At the time I discovered the acid flux, I was in the doghouse and Dave McCorkle had produced a couple of multimedia amplifier designs for Compaq. In one of them he rediscovered an approach for producing a center-tapped power supply equivalent using a single a.c. adapter (just a plug-in transformer) that allowed the two loudspeakers to be d.c. coupled to the power amp outputs. We thought it was clever, and it seemed to work (although a cost analysis later suggested that it really didn't save all that much or perform a lot better than a conventional single-ended cap-coupled arrangement).
But a tech reported problems, and I inspected. As one channel was driven at a very low frequency, or merely drifted a bit, the other loudspeaker's motion mirrored the driven channel! I began to wonder if this was a flaw inherent in the design, and consulted McCorkle. To his great credit he asked about the possibility of flux contamination, which led us to do a thorough aqueous as well as nonpolar solvent washing and drying of the board. His guess was right, and the problem went away. But I'd never seen anything like such conductivity---I estimated it to be of the order of a 10k resistor between points A and B. Finally I spotted the roll of solder, which at first glance didn't look any different than rosin core solder.
I was unpopular that day (not as much as when I pleaded for the music to be turned down, as I couldn't concentrate) as I went round and confiscated a number of rolls of the acid flux product, which people admitted they'd stolen from the speaker lab.
The HF knife that was cuttin through my eardrums was not directly the speakers. I have heard varying MBL systems 5 times now. Twice vinyl sourced, once RtR sourced, once unknown and this last time server sourced. Never on previous listenings did it drill HF like this last time.
MBL is just one example though.
What is confusing me is the use of the term stiction.
The definition of that used here has to do with static friction and the ways around it. In my ultrasonic welding work, a low level of U/S energy will overcome the static friction (stiction), and we end up with a virtually frictionless system.
In a speaker, I do not see where there is static friction to overcome. Spiders and surrounds will be linear for small signal motion...if you reduce the drive by a factor of 10, the motion reduces that much, down basically to brownian motion. The use of "stiction" implies that at some small level, there will be hysteresis. Yet the materials at that level are linear.
So I'm confused..😕
John
Having said that I still have a suspicion about low-level "granularities"---but unsubstantiated, and I don't want to start sounding like Hawksford 🙂
EDIT: I guess the test would be if a voice coil current could be so small as to produce no change in cone position whatsoever.
EDIT: I guess the test would be if a voice coil current could be so small as to produce no change in cone position whatsoever.
Last edited:
This beings to mind a piece of test equipment I was redesigning for someone. It was (or will be as the guy got two days into ordering parts and complained that it was taking too much time [hoping that I would say Well then I'll do it, which I did not] and thus the PCB remains unstuffed a few years later).
The device is supposed to allow one to see the d.c. voltage across a voice coil using a small probe current in the presence of high-level a.c. excitation. I pleaded for the use of film capacitors to a.c.-couple the external amplifier to the loudspeaker under test, to no avail, as he had already purchased a bunch of electrolytics which were arranged back-to-back, and the film caps were slightly more expensive. I invested in a hundred 10uF film caps and placed them in parallel for a net 1mF. I saw a tiny d.c. offset in the absence of the probe current and puzzled over it. I eventually made the various low pass filter stages lower and lower in distortion, and bit by bit the offset went away. So it was residual second harmonic, mostly.
The device is supposed to allow one to see the d.c. voltage across a voice coil using a small probe current in the presence of high-level a.c. excitation. I pleaded for the use of film capacitors to a.c.-couple the external amplifier to the loudspeaker under test, to no avail, as he had already purchased a bunch of electrolytics which were arranged back-to-back, and the film caps were slightly more expensive. I invested in a hundred 10uF film caps and placed them in parallel for a net 1mF. I saw a tiny d.c. offset in the absence of the probe current and puzzled over it. I eventually made the various low pass filter stages lower and lower in distortion, and bit by bit the offset went away. So it was residual second harmonic, mostly.
Polymers can exhibit a hysteresis loop, Stuart must know more about it. Of course hysteresis is not stiction, but some of that may be there too.
Edit: isn't stiction the same as 'static hysteresis' but pretty much the same mechanism? Not my field.
Last edited:
Nonlinearity in suspension != stiction. I think this is the fundamental premise. Doesn't invalidate the complexity, but at least refines what we're talking about.
You'll lose that one, Mettler makes force balance scales using essentially a voice coil they are SOTA, and close to my heart LIGO is essentially a mechanical device.
For that matter I think an important distinction is homogeneous materials vs a structure of differing materials and possibly a poorly defined interface. Dropping a pin onto a 10kG load cell and realizing the 1/2 thick aluminum structure is bending in response very linearly is an eye opener.
Last edited:
Can and do. Spiders are a bit more complicated, of course, being composites. But no common polymers I'm aware of have a discontinuity about zero strain.
I would like to see real numbers for real audio instances of "Barkhausen Noise"
undoubtedly a real physical effect but people keep using authorative voice claiming it limits loudspeaker motor circuit, XO inductor, transformer performance
I even have the van der Veen papers
undoubtedly a real physical effect but people keep using authorative voice claiming it limits loudspeaker motor circuit, XO inductor, transformer performance
I even have the van der Veen papers
Of course, demonstrating it is an easy science fair project nothing subtle here. I think the usual technique is operative here, if it's operating principle is magnetic it must have Barkhausen noise. Take a microphone or MC input transformer and terminate one side with the appropriate resistor, jee wiz all you see at the other terminals is the transformed Johnson noise of the resistor.
Last edited:
Interesting.
If I do the numbers on the Kii three, including the fancy cabinet, DIY may be running its course.
I've distantly enjoyed a playful tormenting of a friend who hates solid state and loves circuits of tubes, inductors, and transformers (and Rs and Cs). He didn't know about Barkhausen noise. I just wanted to instill some doubt that eliminating sand state would prima facie ensure perfection.
Before he got even more intransigent about transistors, he was an apostle of simplicity at all costs, and said an MC stepup circuit he built sounded better than a Klyne packaged pre-pre. His circuit consisted of a 2SK170, a drain resistor, and a battery. He asked me why that wasn't wholly sufficient. I have forgotten how many reasons I gave him.
New experience today, on the punch list was replacing the covers on the outdoor outlets. The last one had a hornets nest behind it, and I got the classic assault of angry hornets. Good thing I was not too sensitive, the folk remedy of ammonia worked as I grabbed a bottle of Wisk.
Somebody in the disk drive business told me they use mono-crystalline ferrites in the disk heads so there are no Barkhausen domains, because the noise was killing them.
Yuk. Ammonia neutralizes the formic acid in the sting, so works out pretty well. Baking soda also is effective.
Good thing you're not too allergic.
A discontinuity about the resting position would manifest similarly to crossover distortion in an amplifier, no? (Further proof this isn't the case)
Good thing you're not too allergic.
A discontinuity about the resting position would manifest similarly to crossover distortion in an amplifier, no? (Further proof this isn't the case)
- Status
- Not open for further replies.