John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
The buzz is not from being loose, but saturation is a strong possibility. 200mA should not be passing through ground at all in any situation, so gluing the ferrites together would be rather missing the point. I will be fixing that, although in a way it was a useful test of my signal isolator.

There might be saturation, but not clear how that would produce a noise. More likely, it could be the wire moving relative to the ferrite, or magnetostriction of the ferrite itself. In either case, it is an inadvertent loudspeaker. That suggests a fair amount of current.
 
RNMarsh said:
You did not ask me. You assume too much and jump to conclusions... a fad here, it seems. That was not the reason I did it. I did it to contain the fields and lower the Ls.
My apologies. I'm curious: why choose 10 cables?

scott wurcer said:
I don't see that nor do I quite get your point. The impulse response contains all the information at all frequencies, the two halves of your sentence seem contradictory.
I know that. You know that. Mr. Neutron, it seems, does not. He will almost certainly deny this, but his claim that you have to take account of impulses bouncing back and forth along an audio cable (rather than merely calculating the response for audio frequencies) is equivalent to saying that the impulse response and the audio response give different results. I seem to recall that you produced some graphs demonstrating that the impulse response converges to the audio response?

We've been over this before, the characteristic impedance graphs are fairly useless at tiny fractions of a wavelength but characteristic impedance and speed of propagation are intimately tied to L and C.
And R and G. At audio frequencies they affect characteristic impedance and speed of propagation. At RF frequencies (and, perhaps, pulses?) they affect attenuation. The two apparently different effects must give rise to the same result, when correctly taken account of. Of course, when speaking of pulses one must remember that a pulse has a wide spectrum so it may be incorrect to model a pulse using purely the RF transmission line model because part of the pulse's energy will be at low frequencies.

This is not a coincidence so all forms of analysis must converge to the same answer unless incorrect values are used.
Exactly.

Despite jn's denials, this looks like the same argument we had before. I don't intend to re-run it.
 
Why does it'd go kaput at wavelengths smaller than the distance between your ears? That is inconsistent with the data. Heck, even Nordmark had 1.2 uSec at 12 kHz.

John

ITD does no longer function at those short wavelengths because ambiguities start to occur when the period of the sound is about twice the maximum possible ITD, i.e. when the period is about 1380uSec and the frequency is around 725 Hz, see Moore "introduction tot the psychology of hearing". Etc.
 
what ever that means. The data you are curious about is not going to be told. Just wondering what idea you might come up with for possible solutions. The HF all concentrated at the edge.....is the point. Consider the return a large ground plane under the plate (insulated by thin insulator. How to normalize the current density across the plate conductor. Brain storm.. I am sure you can come up with a few ideas even with limited data given.


THx-RNMarsh
I already gave you three hints.
1. Modify the conductivity of the plate in the direction of the current, wire EDM
slots is one method.
2. Modify the dielectric environment such that dielectric charging currents fight the propagation.
3. Modify the return current path such that the lower impedance path is not the outer edges of the plate.

If it is a national lab project, have their people contact my people. The clearance people can intermediate if it is required, I'm not cleared.

John
 
My experience with skin effect is high speed digital and mainly high power high voltage SMPS, where we quite often use transformers litzs wound or copper foil wound, stranded or solid copper are no different, in stranded it is a bit worse because the outer layer in misshapen due to the strands. I was reading a paper on this recently that detailed this, I cant remember whether it was put up on this thread or from my own collection. I will try an dig it out.

Edit.
Reading Jneutrons response it seems I am misinformed, though I believe the difference is small, I will double my efforts to find the paper now.
Edit Edit
I also don't get to play with super conductors. But I would be interested in how big a difference there is between stranded and solid and whether it is of concern to us mere non-superconducting mortals...😀

This has some nice pictures of proximity effect, in my case its usually traces on a PCB where the current density is greatest at the edge of the trace that faces the signal return path.

http://www.generalcable.co.nz/getat...d3/AC-Resistance,-Skin-Proximity-Effects.aspx

Edit Edit Edit
Bit of searching regarding stranded vs solid, my initial view was rather simplistic, but we don't often go to the lengths that JN does, but this link explains the difference... I am now curious how JN alters the interstrand conductivity, larger strands with more air gap?

https://www.st-andrews.ac.uk/~www_pa/Scots_Guide/audio/skineffect/page1.html
You are not misinformed, the diff solid/strand is not much.
Sullivan of Dartmouth has some good write ups on the diff for amps apps.
A super is never naked. They are stabilized with copper or aluminum. For copper, they start with a tube of copper, rods of NbTi inside tube, then draw it down. Typically a foot diameter, drawn to .013 to .015 inch final. Aluminum, it can be melted over the super, it's only 600 plus C. The resistivity of the copper at 4.5K is what matters, this is defined by the RRR. An RRR of 100 means the copper will be 100 times more conductive than it's room temp value in LH2. Purity and annealing alter the RRR.
There have been efforts to coat the strands, or oxidize them, or tin them, but those methods are never very well controlled.

