Amplifier "sample test reviews" using Paul's Multitone Analyzer software

[pma]

Multitone - Loopback Analyzer for Audio

Paul @pkane is doing great job with his software development for audio measurements, more

Multitone - Loopback Analyzer for Audio | Distort documentation

Multitone is currently in beta test. Use at your own risk!

distortaudio.org

Yesterday I have finally downloaded and installed his Multitone - Loopback Analyzer for Audio and this is my first attempt to use it for amplifier testing.

Multitone SW offers incredible range of test signals and methods, from individual sine waves, through IMD signals, TIM tests, multitones, to level sweeps of all of them and frequency sweeps. All we need for amplifier testing is here.

I have quickly prepared the test circuit with Topping D10s DAC as a signal source, my small MOSFET amplifier as a DUT and E1DA Cosmos ADC to capture the signals. I started with spectrum measurements.

Spectrum measurements

1. THD and TD+N at 5W/1kHz/4ohm

2. CCIF 18.5kHz+19.5kHz 1:1 at 5Wpeak/4ohm

3. TIM 3.18kHz square + 15kHz sine 4:1 at 4.5W/4ohm

(BTW it is nice to have this test signal!)

4. 32-tone AP multitone into 4ohm

Level sweeps and frequency sweeps

5. THD and TD+N vs. power at 1kHz/4ohm up to 33W

For this measurement and further level sweeps, I would like to have X-axis calibrated in [W], but I was not able to find out how to do it. Paul @pkane , am I missing something, or is it impossible in MT software?

6. IMD 18.5+19.5kHz and TD+N vs. power into 4ohm up to 10Wpeak

7. IMD 250Hz+8kHz 4:1 and TD+N vs. power into 4ohm up to 10W

8. TIM 3.18kHz square + 15kHz sine 4:1 into 4ohm up to 10W

9. THD and TD+N vs. frequency at 25W/4ohm

Conclusion

Multitone software is a great tool for us amplifier designers. It has incredibly wide set of test signals and useful methods like distortion level and frequency sweeps. On my PC with Win 8.1, the SW is a bit unstable and needs re-loading, but I assume this is because of my historical OS. I would appreciate an option of X-axis calibrated in [V] and [W] for level sweeps, or maybe just an advice .

 

[pkane]

Multitone - Loopback Analyzer for Audio

Paul @pkane is doing great job with his software development for audio measurements, more

Multitone - Loopback Analyzer for Audio | Distort documentation

Multitone is currently in beta test. Use at your own risk!

distortaudio.org

Yesterday I have finally downloaded and installed his Multitone - Loopback Analyzer for Audio and this is my first attempt to use it for amplifier testing.

Multitone SW offers incredible range of test signals and methods, from individual sine waves, through IMD signals, TIM tests, multitones, to level sweeps of all of them and frequency sweeps. All we need for amplifier testing is here.

I have quickly prepared the test circuit with Topping D10s DAC as a signal source, my small MOSFET amplifier as a DUT and E1DA Cosmos ADC to capture the signals. I started with spectrum measurements.

Spectrum measurements

1. THD and TD+N at 5W/1kHz/4ohm
View attachment 262175

2. CCIF 18.5kHz+19.5kHz 1:1 at 5Wpeak/4ohm
View attachment 262176

3. TIM 3.18kHz square + 15kHz sine 4:1 at 4.5W/4ohm
View attachment 262177
(BTW it is nice to have this test signal!)

4. 32-tone AP multitone into 4ohm
View attachment 262178


Level sweeps and frequency sweeps

5. THD and TD+N vs. power at 1kHz/4ohm up to 33W
View attachment 262179
For this measurement and further level sweeps, I would like to have X-axis calibrated in [W], but I was not able to find out how to do it. Paul @pkane , am I missing something, or is it impossible in MT software?

