I'm not sure how it's done 'correctly'; perhaps there isn't a real standard. What I do myself is input 1kHz at line level and connect a an appropriate ohmic load to the output. Then measure Vpp on the output, convert that to RMS and use the number to calculate output power. When reporting figures, I will mention input signal, frequency and the value of the ohmic load.
I guess that's the best you can do, really: describe the conditions you used to measure power. All too often, power numbers are mentioned, but all the vital information is left out.
I guess that's the best you can do, really: describe the conditions you used to measure power. All too often, power numbers are mentioned, but all the vital information is left out.
Hello Troncones
You need to first know what form the measurement instument's numbers represent. In the case of a DMM it will present the output AC voltage level in rms volts. In this case you simply measure the output voltage at the test frequency across a 8 ohm dummy load and then do the math (W = E^2/8).
On the other hand, if you measure the ouput using an oscilloscope the AC sinewave measured from the top of the positive peak to the bottom of the negative peak will be a measurement of peak to peak voltage. In this case you need to do the math by converting the voltage to rms in the eaquation. The math then is W = ((Vpp/2.828)^2)/8.
The loads are usually made up of multiple non inductive resistors paralleled up to equal 8 ohms at a high enough wattage to serve as an effective dummy load.
Amplifiers are tested for output power while being driven with sine waves and done so at multiple frequencies to understand the amplifiers ability to deliver power say at 20hz, 40hz, 80 hz, 120 hz 1,000hz and so forth up to say 20khz. This allows one to fully characterize the amplifier power delivery ability.
Mickeystan
You need to first know what form the measurement instument's numbers represent. In the case of a DMM it will present the output AC voltage level in rms volts. In this case you simply measure the output voltage at the test frequency across a 8 ohm dummy load and then do the math (W = E^2/8).
On the other hand, if you measure the ouput using an oscilloscope the AC sinewave measured from the top of the positive peak to the bottom of the negative peak will be a measurement of peak to peak voltage. In this case you need to do the math by converting the voltage to rms in the eaquation. The math then is W = ((Vpp/2.828)^2)/8.
The loads are usually made up of multiple non inductive resistors paralleled up to equal 8 ohms at a high enough wattage to serve as an effective dummy load.
Amplifiers are tested for output power while being driven with sine waves and done so at multiple frequencies to understand the amplifiers ability to deliver power say at 20hz, 40hz, 80 hz, 120 hz 1,000hz and so forth up to say 20khz. This allows one to fully characterize the amplifier power delivery ability.
Mickeystan
A dummy load can also be built by using wire wound resistors.
I for example have done one by connecting parallel 27 pcs. of 220 ohms 5 W wire wound resistors. The parallel connection makes the total inductance of such dummy load very low. Actually the small inductance it still contains makes this load to resemble a real loudspeaker better than pure resistive load.
Anyhow, the difference is minimal and this sort of load can be used for all testing purposes.
I for example have done one by connecting parallel 27 pcs. of 220 ohms 5 W wire wound resistors. The parallel connection makes the total inductance of such dummy load very low. Actually the small inductance it still contains makes this load to resemble a real loudspeaker better than pure resistive load.
Anyhow, the difference is minimal and this sort of load can be used for all testing purposes.
Thanks everybody.
I don't have a Distortion Meter. I might play around with the RoMac sound card meter. I'm open for suggestions about an economical entry level Distortion Meter.
For now, I was just going to use a 25watt 8 ohm and watch the scope for signs of clipping. This should at least give me circuit A to circuit B comparison.
I don't have a Distortion Meter. I might play around with the RoMac sound card meter. I'm open for suggestions about an economical entry level Distortion Meter.
For now, I was just going to use a 25watt 8 ohm and watch the scope for signs of clipping. This should at least give me circuit A to circuit B comparison.
There are dummy complex loads that can be devised to follow an 'average' varying load reminiscent of loudspeakers. Those can be useful to valve amps DIYers so to evaluate the frequency response changes due to usually low damping factor, and to check power availability across the spectrum as the load plunges and escalates. You would not want to jeopardise a powerful loudspeaker as load for testing accidents, even if in a place without neighbours and wearing ear protection.
You might need very good ear protection if the speaker is attached, and very understanding neighbours. Also, domestic speakers may not be very good at handling continuous sine waves and could overheat.
