• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Headphone Amplifier and Low Power Amplifier Dummy Loads

DualTriode

Addicted to Fun and Learning
Joined
Oct 24, 2019
Messages
893
Likes
593
Hello All,

Now that it has quit raining we have come back here to the sunny north coast of Jefferson where I keep most of my test equipment.

I have been playing with the idea of test loads for lower power amplifiers including headphone amplifiers.

While working with power amplifier test loads it became apparent that the test load resistors can and do originate more distortion than the amplifier. Resistor originated distortion is even more problematic with lower power amplifiers.

A couple of years ago I studied resistor originated distortion with a Wheatstone bridge. Using the Wheatstone bridge and APx555 distortion can be measured down to and somewhat below -170dB’s. The FFT and Wheatstone bridge removes the noise where the distortion is otherwise hidden or well below the limits of the best audio analyzers.

I may be mistaken here, I suspect that the headphone dashboard @amirm uses is a balanced 4 Volts into a 300R load and the load is the 300R internal input impedance of the APx555. That gives us right at 0.053 Watts to give us the headphone dashboard FFT and SINAD.

On my bench today is a Topping A30 Pro headphone amplifier. I tested 3 different 300R loads:

10 Vishay metal film 30R 1/8 Watt 5PPM in series for a total of 300R.

3 Vishay metal film 100R 1/2 Watt RN65E in series for a total of 300R.

3 Vishay wirewound 100R 10 Watt NS-10 in series for a total of 300R.

The 3 load resistors tested much the same with a SINAD of ~120.8dB’s. The FFT’s also looked much the same.

See the attached plots:

A30 Pro Dale NS10 wire wound 3 100R in series FFT Spectrum.PNG
A30PRO 300 R PTF 5PPM FFT Spectrum.png
A30PRO 300 R RN65 25PPM FFT Spectrum.PNG


Coming the middle of next week from Digikey is a package of Caddock MP930 “30Watt” power resistors. Also arriving next week are some good size Avid heat sinks Sil-Pads and such stuff.

I am a bit leery of the thick film Caddock MP930 resistors. Thick film resistors did not test as well as thin film/metal film resistors in my previous tests.

Thanks DT
 
Last edited:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
I think the dummy load issue is not such a problem at these resistances and power levels.
The real challenge is the power dummy loads to test high power speaker amps.
 
OP
D

DualTriode

Addicted to Fun and Learning
Joined
Oct 24, 2019
Messages
893
Likes
593
I think the dummy load issue is not such a problem at these resistances and power levels.
The real challenge is the power dummy loads to test high power speaker amps.

Hello,

The dashboard for headphone amplifiers at 0.050 Watts and the dashboard for power amplifiers is 5 Watts are at opposite ends of the scale.

I won’t argue that selecting a test load for power amplifiers is more difficult or not. You could be right that selecting low distortion load resistors for 5 Watts output may be more difficult than selecting a low distortion test load for 0.0533 Watts output.

So far it looks that headphone and low power amplifiers make less distortion and noise.

I do want to play with Caddock MP930 series TO 220 foot print power resistors. The 30 Watt nominal MP930 resistors are made to attach to heat sinks.

Thermal management may be important to minimizing resistor source distortion. Depends on your theory of resistor originated distortion; there are the absolute value of Voltage model, the Voltage squared model or even the cubic of voltage models. Or in the real world some combination of the above plus some other mechanisms of injury.

Will thermal paste vs. sip-pads or a large cold heat sink make a difference to the measured distortion?

The thermal model is all about rate of temperature change of the heating resistive element. It takes time to transfer heat from the heating resistive element to the resistor case through the sil-pad to the heat sink. The temperature change in not instantaneous, the thermal coefficient of distortion is PPM/degree C.

Will a big heavy cold heat sink cause an attached TO 220 resistor to resist temperature change or cause it to cool faster?

