Hi, searching a solution to measure response curve, distorsion of a bridged class D amp. The problem is not the measure itself, but how to connect the amp to soundcard ? (ground is not the same). Do I need transformer or Di Box to isolate? Thank.
with two probes? differential probes. Or maybe something like
http://www.libinst.com/difprob.gif
what do you mean, input signal is input signal (unless you are talking about some balanced input), sure it will have to be on the same ground as the soundcard...if nothing else, by signal cable
why would you need to insulate it from soundcard? and if you do, wouldn't it be more easy to use portable player, or audio transformer or DAC with spdif (optical)
I mean, there are so many options, you will need to explain more what you want to do
http://www.libinst.com/difprob.gif
what do you mean, input signal is input signal (unless you are talking about some balanced input), sure it will have to be on the same ground as the soundcard...if nothing else, by signal cable
why would you need to insulate it from soundcard? and if you do, wouldn't it be more easy to use portable player, or audio transformer or DAC with spdif (optical)
I mean, there are so many options, you will need to explain more what you want to do
You can also isolate the ground (earth) of your instrument (the PC in this case) for a quick measurement and connect the probe directly to the load. Be careful with this, as any peripheral that is earthed will produce a shortcircuit. This technique is useful with oscilloscopes, although some safety considerations have to be had in mind.
The BTL outputs of a lot of class D amps run hot, that is they run at a fixed multiple of the supply voltage. This will fry most sound cards or opamps if fed directly into them. This isn't a concern when driving loudspeakers as the same hot potential is present on both inputs to the loudspeaker, so no current flows through the voice coil.
DC coupling caps are common on the inputs of many audio devices, but they aren't always present. These will filter out the DC, but if they aren't there they wont and the sound card will explode so it's not wise to assume that they are. In any case, lots of class D amplifiers rely on the bridged output to improve performance by cancelling out some of the switching noise, so measuring the single ended performance isn't going to accurately describe their performance.
The best choice would to build a little circuit around an opamp. You want it configured as a differential amplifier. Keep the resistance values small as to keep noise to a minimum and cap couple the inputs. The caps will filter out the DC component and the opamp will then sum the difference between the inverting and non inverting inputs giving you an accurate representation of the class D amplifiers performance. Depending on the power levels you want to measure you will need to set-up a potential divider before the opamp inputs to keep the input level delivered to them low.
Good resistor matching is advised, use at least 1% components otherwise the cancellation affects/advantages of the bridged output could be reduced.
DC coupling caps are common on the inputs of many audio devices, but they aren't always present. These will filter out the DC, but if they aren't there they wont and the sound card will explode so it's not wise to assume that they are. In any case, lots of class D amplifiers rely on the bridged output to improve performance by cancelling out some of the switching noise, so measuring the single ended performance isn't going to accurately describe their performance.
The best choice would to build a little circuit around an opamp. You want it configured as a differential amplifier. Keep the resistance values small as to keep noise to a minimum and cap couple the inputs. The caps will filter out the DC component and the opamp will then sum the difference between the inverting and non inverting inputs giving you an accurate representation of the class D amplifiers performance. Depending on the power levels you want to measure you will need to set-up a potential divider before the opamp inputs to keep the input level delivered to them low.
Good resistor matching is advised, use at least 1% components otherwise the cancellation affects/advantages of the bridged output could be reduced.
I tried this method. Because of classd ampliers' output have a lot of high freq signal, this method didn't work for me. op-amp didn't work correctly.
Ok this works with pure sine wave, but classd output isn't pure sine.
I have to measure my btl class d amplifier output(thd, thd+n, snr etc.). I couldn't measure until now. Help!!!
I know this is an old thread however does anyone have a decent circuit for measuring btl class-d amps. I understand it would be similar to an AES17 (ebay) but a few hundred dollars seems a little steep. Any input would be appreciated
1st step: divider + 1st order LPF on both outputs.
2nd: Instrumentation amplifier. Eg LM7171 datasheet.
3rd: anti aliasing filter.
2nd: Instrumentation amplifier. Eg LM7171 datasheet.
3rd: anti aliasing filter.
Tank pafi for your response. The amplifier I'm measuring already contains a 20kHz LPF. So I assume this is sufficient. Will the lm7171 provide sufficient accuracy in terms of distortion? I will be feeding this signal into a PC sound card, what purpose does the antl aliasing filter have?
Filters are not perfect, there is significant residual. LM7171 is very fast, but maybe not the most linear type. There are more linear types, but they may not be as fast, so they are better with an additional LPF. Many soundcards have good AA filter, but not all of them. Weeeery low residual is allowed, if you want precise measurement, this because I mentioned AA filter. 10 uV on the input of the ADC can ruin your measurements.
I have measured the output of one output (with an oscilloscope), yes it still contains traces of the 310kHz signal. in fact it took a 2nd order filter to reduce it to an acceptable clean level. I will move onto the differential OPAMP and see what I get there. Have you built this circuit before with the LM7171 ? I'm happy to fork out for something better if you can recommend one.
- Status
- Not open for further replies.