[Brooko]

Hi all - just need guidance on how to measure the maximum output power (and resultant distortion) of an amp under various loads.
 
Fiio posts measurements like these
 
270 mW (32 Ω/THD+N＜1%)
450 mW (16 Ω/THD+N＜1%)
 
Can anyone explain what equipment is needed, and how (step by step) to go about it + any calculations needed?  Ideally I'd like to measure both output power and distortion.
 
Thanks in advance.

 

[cdsa35000]

Gloogle: diy amplifier power measurement

For hp they use fixed 32/16 Ω load resistor for the measurement.

 

[stv014]

You need these for the measurement:
1. a simulated load (or even real load if you do not mind your headphones playing extremely loud sine waves)
2. a signal generator
3. an analyzer with an input that is suitable for testing the output of the amplifier at its maximum level
4. a digital multimeter that can ideally be used for measuring the output level of the amplifier, or at least to calibrate the analyzer (if it is just a PC sound card and software)
This older thread where some of the potential issues have been discussed may be relevant as well.
 
For the simulated load, you can use a pair of resistors (15 or 33 ohms using standard resistances, with a sufficient power rating for the amplifier), and a splitter that allows for capturing the signal on the resistors while they are connected to the output of the amplifier. This may require some DIY work (soldering).
Additionally, if the expected output voltage of the amplifier is too high for the analyzer (sound card), then a voltage divider needs to be built as well, for example the following circuit (only one channel is shown) simulates a 32 ohms load, and divides the voltage by a factor of 4.235:

Onboard audio inputs tend to clip at 1.4 Vrms, and a sound card like the Xonar Essence STX can handle up to 2 Vrms. With the divider, this is increased to about 5.9 and 8.5 V, respectively.
 
The signal generator can be a sound card output, or even a simple DAP playing a test tone. If accurate distortion and other measurements are not needed, and the only goal is to find the clipping level, then the sound quality is not very important. It just should not clip or have other problems. The test tone can be generated for example as a file with free software like Audacity or my testgen utility, or in real time with SineGen or Visual Analyser.
 
If you do not intend to buy a hardware analyzer (there are some basic models like this one for a few hundred $), which may be overkill for power measurements on a headphone amplifier, then you need to use a sound card or external audio interface, and a software analyzer. The above mentioned (freeware) Visual Analyser is suitable for real-time waveform and FFT display, although it may have some bugs. For the purpose of measuring the maximum power output of amplifiers, it is enough to be able to find the level where clipping begins, and the quality of the A/D converter is not important (as long as it does not clip or is broken in other ways).
 
There are some potential hardware issues when using a sound card input:
- it may not be able to handle the output voltage of the amplifier, resulting in clipping, or in extreme cases hardware damage. In such cases a voltage divider like shown above needs to be built
- ground loops could interfere with accurate measurements (of THD, noise, etc.), although this should not normally have a significant effect on the maximum power output. Grounding can be an issue when more than one device in the entire test setup is grounded, or when the signal generator and analyzer have a common (not isolated) ground, for example because they are both on a sound card in the same PC
- without more extensive DIY work, amplifiers with balanced/bridged outputs cannot be connected to an unbalanced line input (note: some devices that use a virtual ground can also be problematic)
 
Alternatively, if you do not mind finding the clipping level by ear, and do not need accurate results, it is also possible to just use headphones in the place of the analyzer/sound card, with a voltage divider or large serial resistor used to reduce the volume to a more bearable level that is not audibly distorted by the headphones. It may be easier to hear the clipping as IMD, using a 2 tone test signal (e.g. 60 Hz + 3 kHz at a much lower level, even just a few %).
 
The multimeter can be a cheap model, it just needs to be able to measure AC voltage with reasonable accuracy. Note that some of the multimeters are very inaccurate at frequencies much higher than 60 Hz. This can be tested by measuring the output of the signal generator at 60 and then 1000 Hz, if the DMM is accurate at both frequencies, then there should not be a major difference (more than just 1-2%) between the measured voltages. Otherwise, there are a number of possible workarounds:
- test the amplifier at 60 Hz instead of 1000 Hz
- use the DMM at 1000 Hz and correct the voltage by the ratio between the voltages measured at the two frequencies (assuming it at least remains linear at 1000 Hz)
- use the software analyzer to measure the voltage, calibrating it with the DMM at 60 Hz
 
Once you find the RMS voltage where the amplifier begins to clip (i.e. the distortion reaches 1%, although manufacturers often use 10% instead), the power output in W can be calculated easily as V * V / R. If there is no excess gain, it might not clip even at the maximum volume, however.

 

[davidsh]

Lazy as I am I merely used my DMM, measuring voltage of a 1 kHz sine wave from my iPhone by a male to male jack, measuring at all 16 steps on the volume. Did an exponential regression on the data as well.

 

[Brooko]

Thanks gents.  Steve your post was brilliant.  Helped a lot and allowed my friend Martin to get the measurements he was after.

 

[cjl]

Lazy as I am I merely used my DMM, measuring voltage of a 1 kHz sine wave from my iPhone by a male to male jack, measuring at all 16 steps on the volume. Did an exponential regression on the data as well.

You may want to double check those numbers with a 50 or 60Hz sine, since a lot of cheap DMMs expect that their AC mode will be used for measuring power, not audio (so they may be incorrect at high frequencies).

 

[davidsh]

Lazy as I am I merely used my DMM, measuring voltage of a 1 kHz sine wave from my iPhone by a male to male jack, measuring at all 16 steps on the volume. Did an exponential regression on the data as well.

You may want to double check those numbers with a 50 or 60Hz sine, since a lot of cheap DMMs expect that their AC mode will be used for measuring power, not audio (so they may be incorrect at high frequencies).

I am aware of that though I thought it wasn't worth mentioning in my very short reply.

 

[arnyk]

A couple of tips.
 
Many resistors sold for power amp testing are junk. They heat up and their resistance goes up like  rocket. You test is not what it seems because the load is not an 8 ohm resistor, its a 20 ohm resistor now!
 
This includes items sold by people who cater to audio enthusiasts.
 
If you search the web you'll find something that may tax your imagination. Some people use electric water heater elements for testing amps. It works. They are cheap, readily available, take a ton of power, and if you dunk the elements (but not the terminals, natch) in water or oil, their power handling capacity goes waaaay up. Of course they are waterproof. But just sitting on something that can take heat, they may be able to take as many watts as you have.
 
The trick is using Ohm's law to figure out what their resistance is from wattage and voltage. QED or ask.
 
Plan B is to get some REAL non-inductive power re$i$tor$. There are some on eBay, and the big distributors like Newark, Mouser or DigiKey may have them.
 
If you want a relatively inexpensive DVM with great accuracy, durability, and frequency response, check out the Uni-T UT-61 on Amazon or eBay. About $50. It even has an isolated USB interface.
 
When I use a PC to check out power amps (all the time), I use a 5K or 10K 2 watt potentiometer to cut the voltage down to something that audio interfaces are happy with. 
 
There's a freeware program called RMAA or The Audio RightMark that is used by lots of folks to check out sound cards. If you put a power amp between the inputs and outputs of your audio interface and that audio interface is really good (e.g. MAudio AP 24192, etc), then you automatically get a nice tech report on the power amp.
 
Warning about low cost and even mid-priced audio interfaces: Their performance is asymmetrical. The output side is usually the far better part unless both sides are comparable.