Difference between left and right channel (QA403)

I recently received my unit and am in the process of learning all the features.
One thing bothers me though, the loopback measurements from the right channel are worse than the left channel. THD is worse by about 4-5 dB.
I have determined that this is the fault of the generator. Both inputs perform equal.

Is this normal? Or is something wrong with my unit.

an example measurement below.


Hi, my QA403 that I have been using for almost a year also has a difference between left and right channel. This difference is normal for the QA403 and is within specification.

this is my QA403 in the same loopback configuration as you

Thank you! That’s what I thought too, but it has been bothering me…

1 Like

My right channel is significantly worse than the left and in line with your results. It seems to be an acknowledged and unfortunate issue with the QA403.

Since the left generator channel has better performance, would it be a good idea to use a bnc y connector and feed both DUT channels from the left generator channel?

The numbers are so small, you really only know there’s a difference in loop back. Once a real DUT is in the loop, it’s academic.

The “problem” if you can call it that, is the QA as a system is so highly resolving, any tiny differences stand out.

Hi @yegolebi, building on what @restorer-john notes, you first need to decide what level of performance you need for your DUT.

Looking at your first plot, your 50 Hz seems a bit high and might be related to the leakiness of your BNCs and/or the position of the QA403 relative to powerline sources on your bench. The 50 Hz probably isn’t impacting your THDN (and doesn’t change your THD, of course), but you might try to see if re-orienting things on your bench and/or different BNCs can clean up the powerline. This will help for other measurements. The 50 Hz in your plot is -120 dBV=1uVrms, so it’s extremely small. But it’s also responsible for the bump at 100 Hz, 150, etc.

Shahriar at The Signal Path spends a bit on time looking the left/right discrepancy THD numbers at the link below. We have data on all the units shipped, and there is indeed a shift between the left/right. And since the left/right use the exact same signal paths, the conclusion is that the differences are probably from the underlying silicon (DAC and ADC). And these discrepancies in dual channel silicon performance aren’t uncommon at all. Video is queued up below.

Please pay attention to the excellent comments by the presenter in the video as far as the text disappearing, the text size. And whether or not these types of issues have been resolved. These types of problems, and bugs played heavily in our decision to Return, not purchase the QA 403 until the platform had matured extensively.
My overall take was:
OK look this box could be good when it matures, AND, the concept That the QA 404 was marketed, and not ever even released: OMG.
Therefore, we purchased the PRISM Dscope three, which definitely has matured, and rocks mightily. And has support. My take was: well, maybe in a couple of years. Maybe.
The QA 403 has a very quiet noise-floor but still.
We can’t have these kinds of Issues going on in a BUSY day-to-day repair and research environment. There’s no time for it.
We can’t have it flaking out, from a software perspective, or whatever, plus the concept of having no one to talk to, other than some Users forum. I mean think about it: It’s like having the end users as a beta test site, with the users responding in a forum, the creator responding, maybe or maybe not. So maybe in a couple of years, perhaps.
To the creator, and end users:

If you’re thinking about buying this unit, and your plan is to hit the ground running:

You will need an extensive patch panel, at least in a busy repair environment.

That includes I/O:


Quarter inch T/S




In addition to BNC.

And the ability to ground lift.

Or a whole lot of adapter cables.

And to the creator:
It would be doing quite a service in offering this accessory. And users with other platforms would purchase it also.

CESAUDIOPRO, what you are proposing is 10 times more expensive. Since I am a hobbyist, this is not an option for me. I am quite satisfied with the qa403…

@CESAUDIOPRO- I am sorry to hear that the QA403 did not work out for you, but just had to comment on your post. I have been using the QA402 and recently, a QA403 and they are the greatest tools for audio work since sliced bread. Not sure how it could not be useful for a day to day repair business- I would have loved to have a QA40x when I had my repair shop. Can’t answer about the research aspect other than if you need to measure very state of the art DACS, it should be useful, particularly at its price point. Like any piece of gear and software, it does have a learning curve. The owner’s guide tells you the basics you need as far as adapters and such, but if you are in this business or hobby, you should be able to figure out how to make your own cables and such as you need them (if they are not readily available). Am not sure about the text problem you were describing, but I have not had any issues with test for the titles and axes. Not sure about the patch panel need- cable go into the unit and out of the unit- if you are testing a preamp the inputs to the QA40x are different than a power amp, but to change everything takes me <2mins. As for support, I have found it to be great as the creator, aka Matt, monitors the forums and is quick to implement software changes if those proposed make sense and can be done with a reasonable amount of effort. The users in these forums are pretty good and while a lot are “newbs”, there is a great exchange of ideas and knowledge. Just my few cents worth…

1 Like

The dScope III is a very nice device, offering more features than current QA40x devices. Of couse XLR inputs and outputs are nice, digital inputs and outputs too. The software is sophisticated and you get proper support.
At a fraction of the price of the dScope III you get the QA403. The QA403 offers 10dB better audio performance, a software which fulfills the needs of most professional and enthusiast users and you get Matt who answers most questions in the forum with his very profound knowledge. Matt designed the QA products with a mindset of bringing very high quality, the right features at an interesting price point.
My personal ranking of audio measurement tools would start with AP followed by QA. Probably followed by dScope, SR1 or R&S (but they discontinued the audio product line).
Although I never worked with the dScope, it seems an interesting products which has cool features required by some users. For most others on the QA forum, I think Matts products are great an we can easily deal with the few weaknesses.

