Naive question: is there any correlation between the voltage and current noise of these devices?
Looks impressive until one contemplates about where the peak undistorted and reference level signals from a MM or MC cartridge would appear.
Some place around -60 dBv depending, methinks. Even lower for the reference level from a low output MC cartrdige.
Stuart, you can find expression for correlation coefficient in Leach paper
Usually correlation between Vn and In is neglected
Usually correlation between Vn and In is neglected
Generally, that's how I treat it in my design work and articles. But I was curious if, at least in this particular case, one might run a cross correlation between the two noise sources to confirm that they are entirely uncorrelated and thus add as power. Alternatively, one can do what I usually do (lazy) and predict total noise assuming zero correlation and compare that to experimental total noise measurement.
Aha, nice work. Superglue though? I'm not sure what the glass transition temperature is for superglue. Given my history with tweeters, I go for the highest transition epoxy I can find. West was the last one I used, lasted a few decades. Used the tropical catalyst though, didn't know how long it would take for me to wind.. And I pay very close attention to the ratio...no uncured resin, no uncommitted hardener for me, I want max strength with no flexibility.
It was good therapy wasn't it? Sometimes I find this kind of stuff best at keeping my mind from wandering.
Or how about "bars where the people meet". Never knew you were a Debby Harry fan. 😀
Nice, thanks. Looks like they see where skin depth rears it's ugly head.
I didn't review it all, I assume they discussed bright vs matte finish also..
John
JN,
The voice coil is covered with ferrofluid, so the temperature shouldn't be much of an issue. The failure mode was that the old coil was not centered in the gap and one turn on each was not in the fluid and that is where the failures were.
Marce,
Had a bit of fun learning a few bits. The vias in pads turned out to work very well with hand soldering. They create a bit of a bump and that helps to position parts. Of course it would be very nasty if the parts were wave soldered. The other bit was that all of the components were magnetic. As I could not find my beryllium copper tweezers that was a bit of an issue. Finally the input and output traces could have used a bit more spacing. The nice rows of parts made hand placement and soldering much easier. In the build I had two bad open joints and one bridged short. All easily found and fixed.
Now there is one more bit I want to check out about the influence of the power supply on the circuit and should have those measurements tomorrow.
The voice coil is covered with ferrofluid, so the temperature shouldn't be much of an issue. The failure mode was that the old coil was not centered in the gap and one turn on each was not in the fluid and that is where the failures were.
Marce,
Had a bit of fun learning a few bits. The vias in pads turned out to work very well with hand soldering. They create a bit of a bump and that helps to position parts. Of course it would be very nasty if the parts were wave soldered. The other bit was that all of the components were magnetic. As I could not find my beryllium copper tweezers that was a bit of an issue. Finally the input and output traces could have used a bit more spacing. The nice rows of parts made hand placement and soldering much easier. In the build I had two bad open joints and one bridged short. All easily found and fixed.
Now there is one more bit I want to check out about the influence of the power supply on the circuit and should have those measurements tomorrow.
I don't understand how "not centered" causes one turn on each to leave the fluid.
How did the coils fail? Typically, the ones I've fixed had flexural fatigue failure where they leave the coil pack to travel out of the gap, and it was always the return outer layer end wire, not the start lead.
When the wire leaves the wind and is unsupported, but still within the gap, any clipping signal generating U/S harmonics causes the wire to resonate and fail as it generates side forces but is not clamped.
John
The one turn not in the fluid completely melted leaving little balls of copper. The rest of the coil showed no damage. it was the same for both coils. Nothing came apart. It was clear what was in the fluid and what was not by the staining of the insulation.
Speakers have been working in the office now for a few days, no problems.
Speakers have been working in the office now for a few days, no problems.
For me, the brute force method is ok. It's predictable, with low cost, easy to get parts,
has nice PSSR, needs no matched resistors if I insist in some PSSR.
Well, there is a point of diminishing return. My personal boredom limit was 20 op amps.
I have some 3602 on my Mouser list, but Digikey happens to have more of the stuff I
really need, so that may take some time for the critical mass.
Hey, I'm trying the same game with oscillators, also. 16 pcs. 5 MHz SC cut MTI-260
sloooowly phaselocked to one master, and the 16 outputs Wilkinson-combined
to a common coax. That should give pretty low phase noise. 🙂
Well, that statement about low cost does not really apply here.
regards, Gerhard
Too close to the diaphragm. Really left nice copper bits still glued in place.
Does that really work? It seems like it should since the signal is in phase and the noise is random but its really expensive AND the phase lock's noise may be a limiter. Must be the only way to make something more expensive than a BVA if it works.
Just thinking of the design challenge, the massive parallelism is not really practical as an IC or modular product.
I just saw that oscillator, MTI 260, on eBay for a really good price. Not so crazy an idea after all.
🙂 Yep for solder paste the vias in pads would have to be filled and plated... or just go laser drilled vias... Trying to shake small SMD components of tweezers is a fun hobby. Glad it all worked.
I almost forgot to mention my luck with surface mount film capacitors ain't so good. A bit too long with the soldering iron and they change values. Have to check each one after soldering. Screw up a few every time. The ceramics don't have the problem.
I am fast enough with the tweezers that surface mount cards actually go together faster than leaded ones!
Gerhard,
Thanks for the comments. I am really impressed with your low phase noise oscillator idea. My suspicion is that when combined there will be a very low frequency AM component related to settling time. So it may be advantageous to mix up the phase lock time constants if possible.
But I certainly defer to you on these issues.
Scott,
As the low noise dual Interfet FET is $48 your version would certainly cost more. What effect would the input capacitance have on such a design.
The other interesting question is what effect does positive feedback have on noise?
ES
I just have images of you now going 'who's a good little low noise amplifier' to see if that makes it quieter 😛
Already said, not much of a design challenge. You need some context on the positive feedback issue. In general there is no free lunch.
Positive feedback increases gain, so go figure what effect that has on noise.
- Status
- Not open for further replies.