Let's Talk About Amplifier Clipping.
- Thread starter ken kantor
- Start date
ken kantor
Fowl Humor
Wow... I'll add that to my list of theories... didn't know that amplifiers had Napolean complexes...
This all ignores the (sometimes pretty terrible) behavior of some (more than most people think) amps at clipping. Especially those, with very high levels of feedback, and complex output architectures (think Class H, and such).
I've seen waveforms, in a couple of cases where inductive output filter networks, interacted with the transfer function of a feedback loop, once the amp clipped. The result was HUGE SPIKES of energy, at the leading (and sometimes trailing) edges of a flat-top waveform.
Given, especially before the amplifier industry truly had a handle on low-IMD and amplifier stability in high-feedback amps, that an amp that behaved like this, would almost be GUARANTEED to blow up a tweeter... this is PRIMARILY where the recommendation of NEVER clipping an amp came from. And, where the notion that "tube amps don't blow up tweeters" came from- most tube amps have very low levels of feedback, and therefore, pretty "benign" clipping behavior.
Even now- I've seen professional mixers, equalizers, and other processors, with HORRENDOUS looking waveforms at clipping. Even relatively modern stuff. And the resultant is blown horn drivers. I've replaced enough of them, over and over, to see the end results, for people who don't listen to us folks who try to get them to properly gain-structure their systems...
So, yeah- if you want to oversimplify the situation with graphs and calculations of "pure" square waves- go ahead. Those of us who live in the real world, know that's not all that's going on, in MANY cases...
Regards,
Gordon.
I've seen waveforms, in a couple of cases where inductive output filter networks, interacted with the transfer function of a feedback loop, once the amp clipped. The result was HUGE SPIKES of energy, at the leading (and sometimes trailing) edges of a flat-top waveform.
Given, especially before the amplifier industry truly had a handle on low-IMD and amplifier stability in high-feedback amps, that an amp that behaved like this, would almost be GUARANTEED to blow up a tweeter... this is PRIMARILY where the recommendation of NEVER clipping an amp came from. And, where the notion that "tube amps don't blow up tweeters" came from- most tube amps have very low levels of feedback, and therefore, pretty "benign" clipping behavior.
Even now- I've seen professional mixers, equalizers, and other processors, with HORRENDOUS looking waveforms at clipping. Even relatively modern stuff. And the resultant is blown horn drivers. I've replaced enough of them, over and over, to see the end results, for people who don't listen to us folks who try to get them to properly gain-structure their systems...
So, yeah- if you want to oversimplify the situation with graphs and calculations of "pure" square waves- go ahead. Those of us who live in the real world, know that's not all that's going on, in MANY cases...
Regards,
Gordon.
ken kantor
Fowl Humor
Very true. In college I built a Dynaco 410 that used to FREAK OUT LOUD when overdriven into my LST's autotransformer. (Perhaps the xformer saturated.) Still, for all the noise, I never blew a driver that way, even with 300+ W/ch on tap. I suspect the events were just too short. Blew several dozen drivers playing the speakers too loud, though. (I'd hitch a ride out to AR, where the service guys would hand me new tweeters and a few spares, for free! Great customer service in those days.)
Later in life, I had a test rig that would put one-shot adjustable rectangular pulses up to 2,500W into drivers to test for weaknesses in the wiring. The 1" fabric dome that NHT used, (which cost us ~US$4.50), could easily handle 2.5KW "square" pulses up to a few milliseconds. No magic there... it was too quick for even the tinsel to heat up and fuse, unless there was a bad joint, or crimped wire, or big bubbles in the tinsel glue.
-k
We used a capacitive discharge system to test for weak spots in the coil/varnish. Usually the test was done on bare coils. It was pretty cool to hit a big coil with a pulse and have it center over the metal supports under the wooden workbench. It was cooler to see one fail. On occasion I ran it on a fully built driver. They get a little jumpy.
keilau
AK Member
A very informative thread
I was led to this thread by the recent post on"Bostom is bad for speakers". I went through the whole thing in one setting. It is very refreshing to have the argument based on sound technical knowledge and mutual respect, spiced with a good dose of humor.
The atmosphere here is so different from another forum specicalized in speaker. I went there to ask some questions about crossover optimization. Before I can finish my questions, I was shouted at to go learn more first. The biggest offense was that I said that polyster fiber behaves very differently than long fiber wool in a transmission line cabinet.
We are indeed very fortunate to count Ken Kantor as a forum member. But there are many other contributors too.
I was led to this thread by the recent post on"Bostom is bad for speakers". I went through the whole thing in one setting. It is very refreshing to have the argument based on sound technical knowledge and mutual respect, spiced with a good dose of humor.
The atmosphere here is so different from another forum specicalized in speaker. I went there to ask some questions about crossover optimization. Before I can finish my questions, I was shouted at to go learn more first. The biggest offense was that I said that polyster fiber behaves very differently than long fiber wool in a transmission line cabinet.
We are indeed very fortunate to count Ken Kantor as a forum member. But there are many other contributors too.