I am refurbishing an amplifier I built a few years ago. I used this QA401 at the time. This time, when I connect it up and power the amplifier using a Variac with both voltage and amperage meters, the amperage increases with the voltage, just like a dead short. I did a loopback on the QA401 and it looks just fine. I look at the QA401 output on a Keysight DSO, and again the sine waveform looks just fine. I use my HP Function Generator to input a signal to the amplifier, and that arrangement powers up just fine, and the amplifier seems to be working just fine.
I am using the QA401 from an HP ProBook 4545s running on batteries.
Hi Ceulrich, it’s likely you’ve bridged an output and ground. If the QA401 looks fine in loopback, then is it possible that you have shorted an amp output to chassis ground? For example, if your amp is push pull, and you connect a scope probe center to the speaker +, and then clip the scope ground to speaker -, and then drive the input of the amp with a BNC or RCA, then that setup has tied the speaker - to ground.
I have attached a drawing of the equipment connection setup, and a schematic for the grounding topology used in the amplifier. The QA401 negative inputs and outputs have shorting caps.
I understand what you are suggesting. But I am surprised, because, I think I have been using this arrangement for many years, and have not had this problem previously. Why would it not show up when using the HP Function Generator (HPFG)? I should note that when using the HPFG, I put a 75-Ohm decade attenuator between the HPFG and the amplifier input, because the minimum output for HPFG is around 100 mVrms.
Hi @ceulrich, I think your arrangement is fine. What might be misunderstood is that the output BNC shell on the QA401 is tied to the input BNC shell.
So, in your drawing, on the QA401 input you show “BNC to Dual Banana Coax”. The minus leg of that Banana Coax is at the same potential as your input ground. And so, if you connect it to a push-pull amp (modern class D or tube amp with push-pull outputs via transformer) then you are introducing a short.
Now, your scope likely has it’s BNC shell tied to earth ground. If you amp is powered by a floating supply, then the ground is getting established at the scope.
In short, it’s hard to know what might be going on. But my guess is that somehow grounds and signals have gotten tied together. The first time I measured a TPA3255 class D amp, I was mystified because I failed to grasp that both the speaker outputs idled at half the supply voltage. And I tied the BNC shell to the amp minus and instantly caused a fault. Fortunately, the TPA3255 is very smart, and can cope with such mishaps.
You might try using scope probes on the QA401 + and - inputs (no ground clips!) and then clip the tips of the probes across teh 4 ohm load and see if that changes things.
If he has no ground reference the 'scope traces will be all over the place.
He should activate “add” and “invert” for x/1 and y/2 channels on the 'scope to do a differential measurement with the 'scope on a floating source.
Hello Matt – at least I spelled your name right this time,
Thanks for your continued support.
I am beginning to think this may be interesting enough warrant some more background information.
I have attached a schematic for the amplifier. Also attached is an oscillograph of the amplifier output with a shorted input, and the associated distortion spectrum. My next measurement was to be a distortion spectrum of a 1KHz input for a 1-watt output. I powered down the amp, connected the QA401 output to the amp’s input, and powered up, only to discover the fault, as described in my first post.
Some interesting observations made in the last few days: I understood from the beginning, that the connector shells for the QA401s’ input and output were connected. I found that the fault condition ONLY occurs when both input and output of the QA401 are connected to the amp. It does not make any difference if the QA401 is powered or not. If the DSO is disconnected, the fault persists. I was totally confused. Matt’s comment about amplifier topology caused me to start thinking more about the amplifier rather than just grounding topology. So, I removed the bolts fastening the MOSFETs to the heatsink, and gently prying them away from the heatsink. And just like that, the fault was gone. I re-bolted the MOSFETs back to the heatsink, one at a time, checking for the fault for each. Still fault free. So, I assume there was some sort of fragile defect in one of the insulators. Replacing the insulators is probably the next step.
