Usable distortion analyser for Pre-/Amps

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

I've just searched and found several threads on this topic...lots of links
to free SW-packages and surplus distortion analysers, like the HP33x.

But I wonder if these are really usable for audio electronics? Most of
the SW-packages I found are basically designed for loudspeaker design
with measurement of electronics only as a goodie.

I wonder if a resolution for THD of 0.01% is really usable for SS-amps
and esp. pre-amps where you can (more or less easily) reach THD
values of <0.05% over the complete audible range...? :confused:

Are there any recommended computer based measurement solutions,
really *usable* for serious diy-electronics, not necessarily free but also
not in the price region of an AP-Analyser??

Regards,
Michael.
 
Just another Moderator
Joined 2003
Paid Member
rmaa 5.2

rmaa 5.2 gets my thumbs up..... Not that I'm experienced in this area (in general) but I've been playing with it, and as far as I can tell it is a usefull tool, and I think if you have a good enough soundcard the resolution is probably not too bad. (at least for testing power amps).

here are links to tests I have done with RMAA. Note that the actual measurements are much higher resolution than these html reports would indicate, with full zoom capabilities.

1st loopback test on my sound card to get a base level to work with.

http://home.swiftdsl.com.au/~tonywww/rmaa/audigy II ZS loopback.htm

second burned the test tones to cd and tested my DVD player, to my surprise the distortion figures were lower, which means that the DVD player does a better job of generating the test signal than the sound card. Note that there is something very odd going on with the crosstalk on one channel, did this when I tested the computers cdrom drive too, so suspect something went screwy with burning the cd.

http://home.swiftdsl.com.au/~tonywww/rmaa/marantz DV18 44_1Khz.htm

finally a test of my amp which you can see has "some issues" !!!!

http://home.swiftdsl.com.au/~tonywww/rmaa/series_200_dv18_approx_25W.htm

RMAA can be got here: http://audio.rightmark.org/ Oh goody 5.3 has come out and apparently the audigy II ZS performance is even better still!!! Time for some more playing me thinks :devilr:

Tony.
 
Just another Moderator
Joined 2003
Paid Member
Oh well...... I'ts been usefull for me in showing up gross problems in my amp :)

Not 100% sure what you mean (as I said I'm not really experienced in this stuff). Are you talking about the combining of Noise + THD results?

edit: actually I just went back and looked at the results, and the A-weighted results are worse than the other results, which I do remember confused me at the time, since I allways thought that A-weighted results generally looked better (again my naivety may be showing here).

I guess the other thing I would say is a limitation with RMAA, is that RMAA (for the HTML reports) only chooses one channel for doing it's "analysys" I think it's better looking at the "raw" graphs.

Tony.
 
With a good soundcard and a 10 dollar oscilator I have achieved THD resolutions down to 0.00165 % about -97dB. I think that's about as low as you can get with a soundcard, allthough I've heard someone reaching -105dB THD. I use a Creative Audigy 2 ZS soundcard. The strange thing is that when I use RMAA 5.2 for 24bit support measurements, I cannot improve behond the above number... 16bit or 24bit give the same results both in THD and dynamic range.. hmmmmm makes me think..


Regards,
Thijs

PS

allmost forgot, the software... I think it's been said before:
RMAA 5.2,
DAZYWEB LABORATORIES SA-3002 V1.33,
spectralab ver 4.32.17,
http://www.audiotester.de/ .... etc
 
Hi,

I knew someone would post a sceptical remark... ;) I was sceptical myself too at first...

I haven't done any comparison against a calibrated device. I have compared the various software solutions againstb each other and they give the exact same results. I trust the mathmatics is done correct by the programmers, and that my AMD CPU doesn't produce invalid numbers and that the underlying nummerical-algorithmes are correct.

So once the bits are generated, I have nothing to worrie about... It's only in the ADC process that you might get worried.. the quality of the ADC is the bottleneck. But dittering and averaging the signal, together with carefull signal amplitude adjustment makes it possible to achieve such resolutions..beyond 16bit

Using a good notch filter, one should be able to measure even lower THD, only limited by the oscilators residual THD..

