Exploring Visual Analyser (VA)

Hi, all -- just found this forum. A note on comparing software THD analyzers with analog analyzers like the HP 8903B, HP 339A, etc. The measurement floor of the 8903 series is about 0.0009% at 1kHz if in good cal, due to spuriae from the notch filter. The results for the HP 339A are similar and for the same reasons. Upgrades to opamps and other parts can significantly improve these analyzers, but using a software analyzer is so much easier if it works. If a software analyzer shows lower THD, especially measuring the 8903 or 339 oscillators, which have *very* low THD, then the software analyzer is more nearly correct.

The VA 2012 Beta link seems to be dead now. I've used ARTA regularly, but am going to look at the 2011 version of VA. If and when that gets to the bench, I'll post back here.
 
It would be good to see a comparison of a hardware bsed THD analyzer to the software versions.

Of course a quality soundcard should be used and the best amps amps installed if possible.
I will take a look at the site, But i think you have to sign up to the forum inoder to get to the latest bets versions.

I have been using VA11 and VA12 and sometimes they seem to be a bit flakey in Win7 and seems to be more stable favoring WinXP, But this is an intermittent issue so I can't make a solid judgement on this.
For the most part the work pretty good.
Wave Spectra is another great piece software that displays THD as well.
It is a faster running FFT and runs with no glitch's.

If you can get a hold REWv5 Beta 17 it works great as well but only the Beta versions have the THD function unlocked.

FWIW

jer :)

P.S. I guess it is down, I hope that it comes back sometime soon !!
 
Last edited:
Software THD analyzers are limited to the performance of the ADC used, as well as the residual of the source -- in the case of the EMU-0204 running at 24bits, for example, with the HP 339A as source, the 1kHz full-scale referenced THD will be in the 0.0002 - to 0.0004 area as shown by ARTA. Notching out the fundamental by even 20dB will then result in THD below 1 part per million. Almost no hardawre analyzers can get anywhere near that. Nevertheless, I still use an HP 339A for quick surveys and for most amp testing.

For more detail, see my web pages on modding Herath IG-18 oscillators and building/using an active Twin-T notch filter at Home -- article index
 
Those are some really neat articles !!
I may one day build a hardware version THD analyzer sometime with the new opamps that are now available.

TI's latest greatest audio opamp the OPA1664 has a rated THD of .00006%!!
This is 100 times better than what my current sound cards are capable of.

http://www.ti.com/lit/ds/sbos489/sbos489.pdf

One problem I have been having is getting VA to run in 24bit mode reliably.
Sometimes I can get it to work only for a little while and then it crashes.
Sometimes it won't run at all!!

Eventually, I will look into building the actual interface that was designed for it.

Most of the other programs are limited by their sample rate and the good ones that aren't limited cost a bit of money.

jer :)
 
Last edited:
You can use ARTA for free but not be able to save setups, etc. But in free mode nearly everything works, and works beautifully. It can handle 192kHz and 24 bits. After using it for a few days, I happily sent the registration fee -- I remember it as less than $120 US -- for the full program. As I mentioned, I use it with an EMU 0204.

I also like the QuantAsylum 400 system a lot. But it only works with it's own HW & SW as a complete package. ARTA will work with nearly any sound card/USB box.

What I would like is to have that functionality in a Macintosh app that runs 192kHz/24 bits.
 
Those are some really neat articles !!
I may one day build a hardware version THD analyzer sometime with the new opamps that are now available.

TI's latest greatest audio opamp the OPA1664 has a rated THD of .00006%!!
This is 100 times better than what my current sound cards are capable of.

http://www.ti.com/lit/ds/sbos489/sbos489.pdf

Very interesting. I'm a new convert to op amp rolling, having just purchased some single and dual op amps from Dexa as replacements for the stock op-amps in my Eastern Electric DAC.

Anyway, my real reason for posting here is that I am after some advice for assembling a probe to attach to the soundcard to perform voltage ripple tests using VA.

I tried last night by stripping down a cheap microphone with a 3.5 mm jack and soldering a couple of old voltmeter probes to the exposed wires (two in total), assuming it would be that simple. But the volt readings were extremely wild (calibration attempts made no difference), leading me to think that my DIY approach was simplistic.

Apologies in advance if this is a somewhat naive question! But I can't find any guidance in this area.

FWIW, I was trying to measure a regulated 5v supply, which may be too much - I don't know - but VA does offer V as well as mV as acceptable reading parameters so I assumed it (or, more accurately, the soundcard and the PC!) would be ok with it.

I am running VA 2011 on Windows 7. I doubt the OS is the problem, nevertheless I thought I would mention it.
 
...by stripping down a cheap microphone with a 3.5 mm jack and soldering a couple of old voltmeter probes to the exposed wires....

You need shielded cable and a good, shielded plug. NYS231BG REAN / Neutrik | Mouser Otherwise, RFI and mains hum will be picked up, rendering that great THD useless.

A real stereo sound card would be good. Then you will have two channels, too. They are expensive but much cheaper than a real scope. In fact, with VA, the soundcard is the scope, so to speak.
 
You need shielded cable and a good, shielded plug. NYS231BG REAN / Neutrik | Mouser Otherwise, RFI and mains hum will be picked up, rendering that great THD useless.

A real stereo sound card would be good. Then you will have two channels, too. They are expensive but much cheaper than a real scope. In fact, with VA, the soundcard is the scope, so to speak.

Many thanks for the sound advice.
 
Use a unity gain buffer on the input of your sound card.
Some if not all sound cards impedance varies with the input frequency and this can cause false reading especially if you are using a resistor divider between your probe and the input to the card.

I show such an interface here that I use to measure high voltages from step-up transformers and amplifier outputs as well,

A TEST JIG FOR FINDING ESL STEP-UP TRANSFORMER PARAMETERS

Below I have posted the schematics from that thread.

You can omit the dividers if you wish as long as the A.C. content of the signal is below about 1v to 2v peak and the D.C. voltage is within range of your opamps.
Your sound card will be already DC blocked by its internal DC blocking capacitor.

Some opamps are not unity gain stable in non-inverting configurations as well and may cause issues as I have had this problem occasionally.
I plan on revising this circuit with something like the LT1007 which is a high speed precision opamp and it is unity gain stable in non-inverting mode.

I used this same technique in my regulated HV supply as well here,

http://www.diyaudio.com/forums/plan...tor-insulation-mylar-coating.html#post2848194

This same technique can be found here in its simplest form without the divider resistors,

http://wb6dhw.com/RLC_Meter/RLC_Meter-sch.pdf

This is the link to the original website,

WB6DHW

I have used this interface a lot and with great success and I have found it is necessary to use the due to the fore-mentioned issues.

In this thread I show how well VA does with very low level signal buried in the noise floor compared to my scope and it basically puts the scope to shame in the particular test!!!!

http://www.esldiy.com/index.php?topic=176.msg310#msg310


Cheers!!
:cheers:

jer :)

P.S. Thank you xstreme, For the heads up on the latest version, I will be giving it a try shortly!! :)
 

Attachments

  • TRANSFORMER TESTER.jpg
    TRANSFORMER TESTER.jpg
    303.6 KB · Views: 425
  • Updated Test Jig Schematic.jpg
    Updated Test Jig Schematic.jpg
    68.5 KB · Views: 419
Last edited:
I just gave VA 4.2 2012 Beta a small test run and it seems to work much more smoothly than earlier versions.
I didn't give it a full test with tones yet, as I will do this later.

However, I could only get it to work in Win7. :(

I don't have any sound cards that will work in Win7 at this time except for what is on the motherboard.

I figured that I would give it a quick try before I wiped out my Win7 drive as I have switched every thing back to XP64 at the moment because I am tired of incompatibility issues I have with my older hardware and Multiplicity.

I guess I will have figure out a multi-boot system as I use linux as well and I have done it before, but it is a PIA to do sometimes.

Anyhow I did get VA4.2 2012 beta to lock up under one condition and is probably expected by doing a calibrate function using 0v and no signal was applied only using the noise floor of the onboard sound system.

Also I didn't have anything plugged into the input jack so they were probably shut off, as it does this in Win7, and is what could have caused it to get lost and lock up.
Let alone trying to calibrate to Zero.

I will give this another full test once I get my machine reconfigured with a Multiboot of the Three OS's that I normally use.

It sucks having to keep Win7 around as I just don't like it!!!

But this is not the only program I have run into with this problem of only working in Win7 these days. :(

Cheers !!!

jer :)
 
Last edited:
I guess I will have figure out a multi-boot system as I use linux as well and I have done it before, but it is a PIA to do sometimes.

With GRUB, it's easier to install Linux over an existing Windows installation. GRUB will automatically take over the BCD in Vista and higher, and the NBR in XP and lower. I had some driver issues with XP64, at any rate.

Anyhow I did get VA4.2 2012 beta to lock up under one condition and is probably expected by doing a calibrate function using 0v and no signal was applied only using the noise floor of the onboard sound system.

Also I didn't have anything plugged into the input jack so they were probably shut off, as it does this in Win7, and is what could have caused it to get lost and lock up.
Let alone trying to calibrate to Zero.