6. IMD 18.5+19.5kHz and TD+N vs. power into 4ohm up to 10Wpeak
View attachment 262181

7. IMD 250Hz+8kHz 4:1 and TD+N vs. power into 4ohm up to 10W
View attachment 262182

8. TIM 3.18kHz square + 15kHz sine 4:1 into 4ohm up to 10W
View attachment 262183

9. THD and TD+N vs. frequency at 25W/4ohm
View attachment 262184


Conclusion

Multitone software is a great tool for us amplifier designers. It has incredibly wide set of test signals and useful methods like distortion level and frequency sweeps. On my PC with Win 8.1, the SW is a bit unstable and needs re-loading, but I assume this is because of my historical OS. I would appreciate an option of X-axis calibrated in [V] and [W] for level sweeps, or maybe just an advice .

Thanks for testing, Pavel! Sweeps are missing V/W/dBW options at the moment. I'll add this ASAP. The option for measuring using these units is already available in the other plots:

 

[DonH56]

Thanks for testing, Pavel! Sweeps are missing V/W/dBW options at the moment. I'll add this ASAP. The option for measuring using these units is already available in the other plots:

View attachment 262202

Note that the ask is for the x-axis, not y-axis, to have those units during a level/power sweep (sanity check).

 

[pma]

Audio Buffer test with Multitone Loopback Analyzer

I am attaching some measurements made on my Audio Buffer https://pmacura.cz/buffer_en.html
In fact they almost do not differ from the DAC - ADC loopback itself. I believe that the notes in images are self-explanatory.

Great tool!

 

[Blumlein 88]

Audio Buffer test with Multitone Loopback Analyzer

I am attaching some measurements made on my Audio Buffer https://pmacura.cz/buffer_en.html
In fact they almost do not differ from the DAC - ADC loopback itself. I believe that the notes in images are self-explanatory.

View attachment 262305

View attachment 262306

View attachment 262307

View attachment 262308

View attachment 262309

View attachment 262310

View attachment 262311

Great tool!

I'm disappointed. The results are not quite perfect. Sigh!

I suppose considering the free software and modest hardware expense it is fairly good.

 

[DonH56]

@pma Is the text at the bottom of the graphs from MT or pasted into the image? It obscures the x-axis label.

 

[pkane]

@pma Is the text at the bottom of the graphs from MT or pasted into the image? It obscures the x-axis label.

That's generated by Multitone. When producing an image, you have a choice to select the resolution of the image. When a low resolution is used, that measurement detail line (it's optional and can be turned off) can overlap other text on the bottom. I can make the text much smaller, though that'll make it harder to read.

 

[DonH56]

That's generated by Multitone. When producing an image, you have a choice to select the resolution of the image. When a low resolution is used, that measurement detail line (it's optional and can be turned off) can overlap other text on the bottom. I can make the text much smaller, though that'll make it harder to read.

Got it, thanks. I'd leave it as-is, given the circumstances and conditions, and doesn't seem like an issue in normal operation. Plus my ability to read the fine print has dwindled greatly over the years...

 

[LTig]

Audio Buffer test with Multitone Loopback Analyzer

I am attaching some measurements made on my Audio Buffer https://pmacura.cz/buffer_en.html
In fact they almost do not differ from the DAC - ADC loopback itself.

That was my first impression as well. Can you show the DAC - ADC loopback results?

 

[pma]

That was my first impression as well. Can you show the DAC - ADC loopback results?

Please note the infamous ESS distortion hump. And, also note that the 20 years old analog design (with output capability of 250mA) is still competent in terms of contemporary DAC parameters. It is still able to follow and disclose DAC nonlinearities.

Max DAC output is <2V and DAC range setting is 4.5V, so the noise is a bit higher. X-Axis is in Generator dBFS.

 

[pma]

I have made a comparison of DAC-Audio Buffer-ADC loop measurements from Multitone SW and REW. It is interesting. Though spectrum measurements are almost equal, including measured values of THD and THD+N from both SW, IMD values and stepped level measurements do differ, maybe different definitions are used to calculate IMD, and TD+N in both software packages. Stepped THD and THD+N/TD+N seem to be almost equal as well. Please see post #4 to compare REW results with previously posted MTA results.

REW measurements

 

[Rja4000]

Audio Buffer test with Multitone Loopback Analyzer

I am attaching some measurements made on my Audio Buffer https://pmacura.cz/buffer_en.html
In fact they almost do not differ from the DAC - ADC loopback itself. I believe that the notes in images are self-explanatory.

