• Welcome to The Audio Annex! If you have any trouble logging in or signing up, please contact 'admin - at - theaudioannex.com'. Enjoy!
  • HTTPS (secure web browser connection) has been enabled - just add "https://" to the start of the URL in your address bar, e.g. "https://theaudioannex.com/forum/"
  • Congratulations! If you're seeing this notice, it means you're connected to the new server. Go ahead and post as usual, enjoy!
  • I've just upgraded the forum software to Xenforo 2.0. Please let me know if you have any problems with it. I'm still working on installing styles... coming soon.

How to build amp measurement gear?

PaulyT

PaulyT

Behind the Curtain
Staff member
Administrator
Moderator
Superstar
I would like to do basic measurements (like distortion) on an amplifier. How do I approach constructing some sort of interface that can serve as a dummy load for the amp, but also provide a signal that I can hook back in e.g. to the mic-in or line-in of a sound card (like the MobilePre)? I assume it's mainly a matter of an appropriate network of resistors, to provide both the simulated load and to scale down the voltage and prevent excessive current flow through the measurement device? I'm not sure how to figure out the resistances needed, though... Anyone know of a reference somewhere that I could use to build something like this?

My first attempt at this would be with a headphone amp, though eventually it'd be interesting to try on a full-size speaker amp as well.
 
Well, first you need a load resistor for the output of the amp. Digikey has some 16 ohm, 50 watt resistors which you can put in parallel to get an 8 ohm 100 watt resistor. Here is the page for the resistor:

The resistors need to be mounted to a heatsink or large metal plate - the larger the better. The resistors will overheat without being mounted to a heatsink. This will not be necessary to test a low wattage headphone or SET amp however since the resistor won't get very hot.

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=RHRD-16-ND

Then you need a 10k ohm potentiometer of any wattage. Wire the input terminal of the pot to a 100k ohm resistor (any wattage), and then the resistor to the hot side of the load resistor and the ground side of the pot to the ground side of the load resistor. Then take the middle (wiper) terminal to the input of your soundcard. This gives you 20dB of voltage attenuation on top of the range of the pot. You adjust the output of the pot to a reference level on your test program (it should be as high as possible but not clipping the input of your sound card).

You can test amplifiers up to 100 watts output with this arrangement. If you're testing a very low wattage amplifier, you may want to substitute the 100K ohm resistor with something lower, like 10k ohms so that you can get sufficient voltage to the input of your sound card.
 
Ok cool, thanks, that's pretty close to what I was thinking, but the variable pot is a good idea. Is it important to match the load resistance to the amp? That is, the headphones I normally use are more like 25-50ohm impedance; low impedance for headphones, but higher than speakers. Though I guess I would really need to know the output impedance of the amp, rather than the headphones?

The headphone amps I've got are a 6AS7G-based tube amp that probably does something like 500-800mW, and a hybrid tube (12AU7) preamp / ss power amp that's probably closer to ~2W.

FYI, the reason I'm interested in doing this in the first place, is that in past months I haven't been using headphones much, but recently started again (changes in home office situation, long story). When I recently went back to listening to that hybrid amp (with planar headphones), I noticed some fatigue, sounds like a bite in the HF that I don't much like. I switched back to my tube-only amp (with regular dynamic headphones) for everyday listening. While the difference between these two rigs is obviously not just the amp but the headphones as well, I'm curious to see if the amps have measurably different distortion patterns. Plus I've never done amp measurements, so thought that would be kinda fun to explore in itself.

Of course, I will also test swapping the headphones between the two amps... but the efficiencies of the two headphones are very different, so the comparison is not as straightforward as it might seem; the tube amp plainly under-drives the planar headphones, and the hybrid amp is clearly overpowered for the dynamic headphones.
 
Pauly, did you hear something last few days? Because I was just doing this over the weekend. I measured couple of my amps using bias voltage potentiometer (100 Ohm, 2 watt) I happened to have. I adjusted it to get just under 1V, then I plugged it into my Mobile Pre like SH explained and applied it to RTA program. I found it to be a very crude tool for measuring amps. It's not very precise. I also tried ARTA program which you can download for free. It has little better program for measuring harmonic signature of amps but still, the combination of Mobile Pre and Behringer mic seem to reach the limit rather quickly.

