Frequency Response.
-3 dB from reference at 1,000 Hz. Less then 10 Hz to 20,000 Hz.Distortion Testing.
Power output.
Distortion for just detectable clipping on a scope. One channel driven at a time.23.1 watts each channel.
Power output at 1 % distortion as measured on an HP 334A distortion analyzer with a modified Heathkit IG-18 as the signal source.
22.7 L and 23.8 R
THD at Various Power Levels and Frequencies.
As noted above I think half of rated power is too high for a frequency/power response chart. Rather than use a fixed percent of power output I think for home stereo amplifiers 1 watt is a more realistic power level. In a small living room 1 watt is pretty loud and may be enough to get you in the doghouse with your mate. Frequency = 1,000 Hz.
Power (W) Distortion (%) L Distortion (%) R 20 0.384 0.435 10 0.184 0.205 5 0.107 0.085 2 0.040 0.052 1 0.029 0.048 Distortion at various frequencies at 10 watts output.
Note: I would like to use a lower power for this but my history has locked me in to half of rated power. I think a more realistic power level would be 10% to 25% of rated power. I will likely change this in the future.
Frequency (Hz) THD (L) (%) THD (R) (%) 20 2.12 1.4 30 0.93 0.615 40 0.56 0.375 50 0.375 0.28 100 0.158 0.17 1,000 0.192 0.21 2,000 0.32 0.335 5,000 1.17 0.56 7,500 2.5 1.91 10,000 4.2 3.15 15,000 8.03 6.15 20,000 13.9 8.9 Distortion at various frequencies for 1 watt output.
Frequency (Hz) THD (L) (%) THD (R) (%) 20 0.385 0.235 30 0.18 0.11 40 0.12 0.08 50 0.086 0.061 100 0.083 0.054 200 0.082 0.054 500 0.082 0.058 1,000 0.036 0.044 2,000 0.0905 0.086 5,000 0.425 0.34 7,500 0.86 0.67 10,000 1.33 1.02 15,000 2.45 1.77 20,000 3.95 2.39 You will note that the table for 10 watts jumped from 100 Hz to 1000 Hz. When I was running this one I saw something strange. I ran the left channel first and the distortion jumped from 0.083 at 100 Hz to 0.036 at 1000 Hz. I went back and inserted points at 200 and 500 Hz to find when it changed but it didn't, not by a significant amount anyway. The answer wasn't hard to find. The HP distortion analyzer has a filter that rejects line frequencies but it can be switched in at and above 1000 Hz. Switching the filter out when reading the distortion at 1000 Hz gave a reading of 0.082%. That's the other end of taking the power response at 1 watt. The hum begins to effect the readings. BTW, the hum is 70 dB below 1 watt. That's 0.1 microwatts.
Intermodulation (IM) Distortion.
Some say that intermodulation distortion (IMD) is more meaningful than total harmonic distortion (THD). Yet you rarely see IMD spects for amplifiers sold on the commercial market. Why?It's not because IMD is harder to measure than THD, in fact an IMD analyzer is less complex and less fussy to adjust than a THD analyzer. If you look at the test instrument market there are many more THD analyzers than there are IMD analyzers.
About a month ago I found a combination THD and IMD analyzer on eBay but my bank account was a little low at the time and I didn't win it. So I set out to build an IMD analyzer.
The principle is very simple in concept. (Note: The same can be said of a THD analyzer although the implementation is complex enough that hobbyist built analyzers are rare.) IMD analyzers aren't all that hard to build.
The principle goes like this. Generate two signals of 60 and 7,000 Hz and combine them in a 4 to 1 ratio.
I don't know if Europe has a different standard or not. The U.S. standard used to be 60 and 6,000 Hz but it was changed in the 1950s probably due to beats between the one hundredth harmonic of the 60 Hz and 6,000 Hz signals. I would think this effect would be pretty insignificant but it was changed nonetheless. I don't know if Europe uses a standard of 50 Hz and some higher frequency but I suspect they do because it is convenient to get the low frequency from the power line.The lower frequency is given the higher amplitude. The signal is fed into the amplifier and the measuring circuit is connected at the output. The measuring circuit consists of a high-pass filter that removes most of the 60 Hz signal leaving the 7,000 Hz signal. Any nonlinearity in the amplifier will result in the 7,000 Hz signal being amplitude modulated by the 60 Hz signal. Think of the 7,000 Hz signal as being a carrier that is accompanied by two sidebands that are 60 Hz above and below the 7,000 Hz carrier. The modulation percentage is not very much, the less the better.The 7,000 Hz carrier and its sidebands are passed through a linear AM detector and then through a low-pass filter that removes the 7,000 Hz carrier leaving only the 60 Hz modulation. This is then expressed on a meter as percent modulation in the usual sense of 100 % changing the carrier amplitude from 0 to twice its amplitude.
An IMD analyzer is not nearly as fiddly to operate as a THD analyzer. Here are the steps.
- Use the switch on the analyzer that selects 60 Hz only, 7,000 Hz only, and Both, to verify the amplitude ratio is 4 to 1. Set the switch to Both.
- Connect the signal source output terminals to the input of the amplifier.
- If testing a power amplifier, terminate its output with the proper load resistance.
- Connect the output of the amplifier to the input terminals on the IMD analyzer.
- Using the voltmeter function built into the analyzer set the output of the amplifier to produce the desired power output.
- Set the mode switch to calibrate and the range switch to 100 %.
- adjust the calibrate control for a reading of 100 %.
- Set the mode switch to distortion and change the meter to a lower range as necessary.
- Read the distortion on the meter.
Notice there is no frequency dial or null control to adjust. Heath made a very good IMD analyzer which was the model IM 5248. I don't think it sold many copies because I have never seen one on the second hand market. There was an older vacuum tube version that used frequencies of 60 and 6,000 Hz. I don't believe that HP ever made an IMD analyzer. I am working on an analyzer which I will post on this site when finished. Right now it is on two separate breadboards, signal source on one and detector on the other. It works with an external AC voltmeter. The measurements below were taken with the breadboarded version.
Data were not taken at a power of 20 watts because at that level the scope showed clearly that there was clipping of the 7,000 Hz signal that rides on the 60 Hz signal. IMD seems to be approximately 10 times THD. This is the only amplifier I have tested to date so I don't know if this is typical.
Power (W) IMD (%) (L) IMD (%) (R) 10 2.02 2.48 5 1.17 1.18 2 0.635 0.560 1 0.430 0.460 The next step is to get rid of the 6P3S-E tubes and put 7591As in the output. Tim is especially interested in the outcome of this step and so am I.
Introduction, Circuit, and Specifications.
Distortion Testing. (You are here).
Changing Output Tubes.
Or use your "Back" button to return to where you were.
This page last updated July 11, 2013.