I found that it was difficult, if not daunting to sort out situations
of the differences between what were software bugs, and what were features- IE “Perhaps when it matures, will reconsider, because we would” …specifically in High resolution situations. In our situation, the concept is/was- “OK look- this solution has to work, and we can’t spend time sorting out the difference between measurement errors and actual DUT legit problems”. You start swinging between heavy vacuum tube gear to complex digital gear, with a wide variety I/O,…that is a no go situation. The concept of introducing products like the 404, and keeping potential buyers hanging, for months…especially where they needed something, and chose to wait for the 404 (which they would be waiting for forever)…instead of buying the 403 (which we did)… now THATS problems. It would have been much better to wait and not announce the 404… or say: OK…we decided to not make this up front… and for users to have to ask where is this unit??.. and only get sporadic answers in a forum… that is a broken model. And one of the reasons to wait for the 404 was because of it’s advertised higher input overload protection. You start slamming high voltage PSU rails from shorted large scale power amps… high voltage ultra sonic oscillation situations… all kinds of high frequency/high voltage artifacts from crashed gear under test, and the measurement gear has got to survive on the input and generator sides of the equation… And part of this discussion becomes the concept that (And this applies to FFT boxes like the QA, AP, which we also have, the D scope, what have you), is that they become proofing tools-
What you wind up doing is having a separate rack for the actual repair situation… (That is a fast analog oscilloscopes with no artifacts, solid low distortion signal sources, load system in stereo 2-16 ohms… And having filters can be very helpful-) DSO’s can be helpful…sometimes-
And you need to get to that stuff right away, all of it. It’s like the difference between driving analog and digital mixing consoles. In a lot of ways, there can’t be any waiting on aisle 11. You may only get one shot at the problem.
Maybe the FFT box is connected in parallel on the measurement side, (but don’t crash it on the input side) and then the high-res measurement tools wind up being for proofing, when you’re done, or looking at specific situations i.e. clearing situations with exotic vacuum tube or solid-state power amps, pres, or anything that is on a touring arena stage, plus the consoles, processing etc. and then proof the unit when you’re done with an FFT box like the ones mentioned above.
Now if you’re in busy production environment, and survey says of that is where the QA series of boxes is used lots, where you’re doing the same thing a lot, there is plenty to like about them. Especially from a cost perspective. I think Matt’s brilliant, because HE IS. We’ll look at the QA series of boxes maybe in a year or two again. If it shows in the forums, the release of software where there are bugs in it, then we would be very hesitant. I would expect however that the majority will be sorted out by then. Good luck to all, and good to know you!

See above post for more info

Hi @CESAUDIOPRO, remember the QA404 was being built in preparation for the QA403 seeing an extended outage due to silicon issues. There was nearly a year the QA403 couldn’t be sold due to part availability issues. The specific problem was power supply parts in the QA403…and all of a sudden, over six weeks, we went from having zero of these parts to having nearly 10,000 parts in hand as all the orders at the various distributors filled.

In retrospect, I’m not sure what could have been done. There were daily questions of “When is the QA403 coming back in stock?” and the answer to that was “I don’t know” which upset people too. I didn’t know because Digikey and Mouser didn’t know. And TI probably knew, but couldn’t tell anyone except for the biggest customers. It was a horrible time all the way around. Looking for that same part (SN6505B) that caused so much pain for so long shows there are nearly 1 million parts on distributors shelves around the world just now. TI alone has 308K in stock. WinSource, who was selling the part for almost $300 has 189K in stock and they are selling them for $0.72 each. Far below even TI’s selling price.

It was a strange moment in time. I’m glad it’s over.

I’d prefer to see a maturing/refining of the QA40x software and perhaps a dedicated add-on interface via the expansion connector (if possible) for those who require it for the QA403 for SPDIF/TOSLINK if needed, rather than trying to produce yet another standalone new model analyzer.

The unit is solid, it just needs the bugs ironing out in the software IMO.

Yes, to all above.
Plus an external I/O accessory.
Not the release of new products.
Focus on fixing bugs.
And an accurate delivery/availability schedule that is on the main webpage.