • Posts: 137
  • Country: nz
Re: Audio distortion measurement techniques
« Reply #31 on: September 26, 2016, 07:54:07 am »
First, read the Cordell article on building your own THD meter here- http://www.cordellaudio.com/papers/build_a_thd_analyzer.shtml
I built mine many years ago and it has a mid-band floor of about 0.00015%. Next, get a copy of Doug Self's power amplifier design book. He shows how to routinely build amplifiers at the 0.001% level. You might also like this thread on the 8903- http://www.diyaudio.com/forums/equipment-tools/295895-hp-8903a-audio-analyzer-mods-upgrades.html I use one because it's good enough for most things, but especially because it can be run over GPIB, which speeds measurements dramatically. I usually discover problems or areas for improvement just because it's easier to run routine tests you might not otherwise take the time to do.
Seriously guys, I already gave you the answer. Google QA401. It's a highly competent all in one Analyzer at an affordable price.

Sent from my x600 using Tapatalk

 

Offline Pjotr

  • Frequent Contributor
  • **
  • Posts: 461
  • Country: nl
Re: Audio distortion measurement techniques
« Reply #32 on: September 26, 2016, 09:26:46 am »

Seriously guys, I already gave you the answer. Google QA401. It's a highly competent all in one Analyzer at an affordable price.

Sent from my x600 using Tapatalk

Looks like a hobbyist project USB sound card. But at 2x - 4x the price of a comparable off the shelf card.
 

Offline uncle_bob

  • Supporter
  • ****
  • Posts: 2441
  • Country: us
Re: Audio distortion measurement techniques
« Reply #33 on: September 26, 2016, 02:38:28 pm »

Seriously guys, I already gave you the answer. Google QA401. It's a highly competent all in one Analyzer at an affordable price.

Sent from my x600 using Tapatalk

Looks like a hobbyist project USB sound card. But at 2x - 4x the price of a comparable off the shelf card.

Hi

Make no mistake, it's not a hobby card. The converter on the board and it's support electronics (isolation etc) are first rate. My main knock on the product is that they don't really let you get at the data stream from the device directly. Simply put - you can't easily use it like a sound card. That limits it's utility in really nutty applications (say you want the even and odd harmonics summed up independently).  I got very close to buying one and ultimately decided that software wasn't going to work for me.

Bob
 

Offline Conrad Hoffman

  • Super Contributor
  • ***
  • Posts: 1928
  • Country: us
    • The Messy Basement
Re: Audio distortion measurement techniques
« Reply #34 on: September 26, 2016, 03:39:23 pm »
Similar results can be had nearly for free with a decent sound card and a copy of Visual Analyser, albeit with compromises, but good enough for the majority of needs. The point of reading the old articles is to get a better understanding of the measurement floors and how these things have been done previously. The Self book shows you how to build amps an order of magnitude better than in the past, as the OP doesn't seem to realize just how good one can do, or that measurements at that level can be done with a great degree of confidence. Seriously!
 

Offline Pjotr

  • Frequent Contributor
  • **
  • Posts: 461
  • Country: nl
Re: Audio distortion measurement techniques
« Reply #35 on: September 26, 2016, 03:46:03 pm »
Hi Bob,

look at post#9 on the previous page for what I achieved with an off the shelf M-Audio Audiophile192 card (also with AKM DA/AD). I have modded it with better op amps and distortion became 6 db lower than the picture in post#9 (2ppm). It is similar to the QA401. The M-Audio card was cleared last year for under 80 bucks. Such low distortion is only that low a specific signal level. The QA401 will not differ to that.

If you are at that distortion levels and want to measure lower you can't do without a really low distortion/noise separate sine generator and accompanying notch filter etc. That is not included with the QA401.
 

Offline Messtechniker

  • Frequent Contributor
  • **
  • Posts: 774
  • Country: de
  • Old analog audio hand - No voodoo.
Re: Audio distortion measurement techniques
« Reply #36 on: September 26, 2016, 04:50:27 pm »
Hi Pjotr,

would you care to share details of your Audiophile 192 mod?
There are many descriptions of Audiophile 192 mods. But mostly without
giving any measured data, like you do. I've been planning to modify the two Audiophile 192s
I have been using for measurement purposes for quite a while now.
I expect any Audiophile 192 mod to have only an influence on the measurements
with no audible impact at all.

Yours Messtechniker
Agilent 34465A, Siglent SDG 2042X, Hameg HMO1022, R&S HMC 8043, Peaktech 2025A, Voltcraft VC 940, M-Audio Audiophile 192, R&S Psophometer UPGR, 3 Transistor Testers, DL4JAL Transistor Curve Tracer, UT622E LCR meter
 

Offline Pjotr

  • Frequent Contributor
  • **
  • Posts: 461
  • Country: nl
Re: Audio distortion measurement techniques
« Reply #37 on: September 26, 2016, 07:07:07 pm »
Hi Pjotr,

would you care to share details of your Audiophile 192 mod?
There are many descriptions of Audiophile 192 mods. But mostly without
giving any measured data, like you do. I've been planning to modify the two Audiophile 192s
I have been using for measurement purposes for quite a while now.
I expect any Audiophile 192 mod to have only an influence on the measurements
with no audible impact at all.