For the resistors, I just connected it to woofer of my spare speaker when measuring tube amp (not necessary for ss amp) and had the potentiometer connected parallel. You will hear the tone when you send the signal to sound card but you can use earplug if it bothers you.
 
DIYer said:
For the resistors, I just connected it to woofer of my spare speaker when measuring tube amp (not necessary for ss amp) and had the potentiometer connected parallel. You will hear the tone when you send the signal to sound card but you can use earplug if it bothers you.
Click to expand...

This is not recommended because the impedance of the speaker changes with frequency, therefore you will never know the actual wattage being dissipated. Since a speaker load is an impedance, calculating the actual power dissipated is next to impossible due to the phase angle of the current to the voltage. Sophisticated test gear can measure true power in an impedance, but that is not the case here.

Not to mention the fact that a loudspeaker load is generating a back EMF which can raise havoc with the measurement.

Use a load resistor.
 
DIYer said:
Pauly, did you hear something last few days? Because I was just doing this over the weekend. I measured couple of my amps using bias voltage potentiometer (100 Ohm, 2 watt) I happened to have. I adjusted it to get just under 1V, then I plugged it into my Mobile Pre like SH explained and applied it to RTA program. I found it to be a very crude tool for measuring amps. It's not very precise. I also tried ARTA program which you can download for free. It has little better program for measuring harmonic signature of amps but still, the combination of Mobile Pre and Behringer mic seem to reach the limit rather quickly.
Click to expand...

Wait, what does the mic have to do with it? I'm confused. And was it "not precise" maybe because the A/D conversion of the MobilePre is not fine enough (44/48kHz sampling), or the S/N was too low to measure distortion peaks, or...?
 
DIYer said:
But if I only played 1KHz, wouldn't that be consistent for each measurement?
Click to expand...

Just do it right....use a load resistor like everybody else who tests audio gear does.

Trust me on this one.

Trust me on this one.
 
soundhound said:
Just do it right....use a load resistor like everybody else who tests audio gear does.
Click to expand...
It was my make shift dummy load at the time. I plan on getting the right resistors for the real one.
 
DIYer said:
soundhound said:
Just do it right....use a load resistor like everybody else who tests audio gear does.
Click to expand...
It was my make shift dummy load at the time. I plan on getting the right resistors for the real one.
Click to expand...

The biggest problem with measuring into an impedance is that the readings like distortion cannot be counted on to be accurate, since it is dependent on the damping factor of the amplifier and the nature of the complex impedance of the speaker load. A transistor amp with a high damping factor will measure differently than an SET amplifier which has a relatively low damping factor to deal with the back EMF from the speaker driver. Using a resistor which is devoid of impedance solves this problem (although there is a very small residual impedance in wire wound resistors but it isn't enough to worry about).
 
soundhound said:
Digikey has some 16 ohm, 50 watt resistors which you can put in parallel to get an 8 ohm 100 watt resistor.
Click to expand...
What happens if I use 4 Ohm or 2 Ohm resistor?
 
nuclear-explosion.jpg
 
DIYer said:
soundhound said:
Digikey has some 16 ohm, 50 watt resistors which you can put in parallel to get an 8 ohm 100 watt resistor.
Click to expand...
What happens if I use 4 Ohm or 2 Ohm resistor?
Click to expand...

If the amp is rated for the lower load, then the amp will attempt to dissipate more power into the load.
 
Thanks for the feedback. The reason I asked is, I have one 2 Ohm resistor rated at 20w and a bunch of 1.5K Ohm resistors rated at 10w. IOW, with spare resistors I have at the moment, I can either make a 2 Ohm dummy load or 160 Ohm dummy load (1.5KOhm resistors parallel). Would 160 Ohm dummy load cause too much heat if used for 7w SET tube amp?
 
DIYer said:
Thanks for the feedback. The reason I asked is, I have one 2 Ohm resistor rated at 20w and a bunch of 1.5K Ohm resistors rated at 10w. IOW, with spare resistors I have at the moment, I can either make a 2 Ohm dummy load or 160 Ohm dummy load (1.5KOhm resistors parallel). Would 160 Ohm dummy load cause too much heat if used for 7w SET tube amp?
Click to expand...

Why on earth would you want to load it with 160 ohms? If you load the amplifier with anything other than what the speaker load will be, the measurements will be basically useless.
 
You must log in or register to reply here.
Back
Top