Voltage divider for high power measurements

I want to measure an amplifier that can deliver 100w into 8ohm with the QA401. Seems like my best option would be to make a voltage divider, but my experiments so far aren’t very promising.

This is my setup:

For now it’s just the QA401 connected in loopback with a voltage divider in the middle. The divider consists of a 220k resistor and a 10k resistor.

The divider works, as I get a 28dB attenuation. But the noise is pretty bad:

Just to verify that the cables are ok, this is a screenshot with the hooks connected without any resistors in the middle:

I also tried with the resistors on a breadboard, but it actually made it worse.

Any suggestions on how to get better results?

Check out this from the q402 product page:

Hi @blurpy, the link @VAR posted is the best starting point. In your specific case, you have made a divider with a 10K and 220K resistor, so the impedance of the divider is about 10k (10k parallel with 220k).

If you go to a resistor calculator such as HERE and plug in your values, you will see the noise from a 10k resistor in 20 kHz bandwidth is about -115 dBV, which might seem fine since that has the same noise floor as the QA401. But, in practice you need your divider to have 5-10 dB less noise than the analyzer, otherwise you will degrade the noise floor of the analyzer. So, if we cut our divider impedance in half, we’d see the noise drop by 3 dB. So, you’d really want your divider to have an impedance of no more than 2500 ohms, but ideally around 1k.

In addition to the resistor noise, you need to consider the input opamp bias current noise. For the OPA1612, as it’s a bipolar opamp, its “self noise” will be considerable with source impedances rising beyond 1k. But around 1k or lower, the impact will be modest.

So, thus far, there are two good reasons to keep your source impedance below 1k ohm.

Finally, the higher your source resistance, the more susceptible your setup will be to picking up radiated energy all around. And in your case, that looks to be the greatest offender. Your 10k ohm effective impedance means it’s easy to pick up powerline. If you used a 100k effective impedance, the powerline spikes would be worse still, and a 1k would be better still.

So, when picking a source impedance for a divider, aim to keep the effective resistance below 1k. And the lower the better.

When measuring power amps, the best of all worlds is to build the attenuator directly into the load as the article @VAR linked to shows.

Thanks for the input!

I need to keep the amplifier running for a while to provoke some distortion issues, so I use 200w loads and large heat sinks. I was therefore planning on adding a voltage divider in parallel with the load to avoid sinking much current into the resistors in the divider. I’ve blown a 30dB BNC attenuator already :confused:

So while it looks neat with that circuit board, it doesn’t seem like a good match for my need at the moment, but maybe I will make one for later use.

The thermal noise information was very useful. I adjusted my voltage divider to 20k and 680r, and the noise improved noticeably:

So that keeps the impedance below 1k, if I understood you correctly @matt.

I was wondering if my loopback setup was causing some of that noise as well, so I added my pre-amplifier in the middle like this:

THD certainly looks much better:

So hopefully this voltage divider will work in parallel with the loads connected to the amplifier, and survive.

If it works ok, I might solder up a perfboard. Maybe that will remove some of the powerline noise.

1 Like

Just a quick update. Been testing the amplifier with the voltage divider this weekend.

Pretty close to 100w here:

Running it for a bit at the same settings:
(Update: I see I forgot to set the internal attenuator the same when taking the screenshot because I was changing back and forth to see the difference - it wasn’t much)

It goes into clipping and oscillation, and manages only 25w clean.

I know it’s not the voltage divider provoking the result, as I’ve been seeing this on the scope:

Oh, in case anyone is curious, here’s the physical setup:

QA401 output → BNC to RCA adapter → RCA splitter for mono to stereo → line input on the amplifier
Amplifier speaker output → 200w load + oscilloscope probe + voltage divider - all connected in parallel

The breadboard with the voltage divider is setup like this, in the same order as the image:

speaker + — 20k Ω — QA401 signal in — 680 Ω — QA401 gnd — speaker -

1 Like

It kinda looks like you may be having the same problem I did when measuring an amp into a dummy load. If you look at the “strange behavior when measuring the phono input” thread, check to see if you are connecting your QA401 L- input via a separated cable instead of using the ground from the BNC shell on the L+ input, which is what it looks like your picture is showing… It solved most of my problems

1 Like

Thanks for the reference, that’s an interesting discussion!
You are correct, I do single ended measurements here.