Goodluck,
Thijs
 
Hi,

Found Audiotester one of the most useful progs for doing such measurements. Although it is not free it is not that expensive. What the residual distortion is, can easily be measured by connecting input and output of the soundcard with a short cable. The distortion you measure then is the combined residual distortion of the signal generator and recording device. It is just what it is: the measurement floor of your test set-up.

Looking at the harmonic profile tells you a lot more than THD and/or THD+Noise figures alone. And for this a soundcard+software tells usually a lot more than an ancient second hand standalone distortion meter.

Cheers ;)
 
I have been looking in measuring with "soundcard solutions" for quite some time. After a few internet searches I got confused. So I wrote a document just to order my thoughts. There are a lot of products, software packages for different audio purposes. So by categorising a lot them the picture for me became clear.

What I realy want is a product which gives the posibilities oscilloscope, FFT and wave-generator with a very high precision/ resolution on my PC.

So basicly that is a hardwarebox and some software. Software solutions are available from good freeware to expensive professional.

The most important is the hardware. The old schoolbooks say that the quality of your measuringequipment should be a class better as the things you would like to meausure.

Looking at our "Audio" needs most soundcards specs are to close to be comfortable to the values we like to measure. If you want to have a 100 dB SNR and a DC-100kHz range the number of hardware options are limited. Unfortunately they are absolutely not affordable.:bigeyes:

However 24bit 192kHz AD converters become more and more available. So looking at professional "soundcards" and recording solutions might give a solution in the near future.

Personaly I find a product like the ESI Quatafire most promising(even payable!). I understood the Quatafire must become avavailable real soon Unfortunately they tell that already for more than half a year for their firewire range of products...
:bawling:

So yes I believe a solution is near!:nod:
(and I can finish that document)

Ward
 
Hi,

Attached a loop-through result of what I archived with my (not that expensive) audio card. Averaging was 100 times. THD figure at 16 bits is –96db or 0.0015%, at 24 bits –111 dB or 0.0003%. More than sufficient for DIY projects IMHO.

As far as I can figure out, Audiotester takes into account the first 9 harmonics to calculate the THD number, but I am not sure. At 16 bits there are relatively large harmonics above the 9 th harmonic. So the THD number calculated at 16 bits is doubtful. Dunno with other software but the limit is the measurement bandwidth (< ½ Fs).

For THD+Noise it takes anything compared to the fundamental, so this will be more trustful I think.

Anyway the distortion profile obtained at 24 bits is very useful to judge the qualities of DIY efforts.

Cheers ;)
 

Attachments

  • distortion.gif
    distortion.gif
    9 KB · Views: 670
tschrama said:
Hi Pjotr,

Those are excelent results! What soundcard did you use for that loop-through test?

Regards,
Thijs

Look at the top of the picture :D

Found that at Fs = 96 kHz distortion is worse (doubles). Also at higher levels than –6dB FS distortion becomes worse. I used a 10-cm loop-through cable. At long cable runs a lot of rubbish can enter the audio cables (i.e. from the mains cord), spoiling the results. Something to take care of.

Cheers ;)
 
Hi Pjotr,

I that 24bit -111dB number an absolute number or relative to -10dB. That would mean that the 5th harmonic level is -121dB. Your picture shouws the first harmonic somewhere between -110 and -120dB, not -121dB. Well, upon further inspedction, the red line is difficult to see but seems indeed to be in the -121dB ...

Thanks for the soundcard name.. looked it up.. on sale for 149.95$.. that nice indeed for such a high quality THD analyser...:D

Regards,
Thijs
 
IVX said:
hi,
maybe try to compensate own sound card THD ? I has measurement my class D amp through motherboard AC97 gotten THD are 0.06%, loopback THD are 0.04%(@50hz-0.04%;@1khz-0.01%), and now through audigi i have THD-0.02%. I'll try to compensate FFT's in MsEXEL. What you think about it?


RMAA has a "susbstract results" option, but I haven't tested it
 
Just another Moderator
Joined 2003
Paid Member
Audio Tester V2

Hi Pjotr,

I'd looked at Audio Tester V1.4 before and really didn't like it. Just downloaded V2 and it looks like it's much improved, both in functionality and UI wise. Though it looks like it's got a bit of a learning curve compared to RMAA.

Time to do some more playing me thinks :)

Tony.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.