Yours Messtechniker

Hi,

I only replaced all 5532A op-amps by LM4562 op-amps. Nothing else. IMHO replacing electrolytic coupling caps with film ones doesn't have any use. There is virtually no AC voltage across them and almost no current through, so why should they distort?

Your expectation about the sound is right, sonically I couldn't notice any difference. I doubt this mod is worth the effort, only 6 dB measurement improvement. IMO better use this really good op-amps to build an ultra low distortion spot sine generator. With a good notch filter and amplifying the residual distortion components improves measurement floor orders better.
« Last Edit: September 26, 2016, 07:13:09 pm by Pjotr »
 

Offline Messtechniker

  • Frequent Contributor
  • **
  • Posts: 774
  • Country: de
  • Old analog audio hand - No voodoo.
Re: Audio distortion measurement techniques
« Reply #38 on: September 26, 2016, 08:22:41 pm »
Hi,
thank you for your hint. Just had a look at my Audiophile192k boards only to
find a grand total of 11 NE 5532 SOICs on each board. Since I have to
date never soldered any SOIC chips, I will for the moment shelve this
project and dig it out again should the need arise - which I doubt.
Here once more the law of diminishing returns applies.

Yours Messtechniker

 
Agilent 34465A, Siglent SDG 2042X, Hameg HMO1022, R&S HMC 8043, Peaktech 2025A, Voltcraft VC 940, M-Audio Audiophile 192, R&S Psophometer UPGR, 3 Transistor Testers, DL4JAL Transistor Curve Tracer, UT622E LCR meter
 

Offline echu

  • Newbie
  • Posts: 7
  • Country: tw
Re: Audio distortion measurement techniques
« Reply #39 on: September 26, 2016, 10:01:09 pm »
Such low distortion is only that low a specific signal level. The QA401 will not differ to that.

You can plug the QA401 directly into a 100W class D amp output and not worry about it blowing up. At all. It's fully isolated from the PC. No electrical path between the audio section and the PC interface.

Sounds cards were a reasonable starting point, but there's much better out there today...

http://www.quantasylum.com/content/Home/tabid/40/Post/283/First-Look-Evaluating-Class-D-Power-Amps-with-the-QA401
 

Offline Pjotr

  • Frequent Contributor
  • **
  • Posts: 461
  • Country: nl
Re: Audio distortion measurement techniques
« Reply #40 on: September 27, 2016, 10:09:09 am »

You can plug the QA401 directly into a 100W class D amp output and not worry about it blowing up. At all. It's fully isolated from the PC. No electrical path between the audio section and the PC interface.

Duhhh.........??? PC's that blow the rest of your equipment are likely to blow the user as well. Trash them immediately! PC's have to be grounded to a safety ground anyway. Although it is often ignored, even in a lab environment.

B.t.w. the M-Audio card at hand has differential inputs and outputs. No grounding issues at all nor mains related noise. With those low noise levels a pitfall can be mains induced noise that comes from power cords coupling inductively into signal cables.  Keeping them carefully separate helps a lot.
« Last Edit: September 27, 2016, 11:03:14 am by Pjotr »
 

Offline echu

  • Newbie
  • Posts: 7
  • Country: tw
Re: Audio distortion measurement techniques
« Reply #41 on: September 28, 2016, 06:31:22 am »
Quote
PC's have to be grounded to a safety ground anyway. Although it is often ignored, even in a lab environment.

B.t.w. the M-Audio card at hand has differential inputs and outputs. No grounding issues at all nor mains related noise. With those low noise levels a pitfall can be mains induced noise that comes from power cords coupling inductively into signal cables.  Keeping them carefully separate helps a lot.

M-Audio has differential IO, but the IO isn't isolated from the PC. There's a DC path directly from the audio IO to the PC ground. You really want and need the isolation.

And how would the M-Audio do if you connected it to a class D push/pull stage being driven at +/-40V? Very likely, yes, that's right, smoke. And the lack of isolation means the 40V could kill the front end, jump to the ADC, then to the USB data part, and then into your PC. That is why you want isolation. Isolation is a 1000KV barrier between your DUT and your PC

As I said, the world has moved on from soundcards for audio analysis...

---
"I survived Typhoon MEGI and all I got was this shirt"

 

Offline Gary.M

  • Regular Contributor
  • *
  • Posts: 137
  • Country: nz
Re: Audio distortion measurement techniques
« Reply #42 on: September 28, 2016, 07:12:37 am »
Hi Bob,

look at post#9 on the previous page for what I achieved with an off the shelf M-Audio Audiophile192 card (also with AKM DA/AD). I have modded it with better op amps and distortion became 6 db lower than the picture in post#9 (2ppm). It is similar to the QA401. The M-Audio card was cleared last year for under 80 bucks. Such low distortion is only that low a specific signal level. The QA401 will not differ to that.