I don't think his program is designed to operate with no signal sample at all.

I would check to see what's going on in the bios; that everything is enabled or disabled that should be. When you get the sound card, I would disable the on-board, and enable the addon card.

Most onboard sound cards are lousy, and if you blow it up, it's gone forever.

Try the C-Media CMI8738. I'ts cheap and works perfectly with VA and W7. I prefer used cards because if they still work after all that time, they will probably last forever. One problem, however, may be corroded jacks on the card. I found colored ones 10 for $2.99.

Or you keep buying used ones for $5-$8 each until you find a perfect one. New, I would go with the ASUS Xonar DS as it is high quality and there is a fully-functional, Linux driver for it. It uses either of the PCI slot types (not PCIe), and is about $50. The DG is no good for Linux, as the driver is incomplete, last time I checked 1 year ago.

I also recommend an ASUS motherboard when you get around to it.

Dust buildup on the old mobo can cause "phantom" errors. These are not only more frustrating than finding an intermittent electrical connection, there are no solutions on the net for them (for there cannot be). In fact, I suspect many of these problems we see on the net (that have no good solutions) are in fact caused by phantom errors, or errors caused by mains spikes!

I noticed most people do not bother to acquire TVS protection.....nor does anyone clean their mobo...
 
Last edited:
I strongly urge you to use a USB sound device with 24-bit ADC and DAC, like the EMU 0204 or Quatasylum QA400, for example, if you are at all interested in high resolution analysis. Many of them will work in Win7 as well as XP.

I really can't see much diff between XP and 7 except for the underlying changes and driver changes, but then I'm a Mac user mostly, so the changes wouldn't stand out for me. But Win7 does let me easily dual-boot my hackintosh.
 
I was actually able to get on the forum today and re-download the files.
And I found that the 32 bit version does work in XP64 but the 64 bit version does not.

The card that I am using at the moment is a Gina24 and it has an ASIO driver.
It works great but I was not able to get the software into 24 bit mode as in past tests in fact this version stayed locked in 16bit mode.

I have been studying more about the USB interfaces and I will attempted such an interface as described on VA's home page only I will use some other A/D and D/A chips from either Analog Devices or LT.

I can't get the Gina24 or Creative's X-FI drivers to work in Win7 on my new machine because they are not recognized by the UEFI Bios that it has, as the cards predate it by many years.

I have read of others whom did get them to work but so far I have been unsuccessful. :(

I did however get VA to run the Gina24 in 192Khz Stable this time (maybe, I have not verified it yet) as VA would lock up with the earlier versions.
This time it ran without locking up even though it was locked 16 bit mode.
16 bit mode is fine for now until I get a proper interface built.

Not only do I want 24 bit precision, But I would like to be able to measure DC voltages as well and you can't do that with most sound cards.

At least I did get the new 32bit version working in XP64 for now, and, just having the software stable is 100% better than it was before.

Cheers !!!

jer ;)
 
Last edited:
I couldn't verify it to be running at 192Khz with my hardware although the software could be set that fast, I could only verify a 96Khz input sample rate and at best it would sample up to 44.1Khz flat.

However that is what my current card is specified at, 96Khz.

When I put it in 24bit mode the software stays running but I get no signal displayed from the card, setting it back into 16 bit mode the signal re-appears and is displayed correctly.

Normally the earlier versions would just lock up when any changes were made with those parameters on my system.

As I mentioned, I will do more testing on my newer machine with Win7 once I get it back running again as the on board sound system will do 192Khz, and, after I build my interface buffers into a permanent enclosure as well and start testing stuff again.

jer :)
 
OMG!!

I just tried out the latest VA2014 and so far it looks really good!!!
It still doesn't pick up data from my Gina24 sound card in 24 bit mode but at least it doesn't crash when I set it to it.
Sometimes it will stop when switching modes but it does recover easily.

Here are some screen shots of a simple un-calibrated loopback test, The abnormal Green FFT trace may be due to a crappy cable.

I can't wait to start putting this new version through its paces!!!

You can find it here,

Download

Nice Job, and, Cheers to the developer!!!! ;)

:cheers:

jer :)

P.S. You have to use the 32bit version if you are using XP64. ;)
 

Attachments

  • VA2014 uncal a.jpg
    VA2014 uncal a.jpg
    139.8 KB · Views: 383
  • VA2014 uncal b.jpg
    VA2014 uncal b.jpg
    131.1 KB · Views: 373
  • VA2014 uncal c.jpg
    VA2014 uncal c.jpg
    116.3 KB · Views: 367