Just for fun

Thanks DT
 

3125b

Major Contributor
Joined
May 18, 2020
Messages
1,352
Likes
2,204
Location
Germany
Interesting thoughts and observations - but do they matter in practice?
I just used cheap 5W resistors (33 and 220 Ohms) as HPA dummy loads and that contraption has no issues clearing what my ADI-2 Pro FS is capable of measuring wich is already far beyond audibility threshold.

What's your point of diminishing returns for this? To be "fair" in your measurements you'd have to let the resistors warm up to tightly controlled working temperature and the lower the delta to ambient temperature (due to heat sinking) the bigger the influence of said ambient temperature wich you would then have to control for. This gets cumbersome fast for decimal points far beyond what anyone could ever hope to hear in practice.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,580
Likes
38,281
Location
Gold Coast, Queensland, Australia
@DualTriode

Amir is using this:
 
OP
D

DualTriode

Addicted to Fun and Learning
Joined
Oct 24, 2019
Messages
893
Likes
593
@DualTriode

Amir is using this:

This looks to be only partially true.

If you look closely at the headphone amplifier test dashboard for Balanced 4 Volts both in and out it looks to be using the $30,000 APx555 internal 300R impedance load resistors.

@amirm is able to confirm this, if he will.

Thanks DT
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,580
Likes
38,281
Location
Gold Coast, Queensland, Australia
This looks to be only partially true.

If you look closely at the headphone amplifier test dashboard for Balanced 4 Volts both in and out it looks to be using the $30,000 APx555 internal 300R impedance load resistors.

@amirm is able to confirm this, if he will.

Thanks DT

1644186008127.png


What's not to understand? The statement is pretty crystal clear. He compared Tom's box to the AP's internal 300R loads, found no appreciable difference and now uses Tom's load box wherever we see such plots.

Tom's load box is brilliant, well made, well priced and saves you trying to re-invent the wheel with bags of resistors, tons of experiments and no automated control of the thing. It's a no-brainer.
 
OP
D

DualTriode

Addicted to Fun and Learning
Joined
Oct 24, 2019
Messages
893
Likes
593
View attachment 184699

What's not to understand? The statement is pretty crystal clear. He compared Tom's box to the AP's internal 300R loads, found no appreciable difference and now uses Tom's load box wherever we see such plots.

Tom's load box is brilliant, well made, well priced and saves you trying to re-invent the wheel with bags of resistors, tons of experiments and no automated control of the thing. It's a no-brainer.

Hello,

Please look at the ASR review of the Topping A30 Pro the same headphone amplifier that I am using. The dashboard FFT, SINAD and other test results for Balanced input and output are taken at 4 Volts or the equivalent of 0.0533 Watts into 300R. Amir’s results are reported as having a SINAD of 121.

Remember, SINAD is THD+N without the minus sign

The plot that you copied and pasted above: at 4 volts on the Horizontal axis none of the curves are near -120dB THD+N.

My conversation here is about thin film / metal film resistors and Caddock thick film MP930 Power resistors. The Caddock thick film MP930 Power resistors may perform better at maximum power output for these low power headphone amplifiers, you know at 4 or 5 watts output.

On the other hand an array of RN65E thin film resistors are likely to do better at the dashboard values of Balanced 4 volts or 0.0533 Watts into 300R.

Just for fun

Thanks DT
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,580
Likes
38,281
Location
Gold Coast, Queensland, Australia
My conversation here is about thin film / metal film resistors and Caddock thick film MP930 Power resistors. The Caddock thick film MP930 Power resistors may perform better at maximum power output for these low power headphone amplifiers, you know at 4 or 5 watts output.

Why not get some input from @tomchr on this? He developed his load box using his AP no doubt. He will know the linearity of the Caddocks at representative typical headphone levels i.e. a few hundred mW absolute max. 4-5 watts is well outside 'headphone' levels and well into low powered speaker level testing. I don't believe there is a dynamic headphone transducer on earth that would be able to dissipate 5 watts on a continuous basis.