View attachment 262305

View attachment 262306

View attachment 262307

View attachment 262308

View attachment 262309

View attachment 262310

View attachment 262311

Great tool!

About TIM, for DIM 100, level of IMD should be relative to the 15kHz level.
Similar to IMD SMPTE, where it's relative to the 7kHz component.

 

[pma]

About TIM, fir DIM 100, level of IMD should be relative to the 15kHz level.
Similar to IMD SMPTE, where it's relative to the 7kHz component.

I know the definitions, but it is up to software authors, @pkane and @JohnPM to clarify the computation methods in their SW. REW has quite detailed manual with definitions, as attached.

BTW, here is the DIN IMD, so relative to 8kHz (I understand your comment, though)

Note: the difference in X-axis is due to different level definition in REW and MTA, just shift REW intermodulation level steps right to 0dBFS.

 

[Rja4000]

it is up to software authors, @pkane and @JohnPM to clarify the computation methods in their SW

Sure.
I'm just writing this because it's everyone's interest to use the same definition. Otherwise we'll get (very) different figures and comparing them will be impossible (or will require an in-depth knowledge few people have).

I think Multitone software is a great opportunity to standardize on a measurement software for "amateur" measurements (complementary to REW, and 1000000 thanks to @pkane for the huge work !), so I'm challenging it to be as accurate as possible.

 

[pma]

Sure.
I'm just writing this because it's everyone's interest to use the same definition. Otherwise we'll get (very) different figures and comparing them will be impossible (or will require an in-depth knowledge few people have).

I think Multitone software is a great opportunity to standardize on a measurement software for "amateur" measurements (and 1000000 thanks to @pkane for the huge work !), so I'm challenging it to be as accurate as possible.

I agree, however, at the moment I think that the REW yields proper results and MTA still needs some improvement, though of course kudos to @pkane for his great work and efforts!

 

[Sokel]

I agree, however, at the moment I think that the REW yields proper results and MTA still needs some improvement, though of course kudos to @pkane for his great work and efforts!

I'm quoting @pkane here:

By the way, you can change the method used by just appending CCIF or DIN at the end of the test signal description. For example:

• FM 1k/200Hz/900Hz DIN
• FM 1k/200Hz/900Hz CCIF
• Multitone 20 DIN
• Multitone 64 CCIF

You just type which you want beside the signal.

 

[LTig]

View attachment 262450
Please note the infamous ESS distortion hump. And, also note that the 20 years old analog design (with output capability of 250mA) is still competent in terms of contemporary DAC parameters. It is still able to follow and disclose DAC nonlinearities.

View attachment 262451
Max DAC output is <2V and DAC range setting is 4.5V, so the noise is a bit higher. X-Axis is in Generator dBFS.

I think you can safely assume that SINAD of your buffer is somewhat higher than the software reports. AFAIR the loopback must have a SINAD of 10 dB or more than the device under test to get meaningful results.

 

[pkane]

About TIM, for DIM 100, level of IMD should be relative to the 15kHz level.
Similar to IMD SMPTE, where it's relative to the 7kHz component.

As @Sokel points out, the IMD calculation can be specified right on the definition of the test signal. Just add SMPTE to the end of the signal text. I’ll take a look at TIM again and change it in MT signal list.

 

[pma]

I think you can safely assume that SINAD of your buffer is somewhat higher than the software reports. AFAIR the loopback must have a SINAD of 10 dB or more than the device under test to get meaningful results.

Yes but it is unimportant, to me. I have 4.5V input set because of other, balanced sources and I do not want to toggle DIP switches all the time. I am concentrated to changes in nonlinearity, noise excluded. Small change in noise is unimportant to me, in this case. Noise only masks the existing nonlinearity, if you stick with SINAD or THD+N. I want to have THD and N separated. You do not see the hump in many reviews here because it is masked by N. It is not masked as IMD hump, though it exists as a a THD hump as well.

 

[pkane]

It is not masked as IMD hump, though it exists as a a THD hump as well.

Not surprising, since harmonic distortion is just a very simple case of IMD, with just one fundamental interacting with the nonlinearity.