A dummy load and oscilloscope is the easiest method. Increase the input until the output just begins to visibly clip - it seems that this is not too different from the point where distortion becomes clearly audible. Then remember that the scope gives you peak, not RMS voltage.
A dummy load and oscilloscope is the easiest method. Increase the input until the output just begins to visibly clip - it seems that this is not too different from the point where distortion becomes clearly audible. Then remember that the scope gives you peak, not RMS voltage.
Yes.
I o'scope the output
1000Hz is what I use
An 8R / 30W test resistor I got at Rat Shack for this purpose. For the actual test, I clip the test resistor to a piece of 1X1 aluminum angle stock to serve as a heat sink, since it will get pretty hot otherwise with significant power going into it.
I o'scope the output until there is clipping, then back off a bit. Then read off the Vp-p from the screen. It's simply a matter of figuring what power that represents:
Vrms= Vp-p / srqt(8)
Po= (Vrms)^2 / 8
NBD. The results I get are in line with design nominal values, considering the core and copper losses in the OPT.
Another thing to try is running at a low frequency like 25Hz to test for core saturation. Here I found that you can usually get ~32W out with a 30W rated OPT before you start to see core saturation. That was especially the case with a 6BQ6 design that predicted some 37W out, but with a 30W rated OPT that I had in the junk box.
That was also the case, some 30W out, with an 807 design that promised 26.5W out in Class AB1. Providing cathode follower grid drivers allowed for some excursion into Class AB2, hence the higher than promised Po. That also means good overdrive performance. I prefer to design so that the front end stays out of clipping when the finals either clip and/or the OPT core saturates.
And perhaps this DIY tip is useful as well: In the past, I have encountered situations where I didn't have a resistor of the right value and power rating at hand. In those cases, I took some light bulbs (12V car types or 230V domestic ones) and put them parallel to get the load I wanted. It's an easy (and often cheap) way to make a good-enough load that can handle the power you want. And with a little bit of luck, you can use this setup to read a book when it's dark.
Hmm....
Only trouble is, the 'cold' ohmic value will be quite different to the Hot/lit-up value, so will cause your measurements to be way off.....
A lamp may be say, 5 ohms cold, but 100 ohms lit up....
I think that Miles's method is prolly the best accepted method,--Use it myself...
I use a non inductive power resistor. If you don't have a scope, or even if you do, this works great... also connect a speaker to the output through a 1 watt 470 ohm resistor so that you can monitor the sound at low levels. This way you don't stress your speakers or your neighbors . Play a sine wave tone in the 200 - 500 Hz range. Turn up the volume until you hear distortion, which is a harsh sound. Back off just enough to make the harsh sound gone. With a true RMS meter set to AC volts, measure the voltage across the output terminals. Square the reading and divide it by the resistance. This will be the Max continuous RMS clean power output of the amplifier. If the amplifier has more than one channel, you should connect a resistor to each channel to see the proper value with all channels driven (don't forget to feed the signal to the input of all channels!)
Your ears will be quite sensitive to the onset of clipping at lower frequencies. You may even hear it before you see the tips of the sine wave flatten on a scope. It is harder to hear at higher frequencies.
To evaluate the amplifier at normal listening levels, I play a varied selection of familiar music and simply judge what I hear. No distortion figures needed. If it sounds good to me, that's all that matters. Be aware that some recordings are poor. I have some recordings where kick drum was clipped.
. Play a sine wave tone in the 200 - 500 Hz range. Turn up the volume until you hear distortion, which is a harsh sound. Back off just enough to make the harsh sound gone. With a true RMS meter set to AC volts, measure the voltage across the output terminals. Square the reading and divide it by the resistance. This will be the Max continuous RMS clean power output of the amplifier. If the amplifier has more than one channel, you should connect a resistor to each channel to see the proper value with all channels driven (don't forget to feed the signal to the input of all channels!)Your ears will be quite sensitive to the onset of clipping at lower frequencies. You may even hear it before you see the tips of the sine wave flatten on a scope. It is harder to hear at higher frequencies.
To evaluate the amplifier at normal listening levels, I play a varied selection of familiar music and simply judge what I hear. No distortion figures needed. If it sounds good to me, that's all that matters. Be aware that some recordings are poor. I have some recordings where kick drum was clipped.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.