This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello Chris,

We really need some more info on you amp to answer some of your questions, but here is some general info.

Oscillation in amps can occur at almost any frequency, from sub-sonic (Tube amps with poor output transfomers and NFB) up to many megahertz. If it's in the audible region, well, you can hear it, at least untill your tweeters pop. I suspect you are concerned with high frequency oscillation though.

Another source of high frequency oscillation is RF pickup, where your cables act as anetnna and the amplifier tries to amplify the signal.

Ideally, beg borrow or steal an oscilloscope, and start probing. Other signs are a distorted/harsh sound, or your output transistors getting warm or hot (or blowing) with no load (assuming it's not a class A amp).

Prevention is ideally in the design. I can't comment unless more info is given, preferably a schematic. Poor PCB layout, and long or unshielded cables don't help either.

This is all relatively common knowledge, other may have more detailed info.

Cheers, Adrian
Adrian, thanks for the reply.

Actually Geoff already warned me about the bad wiring of my 1996 version JLH class A would cause occillation about a month ago. My JLH 1996 is done by hardwire, no PCB.

I just want to learn more about the subject before I re-arrange the heat sinks and re-wire the amp.

The problem I have is the first 15 minutes of switch on when the amp is getting warm/heat up. During this time I can hear high frequency screeches occasionally may be 5 times lasted a second each. But after 15 minutes everything seems ok. What worry me is that I might not hear it if it is still occillating without me knowing.

Yes, when I moved one of the power ground wires from the power supply board to the amp board star point around I can hear the high frequency like turning a radio tuning knob. I have moved it to a position so that it is not producing any noise for now. But the power up occillation still occurs.

I only have two multi-meters and two soldering irons on hand, and my circle of friends only know how to change a light bulb. Any help I can get from this forum and Geoff is highly appreciated.

So, is it the high frequency like the one that rodents can hear but not human causes the destruction of tweeters? Or is it the heat thus generated by the high frequency due to occillation? Or is it the DC voltage that destroy the tweeter? Or is it the combination of all the aboves? Would Mid-range woofers also can be damaged by occillation?

hf oscillation may not be reproducible by the mechanics of the tweeters but is still passed by the crossover, so it can destroy them either through gross power overload or loading that they were not designed to handle.

*if* the only non-signal spectral component is above the audible band, then the low-pass sections of the crossover *should* protect the woofer/midrange.

The symptoms prior to the destruction of one one of my (thankfully cheap) tweeters was a high tone that rapidly increased in frequency, becoming silent, followed by the *pop* of the tweeter self destructing.
Contrary to popular opinion, it's not too much power that kills drivers, it's too little. Too little power leads to clipping; essentially forming square waves out of sine waves. Since the 'duty cycle' of a square wave is going to be on the order of 100%, you've got all the juice available running through the voice coil at all times...pure heat, in other words (as opposed to the RMS value, for instance, which is what's usually quoted for drivers). Since oscillations are often full power, i.e. clipped, they're awfully hard on tweeters.
It doesn't matter whether they're audible or not--what matters most is whether they're clipping. It's not always the case, but it's the usual case. Act accordingly.

Hi again,

If it occurs during the warmup period, then the first suspect would be mechanical expansion, maybe solder joints going dry which would be more possible with a hard wired amplifier. Get a soldering iron and re-do all of the joints with a touch more solder (for the flux).

Interittant problems are very difficlt to track down with a scope, but it is still the best tool to use if you can get one.

Cheers, Adrian
Here is the spec for Instek Oscilloscope:

Vertical System Sensitivity: 5mV~5V/DIV ±5%
Bandwidth: DC (AC10Hz)-20MHz MHz (-3dB)
Horizontal System Sweep Time: 0.1µS~0.5S/DIV ±3%
100ns~50ms/DIV ±5% (x10 MAG)
10ns~50ms/ 8% (x10 MAG)
Power Supply AC 115/230V ±15% connector selectable, 50/60Hz
Accessories Detachable power cord, 2 low capacitance probes, instruction manual.

Is this a suitable one for my diy amp building purpose?
Radioshack has this on sale for $250...

Should I get it?

Please advise,
I would buy a Tek 465b on Ebay instead, you could get a warranted and serviced one for that kind of cash, or take a gamble and spend less. 100MHz, solid quality build, the scope that "everyone" seemed to reccomned when I was looking for one. Avoid the mil-surplus 465M as (apparently) its entirely different scope and not as good. All this is second-hand advice, but I followed it without regrets.


This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.