If you are at that distortion levels and want to measure lower you can't do without a really low distortion/noise separate sine generator and accompanying notch filter etc. That is not included with the QA401.
QA401 includes a generator. Google up some loop back tests on it. Pretty good noise floor and very clean spectrum.

Sent from my x600 using Tapatalk

 

Offline Gary.M

  • Regular Contributor
  • *
  • Posts: 137
  • Country: nz
Re: Audio distortion measurement techniques
« Reply #43 on: September 28, 2016, 07:14:56 am »

You can plug the QA401 directly into a 100W class D amp output and not worry about it blowing up. At all. It's fully isolated from the PC. No electrical path between the audio section and the PC interface.

Duhhh.........??? PC's that blow the rest of your equipment are likely to blow the user as well. Trash them immediately! PC's have to be grounded to a safety ground anyway. Although it is often ignored, even in a lab environment.

B.t.w. the M-Audio card at hand has differential inputs and outputs. No grounding issues at all nor mains related noise. With those low noise levels a pitfall can be mains induced noise that comes from power cords coupling inductively into signal cables.  Keeping them carefully separate helps a lot.
You guys must be very short of a meagre few dollars if you need to cobble together sound card solutions when there is such a good ready made option out there. The QA401 is targeted at factory test on production lines.

Sent from my x600 using Tapatalk

 

Offline Messtechniker

  • Frequent Contributor
  • **
  • Posts: 774
  • Country: de
  • Old analog audio hand - No voodoo.
Re: Audio distortion measurement techniques
« Reply #44 on: September 28, 2016, 12:37:00 pm »

You guys must be very short of a meagre few dollars if you need to cobble together sound card solutions ...
[/quote]

Not really, because the sound card solution is more than good enough for aligning tape machines and measurements on audio valve amplifier equipment. Am using suitable input :10, x1 and x100 and output x1 and x10 buffer amps to prevent damage to the sound card. As a rule of the thumb the measuring equipment parameters should be better by a factor of 10 compared to the data expected form the unit under test This is true for the above uses cases.

Yours Messtechniker.
Agilent 34465A, Siglent SDG 2042X, Hameg HMO1022, R&S HMC 8043, Peaktech 2025A, Voltcraft VC 940, M-Audio Audiophile 192, R&S Psophometer UPGR, 3 Transistor Testers, DL4JAL Transistor Curve Tracer, UT622E LCR meter
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 6454
  • Country: hr
Re: Audio distortion measurement techniques
« Reply #45 on: September 28, 2016, 12:57:09 pm »
I think he want to say that when you put together price of good external audio card and handmade atenuator box and tek into account time needed to do it right,  you will come close to 400 € which gives you fully tested and specified solution that woks out of the box.  If you have time and like doing it yourself then go for it.. If you just need something to work right away 400€ is not bad..
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Audio distortion measurement techniques
« Reply #46 on: September 29, 2016, 01:27:37 am »
Quote
PC's have to be grounded to a safety ground anyway. Although it is often ignored, even in a lab environment.

B.t.w. the M-Audio card at hand has differential inputs and outputs. No grounding issues at all nor mains related noise. With those low noise levels a pitfall can be mains induced noise that comes from power cords coupling inductively into signal cables.  Keeping them carefully separate helps a lot.

M-Audio has differential IO, but the IO isn't isolated from the PC. There's a DC path directly from the audio IO to the PC ground. You really want and need the isolation.

And how would the M-Audio do if you connected it to a class D push/pull stage being driven at +/-40V? Very likely, yes, that's right, smoke. And the lack of isolation means the 40V could kill the front end, jump to the ADC, then to the USB data part, and then into your PC. That is why you want isolation. Isolation is a 1000KV barrier between your DUT and your PC

As I said, the world has moved on from soundcards for audio analysis...

---
"I survived Typhoon MEGI and all I got was this shirt"

  Sorry, how many KVs? I take that as a typo...

  QA has some good stuff but there are a lot of things that can be done with just a soundcard, a lot! I've used my soundcards to do the task many times, of course you want to be careful with what you connect, the inputs are rated for audio gear, line level at most, not for power amps driving them directly. That being said, you don't want to blow your soundcard or your computer, so you double check before connecting, DC, ground currents/loops, etc. I guess that should be standard procedure anyway, poke with a :-DMM before running a full analysis and wandering specs.

JS
If I don't know how it works, I prefer not to turn it on.
 

Offline technotronix

  • Regular Contributor
  • *
  • Posts: 210
  • Country: us
    • PCB Assembly
Re: Audio distortion measurement techniques
« Reply #47 on: September 29, 2016, 07:51:29 am »
You need a source for a signal that has a low distortion and then see the difference between signal and output. May be it is not 100% accurate.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf