I'd like to do some more thorough testing of repaired amplifiers. What are people using as standard practices when confirming an amp is good to go after a repair?
Currently I'm limited by a 13.5v 21A power supply, some low wattage bookshelf speakers, and a couple 100w 4ohm wire wound resistors.
Currently I'm limited by a 13.5v 21A power supply, some low wattage bookshelf speakers, and a couple 100w 4ohm wire wound resistors.
Sure, but let's say I grab a larger supply or battery, what would be a sufficient test procedure? eg Run at full power just below clipping into a dummy load for X amount of time?
For most class B or AB amps, I typically ran them until they went into thermal protection.
For class D amps that won't run into thermal protection, I ran them for 3 hours driven with music into a dummy load with a resistance 2x the lowest rated impedance for the amp.
Batteries aren't typically well suited for this because they become drained fairly quickly. You'll likely have to use modifies server power supplies for this type of testing. I had multiple 5v supplies in series but the more common 12v server supplies are as good or better.
For the test, I ran music, up to and intermittently into clipping. This song on a loop (below). I monitored at a level to where I could hear it at a very low level (8 ohm speaker in series wth a 40w incandescent lamp, across the dummy load) while monitoring it on the scope as well.
Not many people will test this way but I had a lot of other things to do and the time testing wasn't an issue.
For class D amps that won't run into thermal protection, I ran them for 3 hours driven with music into a dummy load with a resistance 2x the lowest rated impedance for the amp.
Batteries aren't typically well suited for this because they become drained fairly quickly. You'll likely have to use modifies server power supplies for this type of testing. I had multiple 5v supplies in series but the more common 12v server supplies are as good or better.
For the test, I ran music, up to and intermittently into clipping. This song on a loop (below). I monitored at a level to where I could hear it at a very low level (8 ohm speaker in series wth a 40w incandescent lamp, across the dummy load) while monitoring it on the scope as well.
Not many people will test this way but I had a lot of other things to do and the time testing wasn't an issue.
Hi Perry,
Do you drive class AB amps at full power to try to reach thermal? Minimum load? Sine waves or music?
In the class D scenario above, at what power level do you drive them at when at .5*min load?
In either scenario above, have you ever blown output stages or power supply sections in an amp? If so, then what? Reevaluate your test levels?
I now have huge load banks and a huge power supply at my disposal, but always feel a bit leery pushing amps for extended periods; especially into resistive loads with tones instead of music.
I'm usually just characterizing thdn/output power up to about 3% with my Dscope III. Then I'll run them at half power with tones until they heat up awhile.
Thoughts?
Non-class D:
Music, up to intermittent clipping (light clipping on all bass notes), into a dummy load that 2x the minimum rated load (4 ohms for a 2 ohm amp), generally until the amp goes into thermal protection.
Class D:
Much the same but generally time limited since most won't go into thermal protection.
It's important to not have the bass boosted too much (I use 0 boost with the song above) because it seriously reduces the stress on the amp if it only has to produce significant power on the notes that correspond to the boost frequency.
I don't think many amps would drive their minimum rated load into a dummy load with a sine wave at the threshold of clipping.
There haven't been many times when outputs failed but other problems have shown up. There are cheap amps that I would not take in for repair that may have failed but most amps can survive the tests above and give me confidence in the repair.
If you have a FLIR imager, that can help catch problems for amps that do (or may) fail the burnin. (look for one or more components of parallel groups that run hotter than the others in the group).
I use music up to clipping because it forces the output to the limits (drive into clipping for non-class D and full duty cycle for class D).
Music, up to intermittent clipping (light clipping on all bass notes), into a dummy load that 2x the minimum rated load (4 ohms for a 2 ohm amp), generally until the amp goes into thermal protection.
Class D:
Much the same but generally time limited since most won't go into thermal protection.
It's important to not have the bass boosted too much (I use 0 boost with the song above) because it seriously reduces the stress on the amp if it only has to produce significant power on the notes that correspond to the boost frequency.
I don't think many amps would drive their minimum rated load into a dummy load with a sine wave at the threshold of clipping.
There haven't been many times when outputs failed but other problems have shown up. There are cheap amps that I would not take in for repair that may have failed but most amps can survive the tests above and give me confidence in the repair.
If you have a FLIR imager, that can help catch problems for amps that do (or may) fail the burnin. (look for one or more components of parallel groups that run hotter than the others in the group).
I use music up to clipping because it forces the output to the limits (drive into clipping for non-class D and full duty cycle for class D).
You're probably not going to be able to drive the amp more than 75w/ch with 20A of.current at 14V depending on how efficient the amp is. A typical class AB car amp will be 50-60% efficient at best.
Once the amp is powered up, check individual bias currents on each emitter resistor at the output transistors. They should be even within about 10-15% on average. Once the amp is stabilized and bias is steady, skowly.turn up the levels until the amp clips, then back off a little. Clipping should be symmetrical without overshoot or ringing, which may indicate a tendency for oscillation if excessive.
Let it run at full power for a few minutes and continue to monitor temps of all output devices along with switching fets at the power supply side.
Once the amp is powered up, check individual bias currents on each emitter resistor at the output transistors. They should be even within about 10-15% on average. Once the amp is stabilized and bias is steady, skowly.turn up the levels until the amp clips, then back off a little. Clipping should be symmetrical without overshoot or ringing, which may indicate a tendency for oscillation if excessive.
Let it run at full power for a few minutes and continue to monitor temps of all output devices along with switching fets at the power supply side.
HI, usually, after making a repair (both CLASS AB and CLASS D) I push them to their maximum capacity on the bench with a minimum load supported by the amplifier.
