Re: Instead scopes

On 3/09/2024 1:45 am, john larkin wrote:

On Mon, 2 Sep 2024 17:52:09 +1000, Bill Sloman <bill.sloman@ieee.org>
wrote:
 

On 2/09/2024 12:49 pm, john larkin wrote:

On Sun, 01 Sep 2024 17:43:32 -0400, Joe Gwinn <joegwinn@comcast.net>
wrote:
>

On Sun, 01 Sep 2024 13:17:03 -0700, john larkin
<jlarkin_highland_tech> wrote:
>

On Sun, 01 Sep 2024 15:53:46 -0400, Joe Gwinn <joegwinn@comcast.net>
wrote:
>

On Sun, 1 Sep 2024 17:55:58 -0000 (UTC), Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:
>

john larkin <jlarkin_highland_tech> wrote:

On Sun, 1 Sep 2024 17:45:46 +1000, Bill Sloman <bill.sloman@ieee.org>
wrote:
>

On 30/08/2024 2:21 am, Jan Panteltje wrote:

On a sunny day (Fri, 30 Aug 2024 00:43:39 +1000) it happened Bill Sloman
<bill.sloman@ieee.org> wrote in <vaq1f2$jdj$1@dont-email.me>:

>
<snip>
>

I'll have someone start on a SolidWorks model.

>
I bet you need the standoff, so the lossy FR4 material isn't too
close.  That should be in the requirements as well.

>
If the FR4 losses matter, the printed circuit board under the coil would
darken, which the customers wouldn't like. The Cambridge Instruments
0.5nsec beam blanker did that so we swapped to a different substrate
that didn't discolour.
>

The turns squish down into the gap-pad gunk, which is an OK heat
conductor. The PCB under the pad is a big copper pour, top and bottom,
with a zillion thermal vias.  There's more gap-pad on the underside of
the board to dump heat into the baseplate.
>
At 4 MHz, skin depth is 32 microns, so most of the copper is wasted.
That's why it gets so hot.

>
It's a sawtooth so it has quite a lot of higher harmonic components with
even thinner skin depths. Baxandall's preference for sine waves has
incidental advantages.
>
The turns are wide and flat, which reduces the effect of skin dept.

I tried three of the Coilcraft 1010VS parts in series, but they
smoked, probably skin+proximity effect.  Maybe parallel would have
been better.

>
https://www.coilcraft.com/getmedia/55a4b40a-2e02-4bf5-b0af-2ea5db75b6cf/1010vs.pdf
>
There are five 1010VS parts, all rated at about 25A rms. You haven't
specified which one you used three of.
>
That 25A rms isn't going to include any allowance for skin effect.
>
They don't look as if there would be much cross-talk from one to the
next. Making space for more parts might have been a better approach.
>
With +/-20% tolerance on inductance, putting them in parallel wouldn't
have been a good idea.

 Don't they have the same tolerance in series?

When they are in series, the same current passes through each one.
When they are in parallel the lower inductance parts would carry up to 40% more current and have to dissipate up to twice as much heat.

In parallel, each would get 1/3 the current. But each would need to be
9x the inductance. I suspect that's a wash, something fundamental
going on.
 My coil opens itself up for a lot of air cooling, and bare copper can
run pretty hot.

And melt the solder holding it in place?

I'd specify the coil dimensions, not the mandrel dimensions, which may
be provided as a helpful suggestion only.

>
I could have a mandrel machined or 3D printed, to more accurately wind
the inductor. The improvement would be mostly cosmetic.
>
Inductors are a pain.

>
Particularly when you don't think about what you doing.

 But it works. A big laser company buys them.
 Why don't you design a 1200 volt, 4 MHz pulse generator and we can
discuss it here.

Why would I bother? I wouldn't get paid for it. More to the point, I'd need a lot more information than 1200V, 4 MHz and 25A peak to peak, and your customer wouldn't be happy if you broadcast that much detail.
--
Bill Sloman, Sydney