The reason I posted the plot was Amir is stating he uses the load box for his testing wherever we see a bunch of load lines as depicted.

As you say, have fun- keep up the good work.

Cheers.
 

charleski

Major Contributor
Joined
Dec 15, 2019
Messages
1,098
Likes
2,239
Location
Manchester UK
The dashboard FFT, SINAD and other test results for Balanced input and output are taken at 4 Volts or the equivalent of 0.0533 Watts into 300R.
The base input impedance of the APx555 is 100kohms. I presume this is used for the dashboard results and the dummy load is only used when it’s needed for the power tests. This is why you’ll get better numbers on the dashboard than on the power sweeps.
 
OP
D

DualTriode

Addicted to Fun and Learning
Joined
Oct 24, 2019
Messages
893
Likes
593
Hello @charleski,

If you go to the APx500 software there is Signal Path Setup => Input Configuration => Connection is Analog Balanced => Termination, there is a choice among, as you call it base impedance, 200K, 600R and 300R.

So a user can select the 300 Ohm internal impedance and use it for a 300R test load.

The internal (base impedance) connection impedance is not fixed at 200K.

Thanks DT
 

charleski

Major Contributor
Joined
Dec 15, 2019
Messages
1,098
Likes
2,239
Location
Manchester UK
Hello @charleski,

If you go to the APx500 software there is Signal Path Setup => Input Configuration => Connection is Analog Balanced => Termination, there is a choice among, as you call it base impedance, 200K, 600R and 300R.

So a user can select the 300 Ohm internal impedance and use it for a 300R test load.

The internal (base impedance) connection impedance is not fixed at 200K.

Thanks DT
Ah, that’s interesting, thanks. I must admit looked at the spec sheet and decided not to wade through the 650 page manual :).
 
OP
D

DualTriode

Addicted to Fun and Learning
Joined
Oct 24, 2019
Messages
893
Likes
593
Hello All,

A couple of days ago a few Caddock MP930 30 Watt resistors and Avvid heatsinks arrived. I ran the same FFT plot and SINAD as with the other resistors over the last few days.

The FFT shows much the same 2nd Harmonic as the previous resistors.

The 3rd Harmonic is a dB or 1,5dB higher

A 4th Harmonic has appeared above the noise floor.

The 5th Harmonic is 2 or 3 dB’s lower than the other resistors.

With all the small changes to the FFT the overall SINAD is the same as the other test load resistors measured at 120.8 dB.

This Caddlock MP930 resistor with small differences in the FFT Plot measures the same in THD+N and SINAD as the other resistors with a balanced 4Volts into 300 Ohms. (Using the Topping A30 Pro HPA.)

DIY Fun

Thanks DT

Topping A30 Pro 4 Volts Caddock MP930 30 Watt plus Aavid heatsink and Sil pad  FFT Spectrum.png
20220210_010530.jpg
 

charleski

Major Contributor
Joined
Dec 15, 2019
Messages
1,098
Likes
2,239
Location
Manchester UK
The 3rd Harmonic is a dB or 1,5dB higher

A 4th Harmonic has appeared above the noise floor.

The 5th Harmonic is 2 or 3 dB’s lower than the other resistors.
I think that getting a meaningful comparison for small variations down at -150dB is going to require precise temperature matching.
 
OP
D

DualTriode

Addicted to Fun and Learning
Joined
Oct 24, 2019
Messages
893
Likes
593
I think that getting a meaningful comparison for small variations down at -150dB is going to require precise temperature matching.

Hello,

Close enough for me.

Sitting here on my bench with the window open with the test load resistors sitting side by side with 192K FFT Length and ten averages there are consistent and reliable differences among the load resistors measured.

Real world without the long length FFT’s and ten averages a few dB’s of the 5th Harmonic at -154 ish dB’s is lost forever in the noise floor.

Thanks DT
 
Top Bottom