Not sure what I am doing wrong here - but when I try to use my Aiyima A07 for testing the different features of the QA403 - it doesn’t work unless I feed a signal from another signal generator? The exact same setup works fine if I unplug it from the Aiyima and use for instance - a HK Citation 12 I have on the shelf - but it’s a clients system and I don’t want to take any chances with a clients system when I am learning how to use the QA403 properly.
I seem to recall reading something about it having to do with the fact that it’s being powered from the ac/dc adapter…but can’t find anything here or elsewhere regarding it?
What FFT size are you using? Can you try 48K sample rate with a 64K fft? I ask because some amps and bluetooth have a noise gate, and if you have 2K FFT and a 48K sample rate (2/48=41mS burst) the noise gate never opens.
So, if you use a 48K sample rate and 64K fft, that will give you a 1.33 second burst.
The amp probably has a gain around 25 dB. So, you could set the QA403 generator to -25 dB, and then while stopped (not acquiriing) press the IDLE button. That will generate a continuous tone, and you can measure the amplitude at the speakers with a DVM. It should be around 1V.
Finally, how around you measuring the output? The TPA3255 amps will idle at half the supply rail. That is, if your supply is 50V, then speaker + will be 25V, and speaker - will be 25V. And they will diverge as needed to generate a positive or negative voltage. That is, to create 1V across the speaker, speaker+ will go to 25.5V and speaker- will go to -24.5V. So, you definitely want to make that measurement differentially and NOT connect the BNC grounds on the QA403 inputs.
When I was first measuring the the TPA3255 I tried to make a single-ended measurement by clipping a scope probe ground onto speaker-, and the probe tip on speaker+. And with the input of the TPA3255 connected to BNC ground and the L+ output, that effectively shorted the ground to the 25V. Now, the TPA3255 didn’t care at all. It just turned on its error light and handled it no problem. But it took a bit of head scratching to figure out what was going on.
Finally, I don’t think the AC/DC should have anything to do with it. I’ve ran my TPA3255 from a 48V telecom supply, from a lab supply, etc. All are isolated. But, it is interesting when you use an external generator you might be bringing over an earth ground that is somehow enabling things. Is that a possibility?
I was using a 192k sample rate with a 2k FFT … I have had the QA403 for 6 months but life has been busy and I am just now getting a chance to learn to use it - didn’t even notice the fact that I was using such a low resolution on the FFT I tried with a 48k sample rate and 64k fft - no difference.
I have made cables that have a single BNC which I was connecting to the + Inputs (- inputs have 50ohm terminators) and on the other end both + and - have a Y connection so I can connect one side to a dummy load and the other to the amp.
I plugged the speaker output into the DVM - and at 50% volume measured 1v with a 150mV 1k input from sig gen - with a 158mV signal from the QA403 I am seeing no output…seems that it could be a grounding issue but not sure what the proper way to connect it to use the QA403 output would be?
When time permits, please try this with your QA403 (unconnected to anything)
- File->New Settings. This puts the app into a known state.
- Right click on the GEN1 button and change the amplitude to 0 dBV
- Press the GEN1 button
- Press the IDLE button near the front panel.
Measure L+ out with a DVM. It should measure 1Vrms (some DVMs don’t have the bandwidth required at 1 kHz, so you might be a bit lower. But you can change the frequency to 60 Hz if needed).
Then, do the following:
- Disable IDLE
- Connect the L+ output to L+ input using a BNC jumper
- Press RUN button
You should see the following. This is showing the left trace (blue) at 0 dBV, and the right trace is in the noise (because it wasn’t jumpered). You can swap the jumper around as needed to confirm that all outputs and inputs are measuring as expected.
If you can confirm the QA403 by itself is acting as expected, then we can turn to the grounding.
The page HERE has a picture in the section titled “Connecting to the QA401” that shows how you might connect.
Note the grounds on the QA403 outputs are usually achieved using a BNC to RCA adaptor (see the section “Getting to XLR” for a picture HERE, and then connect to the amp input using RCA cables. This establishes your ground at the input of the amp.
And then, you measure across the load using scope probes WITHOUT the grounds connected to anything. So, the QA403 Output and amp input share ground. The output of the amp is floated and you make a differential measurement across your load.
Please report back what you learn!
OK - so my QA403 is working fine - measured exactly 1v on both outputs using your instructions. I also did another test with the A07 - and with BNC > RCA > AMP out of the QA403, and using a BNC > Banana jack center pin only from positive speaker output to QA403 +input and negative speaker to QA403 -input…it now works properly!
Obviously I was doing it wrong…I was under the impression and I am not sure where I found it but I wouldn’t make it up or take something that wasn’t on your website - I thought the proper connection for a single-ended amp (which is 90% of what I work on) - was BNC connected to +input and the other end connected to both + & - speaker out on the amp, and the -input terminated - this is how I have tested quite a few SS and tube amps and was always getting results. Although I have really only been looking at the real-time distortion/rms/snr readings until I had time to learn how to properly run tests - didn’t want to take any chances with clients amps.
Part of the problem I feel is there is no one place to find the information, what is there is a mix of qa40x versions, the wiring diagrams are somewhat lacking - and there is little mention of the differential cables except the recent thread. I do see in the link you provided that the differential method that I used to make this work is the proper method, but it would be nice to see the various connection options and reasons for using each.
Thanks very much for the quick help @matt !