help with displacement current, distributed capacitance

Ya, haven't built one yet, every time I'm about to start I go off on a tangent like this. But that's OK, I enjoy this. Mikek

e.

s,

are

Dry air dielectric is practically lossless as long as the voltages are fa r from breakdown (corona can steal a lot of power "invisibly"). If you want to get really picky, use uninsulated silver wire and polish off the oxidat ion every once in a while, but seriously, ordinary magnet wire is a good tr adeoff of price, durability, and maintenance against minimal losses due to the insulation.

Also, your coil former can be an issue:

You want to look at the "loss tangent" values.

Mark L. Fergerson

I'd have to check at work for the smallest wire size in my custom litz, but it's #44, or likely smaller.

It certainly looks fragile. But in my experience, treating it almost like it was ordinary wire (not counting the solder-pot termination), surprisingly, it was very robust! I did not suffer substantial degradation over theoretical values. I'd think dirt or grease would not be a factor. The issue is, how many of those fine wires did you break? Or perhaps fail to tie into the RF connection.

This is a fine model for turn-turn activity, but as many have pointed out, and Knight's 2016 105-page document, for the case of a single-layer coil or solenoid, the transmission-line model gives the same SRF answer. So some of them like to discount the importance of capacitance.

But in my experience, making high-inductance high-frequency air coils, where multi-layer winding was required, a low-capacitance design is critical. My idea to use bank winding came from (or was re-inforced by) reading Terman's discussion in his Radio Engineer's Handbook. Examining multi-layer coils of various types made back in the day, honeycomb and the like, requiring exotic special machines to wind, you will see they all knew this reality. It may be that some of these awesome coil- winding techniques beat bank winding, but bank winding, with the aid of machined, 3D printed, or laser-cut forms, is an easier way to lower the overall capacitance.

Thanks Win, I never heard of bank winding. figure 3-11 here?

And what is Knight's 2016 document? (Can you give me more of a hint to find it? :^)

A Q of ~500 @ 1MHz is impressive. Can I ask what you used for a capacitor?

George H.

Huh, I didn't know that either. (stay below SRF/10) So it's currents "sloshing" around between sections that is causing the extra loss due to inter-winding C.

George H.

We don't know the inductance of Win's coil, I think it is large, so it probably is impressive. However the crystal radio guys have ~200uh coils with Qs near 2000, they even have a ferrite that will get them to a Q of 1400. I'm getting a strong argument against distributed capacitance being the characteristic of importance in the high Q coil, (from the crystal radio group.) Seems to be no belief that 'IF' there are distributed capacitance caused currents, they do not add to the losses of the coil. The thinking is that the special winding styles that reduce distributed capacitance also reduces proximity effects and 'that' is the key characteristic that reduces losses for an increase in Q. If you care to read their thinking see it here, starting about four posts in.

Huh.. OK 'Dat's a pretty high Q too. If they are using this for AM radio isn't a Q of ~1000 going to limit the BW to less than 1 kHz?

Well don't believe everything you read on a news group. :^)

Hmm What's a "proximity effect" and how is it different from distributed capacitance? proximity effect is skin effect?

George H.

Google knows more than I do, but simplistically, the magnetic field from one wire pushes on the the electrons in another wire, this causes more current to flow in only part of the wire. If current flows in only part of the wire it is as if the wire is a smaller diameter. That is the reason for spacing turns, it reduces proximity effect allowing current to flow in more of the wire.

Oh sorry, my bad. I should have googled before asking you.

George H.

Why? was my answer wrong? :-) Mikek

That's a poor version, a workaround to use where you don't have a proper coil form.

That was referenced in Martin Brown's post.

That was the easy part. I used a large 15kV adjustable vacuum capacitor. The coil was tuned with the cap to 1MHz and resonantly servo driven up to 10kV. Thankfully I only needed 50W of amplifier drive power.

Ours run at picowatts!

Win, What was the inductance of your coil?

Mike

Cool!

Cheers

Phil Hobbs

If I understand from my old RIS-330 drawings (1998), it had two modes, f and 2f, selected with Jennings vacuum relays: 1175uH and 290uH. My file is marked 15kV litz coil.

Thanks Mikek

Well, and also reading Knight's paper I see these arguments. Some of the arguments seem to be of the type you have to believe, but can't measure. But they don't float my boat, not just yet.

We know capacitive currents do exist, they can be quite large, and the wire is very lossy at these high frequencies, so it's inevitable there will be losses. These are the same currents that increase approaching SRF, and ultimately totally kill the Q. The circulating currents are highly local, and can exceed the external current by orders of magnitude. I'm sorry, but distributed capacitance coupled with the distributed inductance we've created, creates the sharp dramatic SRF resonance. The sharp SRF can not be explained with proximity-effect losses. So there.

But I will read your linked items in the morning.

Only disagreement is with your term 'sharp' dramatic SRF resonance, as I recall self resonance is much lower Q than than a lower frequency tuned resonance. Adding to the theory that distributed capacitance caused currents are causing additional losses. Mikek

Haiku coils always Have low distributed C Winfield Hill agrees.

Cheers