IMHO, if the amplifier sustains the minimum load that can be supported for 30 seconds or even more sessions separated by a short pause in which you pull the maximum it can (observing a minimum of distortion with the oscilloscope) then the amplifier will can consider in 95% of cases, perfectly repaired and working.
Seriously consider that what you make the amplifier do on the test bench will never happen in the car, where the natural raising of the real impedance is king.
If during the bench tests you manage to get the amplifier to warm up quite a bit, you will also get the result of an amplifier that, despite heating up, works perfectly.
At this link, you can find my youtube channel, where I repair and test most of the amplifiers that are entrusted to me, 95% of the time with crazy and reliable results.
https://www.youtube.com/channel/UClecSKJA1P2qj71lrxgxW5Q
Take a look.
IMHO, if the amplifier sustains the minimum load that can be supported for 30 seconds or even more sessions separated by a short pause in which you pull the maximum it can (observing a minimum of distortion with the oscilloscope) then the amplifier will can consider in 95% of cases, perfectly repaired and working.
Seriously consider that what you make the amplifier do on the test bench will never happen in the car, where the natural raising of the real impedance is king.
If during the bench tests you manage to get the amplifier to warm up quite a bit, you will also get the result of an amplifier that, despite heating up, works perfectly.
At this link, you can find my youtube channel, where I repair and test most of the amplifiers that are entrusted to me, 95% of the time with crazy and reliable results.
https://www.youtube.com/channel/UClecSKJA1P2qj71lrxgxW5Q
Take a look.
A company I worked for long ago did similar tests as this one to get big Crown K2's (among others) on their knees.
https://www.prosoundweb.com/real-world-testing-matching-amplifier-data-and-specifications/
They succeeded rather quickly and I was told that later K2 models sported a bigger PSU toroid to overcome that problem.
Hugo
https://www.prosoundweb.com/real-world-testing-matching-amplifier-data-and-specifications/
They succeeded rather quickly and I was told that later K2 models sported a bigger PSU toroid to overcome that problem.
Hugo
If you need a ton of 12v power for cheap, you can always spend $20-25 on ebay for a Delta 1300w server power supply like this: https://www.ebay.com/itm/284845028077?epid=1601413175&hash=item42521632ed:g:dfkAAOSw1P9inlbn
Oops. $28. Bidenflation strikes again.
Outputs 12v:
63.3 amps @ 120v
106.5 amps @ 240v
Considering it's a server power supply it's probably got a screamer of a fan, so you might want to change it out. Delta makes great power supplies, so it should last.
Oops. $28. Bidenflation strikes again.
Outputs 12v:
63.3 amps @ 120v
106.5 amps @ 240v
Considering it's a server power supply it's probably got a screamer of a fan, so you might want to change it out. Delta makes great power supplies, so it should last.
I spent a few years using large server power supplies for testing amps. I now believe that lab power supplies with variable voltage AND current limit are ideal.
One can set the current limit low as he tests and amp,and if things are ok, gradually increase it. Of course, they are very very expensive, but they last for a lifetime, it is practically impossible to burn them.
One can set the current limit low as he tests and amp,and if things are ok, gradually increase it. Of course, they are very very expensive, but they last for a lifetime, it is practically impossible to burn them.
I mentioned the one i did specifically because it was the one used by Derek (Big D Wiz) from Williston Audio Labs in a video of his a long time ago demonstrating how it's done. I'm sure there are better options today. I'm sure there are controller boards out there you can use with this kind of power supply that allows you to adjust voltage as well as current limits, though i'd imagine they'd get quite expensive. From what i've seen though, those server PSU's are the most economical method of getting 60+ amps of 12v.
My only concern with using a server PSU, or a converted ATX, or a high current MeanWell or other similar supply... is that they may get worn out or damaged because of the power requirements of car audio amplifiers. This is all speculation though, as i am a self taught dummy with zero formal EE training who happens to be pretty good at following directions and is decent at soldering.
Yes, they are nice. I have a 0-80v 0-60a variable power supply made by B&K Precision that i got from from a company going out of business that a friend worked for and I was worried i might have overpaid, spending $400 for a used piece of gear. When i looked it up i nearly fell out of my chair. The machine was over $3000 when it was new and the one i got was hardly used, if at all.
I built a huge supply late last year, works great.
https://www.diyaudio.com/community/threads/high-current-supply-for-testing.391159/
However, I wouldn't dream of using for troubleshooting... Ha.
Have had a tp3005 for quite awhile for that purpose. I'm tired of it's limitations and basic functionality though.
Now building a rd6018 for that purpose with the custom code, should be sweet. May not be quite as clean as the tp3005, but I don't need single digit mv noise anyway..
https://www.diyaudio.com/community/threads/high-current-supply-for-testing.391159/
However, I wouldn't dream of using for troubleshooting... Ha.
Have had a tp3005 for quite awhile for that purpose. I'm tired of it's limitations and basic functionality though.
Now building a rd6018 for that purpose with the custom code, should be sweet. May not be quite as clean as the tp3005, but I don't need single digit mv noise anyway..
I have a 12.6v 100a server power supply, converted for this purpose, but always reluctant to use it because of its “tiny” factor, doubting wether it can even supply half of that current cleanly and reliable, even tho that is its main function (PCs being more susceptible to spikes, etc) but after finally digging in and reading posts such as this I think it’s about time I actually use it on the regular versus pulling the amp out to my “test vehicle” and running it there 😅 I know there’s no real substitute for in-vehicle testing for what I do, but still an at desk solution would be dream right now with my workload, lol. And so many links to build something that’s even better, ty guys 🤙🏻