The new Version 1.3.04 looks on my PC like this.
Is it normal o.k. so, the color?
Thanks
Yes, if you check out the What’s New in the Command menu this is my preferred color scheme. You can change it back to what it was if you want. I just like it better than the original and have been using it for months so…
Here are the older color values: #dce4e4", “#ecf4f4”, “#f8f8f8” that go into the Settings values.
Mark
1 Like
Why every version since 1.3.0.1 is using much more CPU when running just response test?
Noticeable 1-2 core CPU usage and cpu fan runs at high speed. No such problems with 1.2.10.1 or lower versions. What changed?
That’s specifically one release - 1.3.0. The one that adds Idle Generation.
As noted in the release notes (4 posts above this one and in the what’s new and in the changelog), it has major changes to support the idle generator. They aren’t yet optimized, and, in fact, I found (after release) problems with tests that average along the sweep (i.e. THD vs Freq).
It would not surprise me if there are cases where 1.3.0 stays free-running.
There’s a next version coming in a day that fixes (most of) the issue with averaging and will also, I think, fix the resource problem. There may yet be more problems with the release due to the complexity of the new module.
I deliberately left idle generation for the very last feature in the original wish-list. I expect over the next month it will end up being substantially better than the prior code - producing better results more quickly - but there may be a bump or two.
Mark
Postscript: after flailing at this threaded code that only fails in release mode in random cases for a week I’ve decided to temporarily back it out since in a few cases the failures do pretty bad things. I posted a build just now (1.3.1.6 I think) that goes back to the pre-idle acquisitions. While I should be able to resolve this there’s no sense leaving a failed release up.
1 Like
I can’t get either QA40xPlot_1_3_1_5 or _6 to run at all - doesn’t open, run from command prompt also not opening or any errors. All previous versions still working.
1 Like
Same here, I get the little version check / installer box that pops up, but the main window never opens.
Mark
Whoops, I left a link to creating my log file in the program and you don’t have that folder. Rather than not build the file it’s apparently just crashing.
I’ll post a new one without that link asap. Thanks for the quick notice both of you. My bad.
Mark
No, thank you. Your willingness to keep this program alive and evolving without funding is honorable to say the least.
5 Likes
Quick Q. I try to do some speaker impedance measure , have a serie res. 0.119 Ohm - “set DUT” -35Db . But I get Y read out as between 0.172 to 0.189 ohm ? The curve look right but not the Zimp
This is 4.7ohm resistor
I’m not sure what you mean about a series resistor of 0.119 ohms. To measure impedance you need a reference resistor that is usually in the 5-50 ohm range. I don’t think a 0.1 ohm resistor would even work since the connector impedance is greater than that. For good accuracy a reference resistor approximately equal to the impedance you test (at 1KHz?) works great.
Have you watched the Impedance video on youtube referenced on the product wiki? You may have the connections backwards or off and the video has a cute but crude picture for reference.
See: Impedance Test · MZachmann/QA40xPlot Wiki
Mark
Ok, I am once again posting a new build (1.4.0) with the idle processing. The idle loop has been completely rewritten (for the third time) and I’m pretty sure it is working reliably but not certain.
Assuming it all goes ok this version works noticeably more quickly and I think the results are more reliable because the QA40x doesn’t keep turning on and off during a repetitive test.
Please let me know if you find issues with the program crashing or the QA40x hanging on the usb processing.
Things to check→
-
I think that the external DAC generator support is still good but that’s not certain with the new code.
-
I only have a few computers to test on and they are all pretty fast. It is possible that a slower computer may have troubles keeping the queue filled with the timing I’m using. If you think you have a particularly slow computer you might try doing a THD vs Frequency plot at 192K/16K with averaging of 2 or 3 and ensure it works - that’s the most stressful setup. Note that setting has poor low frequency accuracy.
I really appreciate everyone’s patience.
Mark
Postscript: External generator support is not working and will take some effort to upgrade into the new idle code.
Okay you do not set it up as QA does , balanced input . But unbalanced and voltage/voltage, not a current/voltage measure which is normal way to do it.
Yes, you’re right I don’t do it that way. It seems like a lot more complexity for a task that otherwise works fine with three alligator clips and a random resistor.
Mark
Okay , it works nice have to twik the serie res. (software) value a bit to get a precise reading. Scale is up to 2Kohm can that be increased ? 20K or higher, yes I know it will give more noise/hum but I have some old guitar it could be fun to sweep 
And thanks for your nice work it really make QA40x easy to use
HI
is it possible to measure also the filter inductances? Which is the max value of H?
Walter
I had no problem measure an air coil and it showed correct Z, phase etc. One thing thought you can not attach read out markers on the curve that would be nice
Extra amplifier is not needed to measure speakers? Can QA403 handle that without extra distortion that could skew measurements? And it works also when you measure a complete speaker with crossover?
If you want to measure higher impedances just use a larger reference resistor for more accuracy. This also adds additional damping for the circuit for reactive loads.
Take a look at the wiki video: I test a 30uF capacitor standalone and a woofer (tuned circuit) and they both work fine. The capacitor test values, even using a 20-20KHz chirp and no amplifier, are accurate down to 20Hz and 350Ohms.
In my experience finding an amp and hooking it on in the middle of things and having it work well with reactive loads is always problematic. Keeping a circuit board around with the testing circuit and using 4 output cables is a bridge too far. Admittedly I’m lazy and using a few test clips and a resistor is easy. If you all decide you have specific use cases where differential inputs is necessary let me know and I’ll add it to the roadmap.
I haven’t spent much time on the impedance test extras, but it’s easy enough to have it calculate Q and peak frequency for tuned circuits. Not to mention calculate T/S parameters. Both of those are on my roadmap.
Adding another impedance test method is possible but would take a few man-days and it’s just pretty low on my list without a specific use-case.
- Increasing the scale above 2K is easy and I’ll take care of that.
Mark
1 Like
have you tested a 20H, for example, coil
‘I have some and I would like to test them
Walter
No, I’ve never tested anything greater than about 50mH. I have a 3.76mH lying around that I just tested and it worked fine with a 10 ohm reference but a 20H would have 5000 times the impedance. If you use a 20K series resistor or so it should work pretty well. No guarantees since it’s very inductive and the input impedance of the QA403 may need to be included in the math. I’d use a pretty large fft size (256K or so) and scan at 48K with a small chirp range (20-100 maybe) to get good low frequency resolution if you use a chirp.
=================
I did just try this experiment: I took out an avel toroidal transformer (Y236850) and connected the input as a load to test. Using my handy dandy $80 impedance meter it measured it at 1.2H.
I then scanned it with QA40xPlot using alligator clips and a random 2.2K resistor I had lying around. It measured the inductor as 6.5H. I thought wtf….
So, for sanity, I generated a 100Hz sine wave and used my trusty VOM to get both AC voltages at 100Hz. .48V total and .3V across the inductor which implies an impedance of about 2.2*3/1.8 or roughly 3.6Kohms at 100Hz or about 5.8H. So there’s that.
Mark
Note I cheated on this chart by using the next version of QA40xPlot with higher upper limit.