John
 
My apologies. I'm curious: why choose 10 cables?


I know that. You know that. Mr. Neutron, it seems, does not. He will almost certainly deny this, but his claim that you have to take account of impulses bouncing back and forth along an audio cable (rather than merely calculating the response for audio frequencies) is equivalent to saying that the impulse response and the audio response give different results. I seem to recall that you produced some graphs demonstrating that the impulse response converges to the audio response?


And R and G. At audio frequencies they affect characteristic impedance and speed of propagation. At RF frequencies (and, perhaps, pulses?) they affect attenuation. The two apparently different effects must give rise to the same result, when correctly taken account of. Of course, when speaking of pulses one must remember that a pulse has a wide spectrum so it may be incorrect to model a pulse using purely the RF transmission line model because part of the pulse's energy will be at low frequencies.


Exactly.

Despite jn's denials, this looks like the same argument we had before. I don't intend to re-run it.
Again, your arguments are disingenuous, you dance around reality in audio.

As to your straw man, I am the one who has always stated equivalence between lumped and t-line.
As you state, it is better if you discontinue discussion as you continue to dance away from the science aspects.

As I said, look up state variable and control theory.

John
 
You are not misinformed, the diff solid/strand is not much.
Sullivan of Dartmouth has some good write ups on the diff for amps apps.
A super is never naked. They are stabilized with copper or aluminum. For copper, they start with a tube of copper, rods of NbTi inside tube, then draw it down. Typically a foot diameter, drawn to .013 to .015 inch final. Aluminum, it can be melted over the super, it's only 600 plus C. The resistivity of the copper at 4.5K is what matters, this is defined by the RRR. An RRR of 100 means the copper will be 100 times more conductive than it's room temp value in LH2. Purity and annealing alter the RRR.
There have been efforts to coat the strands, or oxidize them, or tin them, but those methods are never very well controlled.

John

Cheers.
 
'Kiddie' in this sense does not mean that the speakers cannot sound pretty good. If I were to compare and specify a 'kiddie' sports car, it might be the Austin Healey Sprite, of yesteryear. I had one, and it was a lot of fun. I now drive a medium grade Porsche, and there is a difference.

I'm one up on you, John. Just got back an hour ago from the hosptal, where my spine rings Nos. 3 an 4 got their bore refurbished, which are supposed to be 11-13 mm caliber, but were only 7 mm and inhibited nerves controlling both my legs, which is what with side effects of my Diabetes II prevented me from walking in general, more of a sort of dragging along. The upshot is that I walk MUCH easier and with far less pain, so in a sense, I have been tuned back to normal. I'm not quite up to my Sprite specs yet, but with some physical therapy I am expected to make a more or less full recovery. Consequemtly, I will not be driving a mid sized Porsche yet. 😀 But I will be up and about once again. 😎

If proof is expected, anyone expecting it will have to dish out US$ 150 for an "after" magnetic resonance shot, to compare with the "before" shot. :smash:
 
ITD does no longer function at those short wavelengths because ambiguities start to occur when the period of the sound is about twice the maximum possible ITD, i.e. when the period is about 1380uSec and the frequency is around 725 Hz, see Moore "introduction tot the psychology of hearing". Etc.

You said period twice the maximum possible ITD, then period of 1.4 mSec...

We are not talking about the same thing. You are three orders of magnitude too slow with respect to actual measured, published, and verified human capabilities.

I suspect it is a definition of terms thing.

John
 
Impressed 0.075mm approx. hmm quite small.
Smallest milling bit I ever used was 0.5mm, machining one of the fancy engineering plastics (can't remember which one) got some grey hairs that day (machine was running about 60,000 rpm, very very very slow plunge and traverse rate lots of isopropanol based coolant).
Smallest PCB hole I have had to have mechanically drilled was 0.2mm.

Been looking at copper bus bar profiles to reduce skin effect and provide more surface area for cooling, some interesting stuff.
 
Status
Not open for further replies.