The signal ground, the chassis ground and the speaker negative leads are all the same on the amplifier. The amplifier is a floating design, with only 2 leads on the mains cable. I have mostly done my testing with the oscilloscope, looking for visible distortion based on the input from a function generator. Both connected to earth ground. And to make sure the issues I saw was not related to earth ground I also did the same tests with a battery operated scope and generator. The results looked the same.

The next step was connecting the QA401, and I needed to step down the voltage a bit.

I have now put my circuit on a perfboard, so I’ve done some more testing. These are all with the QA401 in loopback (no amplifier).

First test is the same as my second post, first picture, except this time with the perfboard:

This shows a noticeable improvement to the right of the fundamental, though the numbers don’t change all that much.

Next test was using a BNC to dupont adapter instead of the BNC to grabbers on the output of the voltage divider:

That was actually a big improvement. I tried switching it around, using dupont on the input of the voltage divider instead and it was back to the previous result.

Third test was doing a differential measurement:

The results look slightly better.

And the last test is that I replaced the BNC cable on L+ with a small BNC adapter that lets me plug the dupont adapter right on the input of the QA401:

It got slightly better now as well.

I only have 1 of those adapters, so maybe it would be even better if both the L+ and L- used the adapter instead of a cable.

Here is what the final setup looks like:

There are so many little details that all have different effects on the result here. Not too bad now though!

I will do some tests using this setup on the amplifier as well, when I get the time.

It looks like you should be ok the way you are testing now as the grounds from the left inputs are separated. When I looked at how neat your workbench is I started chuckling. Mine is quite the rat’s nest as my two dummy loads are a bit large.

Hehe, the pictures from my loopback setup is from my computer bench. I don’t work on projects there so it’s not too bad. My electronics bench is a different story :laughing:

That makes sense. My “computer workstation” is part of my electronics bench and has it’s own stereo system, complete with cassette and turntable. It is in what we call an “Arizona Room”, and not my garage like it used to be. One day I will probably do a thread on the loads that I made…

I’m back with some examples using the amplifier, running at 1 watt.

First is from my previous tests using the breadboard voltage divider, single ended:

The second is with the perfboard, single ended:

The line noise increased here, but it looks a little bit cleaner with the harmonics.

The third is with the perfboard, differential:

Nice improvement here on everything.

And the fourth is without the voltage divider, using the internal attenuator, differential:

The line noise improved a bit this time.

In conclusion - the built in attenuator is the cleanest option, though it will only help me to half of the rated power of this amplifier. The differential measurement with the perfboard is the second best option, and it works up to 100w. Maybe it’s possible to shield it better to reduce the line noise.

Looks pretty good. How is it at 100w with the perfboard ?

This is what it looks like single ended:

And differential:

These are both at max power. Even the slightest increase in volume creates oscillation. The confusing part is that the differential measurement displays a lower voltage and power, even though it says 80Vpp on the scope. I see this effect without the voltage divider as well (at lower power). Not sure why it’s like that.

The differential measurement looks better and I am not surprised once you hit max power just a bit more volume will start oscillation- I am seeing that with the amp I am testing. I am not surprised that the two measurements differ slightly since the grounds are not the same…?

The BNC ground is unconnected in the differential measurement.

I would have thought the voltage seen by the QA401 would be identical in both cases. In the single ended measurement the signal should be what’s seen between signal and ground of L+, and in the differential measurement it should be what’s seen between signal on L+ and signal on L-.

But it could be that it expects a real differential signal when connected that way, and L- being connected to ground from the amplifier (since negative terminal is ground) is confusing it?