Goal: Connect a scope or analyzer to the output of a power amplifier and observe amplifier behavior when presented with one or more dummy loads.
Assumptions: I will need a dummy load of the value I expect to drive (2, 4, 6, 8, 16 ohm, or reactive). I will also need to reduce the voltage so as not to blow up my measurement device.
Circuit:
This is what I think the circuit should look like. Signal passes through the dummy load, then through the voltage divider R1-R2 where it is sent to the measurement device. I assume R1-R2 will need to reduce the voltage from a maximum of ~10-20v to ~1-2v.
Am I even close to being correct?
Assumptions: I will need a dummy load of the value I expect to drive (2, 4, 6, 8, 16 ohm, or reactive). I will also need to reduce the voltage so as not to blow up my measurement device.
Circuit:
This is what I think the circuit should look like. Signal passes through the dummy load, then through the voltage divider R1-R2 where it is sent to the measurement device. I assume R1-R2 will need to reduce the voltage from a maximum of ~10-20v to ~1-2v.
Am I even close to being correct?
Most amps would just require a x10 probe for the scope.
Most DVMs can easily measure an amp's output, although they seldom are true RMS.
In your circuit, the dummy load should be directly across the amp output.
The amp output voltage attenuator (if needed) would be also in parallel with the amp output.
It has no direct relation to the load.
Most DVMs can easily measure an amp's output, although they seldom are true RMS.
In your circuit, the dummy load should be directly across the amp output.
The amp output voltage attenuator (if needed) would be also in parallel with the amp output.
It has no direct relation to the load.
Last edited:
I was thinking about running the signal into an audio interface and using a software scope. I assume I would need to apply the 10x reduction myself in that case?
Like this then?
Yes.
You could size the attenuator so the resistors dissipate around 1W for a 100W amp.
Using 720R (rated at over 1W) in series, and 80R in shunt would work, but avoid
wire wounds. Be sure to use a non-inductive dummy load also.
You could size the attenuator so the resistors dissipate around 1W for a 100W amp.
Using 720R (rated at over 1W) in series, and 80R in shunt would work, but avoid
wire wounds. Be sure to use a non-inductive dummy load also.
Last edited:
Excellent. So then I need to pick values for R1 and R2 so that "Out+" is 10x lower than "Amp+" (or use an x10 probe) and I should be good to go.
Rayma says a 10:1 scope probe is all you need to measure any audio amplifier and he is correct. Your
circuit is making the assumption that the output is single ended but this is not always true. The Texas
Instruments class D amplifiers are mostly BTL so the '-' outputs are not ground. I've seen some old H.H.
Scott tube amps that the '-' output is not ground. This means your dummy loads must be 2 identical AND
isolated systems. When I measure BTL amps I use separate probes for each output and add them in the scope.
The load I built uses 4 16 ohm 50 Watt resistors in 4 parallel blocks for 2 8 ohm resistance blocks per channel
that can be run as 4,8 or 16 ohms per channel. The 2 units (4 resistors each) are totally isolated and can be used
for BTL or plain old single ended amps.
G²
A scope probe is designed for the specific input impedance of a scope, so unless you use
something like a Pico scope, a scope probe won't be suitable for this. Just use precision resistors
of adequate power and low enough values, mounted in a box with connectors.
Yes, be careful of bridged amps with two floating outputs. Unless your audio interface has a
balanced input, you won't be able to test them. If it does, then make a balanced attenuator,
for example a series 360R in each line, and a shunt 80R.
something like a Pico scope, a scope probe won't be suitable for this. Just use precision resistors
of adequate power and low enough values, mounted in a box with connectors.
Yes, be careful of bridged amps with two floating outputs. Unless your audio interface has a
balanced input, you won't be able to test them. If it does, then make a balanced attenuator,
for example a series 360R in each line, and a shunt 80R.
Last edited:
But you're gonna be bringing both channels "grounds" together in the BTL example you give if measuring simultaneously.. ?!
(isolated scope or not)
When you try that, connect the ground clips of the two probes together,
but not to anything else, and subtract the channels.
but not to anything else, and subtract the channels.
I'm a little confused by what you mean here. Is a scope probe different from an x10 probe?
Do you mean here to use good load resistors and then not use attenuation but connect to an x10 probe?
And if I wanted to run into a PC sound card, I assume I'd need to do something like 10x attenuation myself to get it down to 2v or less?
The voltage divider you added in your later pics just reduces the take off voltage to the scope to a more manageable level for measurements.
i.e. you only need a fraction of the voltage across the load not the full load voltage - either with a 10:1 probe on a conventional scope, or just straight off your 10:1 resistor divider to your picoscope.
But be wary of BTL outputs, or aware of the difference
i.e. you only need a fraction of the voltage across the load not the full load voltage - either with a 10:1 probe on a conventional scope, or just straight off your 10:1 resistor divider to your picoscope.
But be wary of BTL outputs, or aware of the difference
A x10 scope probe has an internal high value series resistor of 9M, and scopes have a 1M input impedance.
So the attenuator formed is: 1M / (1M + 9M) = 1/10.
Because of the high impedance level, an adjustable compensation capacitor is required across the
9M resistor to match the probe to the scope input, which also has some inherent input capacitance.
Otherwise, the frequency response would droop above 1kHz.
A scope probe is not suitable for an audio interface, it's only good for use with a scope.
A sound card would be easy to overload, so clamp diodes may be needed at the input.
So the attenuator formed is: 1M / (1M + 9M) = 1/10.
Because of the high impedance level, an adjustable compensation capacitor is required across the
9M resistor to match the probe to the scope input, which also has some inherent input capacitance.
Otherwise, the frequency response would droop above 1kHz.
A scope probe is not suitable for an audio interface, it's only good for use with a scope.
A sound card would be easy to overload, so clamp diodes may be needed at the input.
Last edited: