Clever guy John, you know which side the apple's polished.
EDIT - or should I say the turd is polished
Last edited:
Just got back from Los Angeles tonight. ---- I wouldnt expect asymetrical waveforms to have much affect on symetrical topologies using CCS etc. But quasi- and other nonsymetrical topologies should show problems.
Whats the explanation for these results:
Sound.westhost.com/articles/intermodulation.htm
??
THx-RNMarsh
Whats the explanation for these results:
Sound.westhost.com/articles/intermodulation.htm
??
THx-RNMarsh
I know I'm kinda late, but better ever than never, right?
Richard, my hat off to you and Walt Jung for that two part article on capacitors. It is truly enlightning, and such texts appear at the rate of wisdom tooth growth. Bravo!
How I missed it before is truly a mystery to me. Of course, reading it means having to redo some current things a bit, but it's all for the better.
Richard, my hat off to you and Walt Jung for that two part article on capacitors. It is truly enlightning, and such texts appear at the rate of wisdom tooth growth. Bravo!
How I missed it before is truly a mystery to me. Of course, reading it means having to redo some current things a bit, but it's all for the better.
I am not sure - I have measured a very simple 'generic' amplifier. It just has balanced Ic currents in the input LTP and optimally biased output stage. I agree that huge imbalance of input LTP results in sensitivity to asymmetric input waveforms - but we have a serious design flaw then.
This is an interesting read, but the author is obviously wrong when he equates 35% thermal compression with 35% distortion.
Yet, like in resistors, thermal effects might lead to distortion when the signal is of a low enough frequency to lead to thermal modulation. In view of the low heat capacity of voice coils, combined with their large surface area for cooling, this is likely to occur, but I am not aware of any published studies of this effect.
The author correctly identifies current drive as a solution for thermal compression issues, but there are a couple of snags. First, it does does not address the misalignment that occurs in analog xovers as a result of thermal impedance increase. In other words, current drive only works per driver, not per speaker. Second, current drive goes awry at the lowest frequencies, where one (closed enclosure) or two (bass reflex) impedance peaks occur. For a flat response, a driver needs voltage drive in this region. At the same time, it is exactly in this region where current drive would be most useful to counter thermal distortion.
So, in other words, what we would need is a) electronic crossing over, which is what the most sensible ones amongst us probably have already, and b) current drive amplifiers with a FR compensation being the inverse of the impedance peaks mentioned earlier. Might be a nice project once I am done doing what I am doing now.
A lot has been pubished by driver mfr of high power speakers on compression and distortion.
Have you read the book..... "Current-Driving of Loudspeakers" by Esa merilainen
"Eliminating Major Distortion and Interfrence effects by the Physically Correct Operation Method".
first edition 2010.
?
THx-RNMarsh
Last edited:
Nothing strange and nothing to explain. Define your non-linear transfer function and apply your signal as a sum of sin/cos waves, it can be done with trig identities on paper if you wish. Doing a simple case by hand is very enlightening, you will see as Rod says the terms that cancel in the result. This is also useful to see how single side band or quadrature modulation work.
Somehow folks seem confused when superposition does not directly apply, of course if you apply two sine waves separately to a mixer there are no difference frequencies (duh).
Last edited:
Its only problem is that you shouldn't try to lift it, or you might get lumbago.
I don't recall Get lumbago, just his older brother Guy. You know, that new years eve band leader...
Have you tried a zobel at the speaker end to match the cable?
.... how you arrived at 150C,
...
Back in the late 80's, I recall the epoxies I worked with all ran about 117C or so as the start of glass transition. Once that temp is reached, it softens and the coil disassembles. Nowadays, polyimide goes way up there in temp with no strength reduction.
DC in an oven is simple enough, but will never provide measurements consistent with active techniques within a gap. The problem within the gap has to do with the eddy dissipation of the metals in the gap as temperature climbs.. There are no active methods to subtract the dissipation out of the resistance of the wire. It is possible to clamp the high power AC drive signal to a low level testing DC current, but you still have to wait for the mechanical system to stop moving. This limits how close the measurement can get.
I was thinking more along the lines of the test was a constant voltage test, so resistive increases reduced the current pulled. Is the test actually constant power by reading the current and upping the voltage to compensate, or was it SPL at a constant voltage?
jn
It doesn't to me, and Scott stated things clearly enough.
It is one of the reasons that make me think that THD is generally a better indicator of non-linearity than IMD (yes, I know: using contorted and unrealistic methods, you can find cases where IMD is present and HD almost absent, but for the opposite to happen, you do not have to search for impossible examples, they are present everywhere in real life)
In fact, there are cases when this IM products cancellation effect can be quite useful, example here: http://www.diyaudio.com/forums/solid-state/252143-dive-into-past.html#post3839604
Last edited:
Modern drivers do not benefit as much from current drive as some may think.
The current drive erodes basically dynamic unlinearities in the impedance curve.
When BL is symmetrical, so is the suspension and the voice coil is fully compensated with a copper cap there is not much benefit left.
Crappy drivers benefit a lot though.
The current drive erodes basically dynamic unlinearities in the impedance curve.
When BL is symmetrical, so is the suspension and the voice coil is fully compensated with a copper cap there is not much benefit left.
Crappy drivers benefit a lot though.
- Status
- Not open for further replies.