Since I am about to build a high power amplifier using a +/-85V DC power supply, I would like to explore what can be done, so that, the critical phase of first time testing is non-destructive and successful. If I remember well, Mooly instructed me to use a filament lamp in series with the toroidal primary. This in the hope of limiting the current the transformer secondaries, thereby, protecting the amplifier's electronics.
So, the first step is to use an incandescent bulb in series with the primary and check whether the output is correctly at 0V and whether the rail voltages are approximately at the levels they are expected to be. One problem that hits my mind, is even with this precaution, the large smoothing capacitors will still be charged at +/-85V, which mean, in the event of a malfunction, a temporary, but sufficiently destructive current can damage components. So, smoothing capacitors have to be scaled down to prevent this from happening.
Any more ideas, I do not want to destroy expensive powerful speakers.
P.S. I tested a prototype pre-driver circuit with the output for several days and no damage to speakers was done.
So, the first step is to use an incandescent bulb in series with the primary and check whether the output is correctly at 0V and whether the rail voltages are approximately at the levels they are expected to be. One problem that hits my mind, is even with this precaution, the large smoothing capacitors will still be charged at +/-85V, which mean, in the event of a malfunction, a temporary, but sufficiently destructive current can damage components. So, smoothing capacitors have to be scaled down to prevent this from happening.
Any more ideas, I do not want to destroy expensive powerful speakers.
P.S. I tested a prototype pre-driver circuit with the output for several days and no damage to speakers was done.
The incandescent light is only to test the amplifier (no speakers should be attached). Once you verify that nothing smokes or blew up, then you start up the amplifier without the lamp and test the bias and verify there is only minimal offset voltage on the speaker terminals (like 10mV or close to it) - still no speakers are attached to the amp.
Only after you have sufficiently verified the amplifier is stable and not producing any bad behaviors, do you ever plug in speakers. Even then, use cheap ones the first time.
- I would also add, I tend to error on safety and include a speaker protector in the case.
Only after you have sufficiently verified the amplifier is stable and not producing any bad behaviors, do you ever plug in speakers. Even then, use cheap ones the first time.
- I would also add, I tend to error on safety and include a speaker protector in the case.
Another pre-test you can do is to put lamps as load for the amp. Having 85Vpeak, 110V lamps will perform pretty fine. I use 3 of them in star when testing industrial frequency inverters, from 220/380 mains. A gate problem in final IGBT stage is easily seen in the lamp sequence/bright, preventin from use large current spikes with inductive motor) loads.
Make sure that the Variac is properly fused on the output, to prevent its failure. Most amplifiers are not designed for continuous full power operation, since music doesn't require that and the cost is much reduced, so test only briefly near full power. Proper load resistors for high power testing can also be a problem. Some use water heater elements, in water for dissipation.
Last edited:
I design quite a few amps so I built my own test rig. The transformer will only supply a couple of amps. All connections are made by flying leads with crocodile clips. On the output is a speaker protector module in its box with its own power supply.
With no speaker connected, I power up the amp with a scope on output. If the output voltage hits power rail or the transformer grumbles I know there is a problem and turn off quick.
With no speaker connected, I power up the amp with a scope on output. If the output voltage hits power rail or the transformer grumbles I know there is a problem and turn off quick.
+1 for the DBT for initial fault testing.
For very little cash, I built one from hardware store items:
Batten mount light fitting, double power switch (main power and lamp bypass), switch mounting block, power extension cable (cut in half), insulated connector, small piece of MDF to mount, scrap metal shield for lamp.
It's all safely insulated and allows remote operation from the DUT so I don't even need to touch it.
Build once in under half hour and it is always there whenever you need it.
I actually test all my amps with speakers using a DBT. Once I verify that it runs into a high Z load and that bias is stable, I connect a speaker and play music thru it. It will play at a low level, and if anything goes wrong the lamp protects both the amp and speaker. Running it this way allows you to check for instabilities and potential latch-up issues that may occur as the supply voltage falls. What you should get is a graceful drop in supply voltage with volume as the lamp heats up. Sudden jumps in brightness or the amp latching to the rail are indications of potential trouble that might not show up without a load. It will also make turn on thumps much worse, allowing you to see and address the problem. In some extreme cases, I’ve seen the turn on transient cause a latch up with the bulb. The same amp might not do it with no load immediately, but could have trouble down the road as caps age.
A custom high power amplifier can easily produce high frequency oscillations superposed on the output signal. These can damage tweeters.
A method may be to connect a very small capacitance in series with the probes of a frequency counter which should, I presume, tell the frequency of such oscillations. Another way is using a small torch incandescent lamp in series with a very small capacitance. If the bulb lights up, it would mean, there are high frequency oscillations.
A method may be to connect a very small capacitance in series with the probes of a frequency counter which should, I presume, tell the frequency of such oscillations. Another way is using a small torch incandescent lamp in series with a very small capacitance. If the bulb lights up, it would mean, there are high frequency oscillations.
If you have HF oscillations, not only does the bulb light up but you don’t get much power out.
One thing you commonly get with amps running off high voltages is breaking in and out of oscillation with a speaker-ish load (but not with a dummy load or no load). The bulb oscillate between dim and bright, making the characteristic “szhmmmp szhmmmp szhmmmp” sound. You often need the speaker to see if an amp will do this, and need to address it if it does. Do it without the bulb and you kiss a speaker and a set of output transistors goodbye. A raw unhoused 10 or 15 inch driver works just fine - can handle any residual garbage that gets thru the limiter.
One thing you commonly get with amps running off high voltages is breaking in and out of oscillation with a speaker-ish load (but not with a dummy load or no load). The bulb oscillate between dim and bright, making the characteristic “szhmmmp szhmmmp szhmmmp” sound. You often need the speaker to see if an amp will do this, and need to address it if it does. Do it without the bulb and you kiss a speaker and a set of output transistors goodbye. A raw unhoused 10 or 15 inch driver works just fine - can handle any residual garbage that gets thru the limiter.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.