Re: tiny dc/dc

On Sat, 29 Mar 2025 20:22:48 -0000 (UTC), Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:
 

john larkin <jlArbor.com> wrote:

On Sat, 29 Mar 2025 16:30:23 -0000 (UTC), Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:
 

john larkin <jl@glen--canyon.com> wrote:

On Fri, 28 Mar 2025 16:49:23 -0400 (EDT), Martin Rid
<martin_riddle@verison.net> wrote:
 

john larkin <jl@glen--canyon.com> Wrote in message:r

Check out UCC33420. It's a tiny cheap isolated dc/dc converter. It
switches at 64 MHz!

 
Why does it have such poor esd ratings?
The ucc33421 is better, but not 15kv.
Cheers

 
Right now I dodn't need kilovolts of isolation. I want to make a GaN
totem-pole driver and want a floating power supply for the high side,
but I want very low switching noise for low jitter on the rising edge.
 
I might try one of these with a bunch of added filtering. It might be
better than some dc/dc that works in the 100 KHz sort of ballpark.
 
Or it might be much worse.

 
I bet the gain bandwidth of a GaN FET in its linear range at high current
is pretty impressive.
 
I?m using SiGe BJTs to speed up the edges of my TDR pulse generators.
Driven from a 500-ps comparator edge, it?s fast enough to oscillate on the
falling edge if the layout isn?t right.  That?ll make really entertaining
jitter.
 
Cheers ?
 
Phil Hobbs

 
One part that I like is the EPC2037, the tiny BGA thing. I think you
have used it too.
 
Transconductance is about 1.5S and it's on hard by 3 volts on the
gate. Capacitances are absurdly low compared to a silicon mosfet...
reverse transfer is 0.1 pF.
 
But a little noise on the gate drive will surely jitter the output.
 
I never had much luck getting bipolars to make fast output edges, but
I haven't tried SiGe.

 
I get about 30-40 ps falling edges, which are all that matter for TDR.  The
rising edges aren’t nearly as good, because I’m saturating the poor thing
completely.  It’s a 65 GHz transistor, though, so even an edge like that
has room for a snivet.
 

 
That tiny EPC thing is rated for 100 volts and can conduct 2 amps!

 
Yup, a very nice part. I use it in my ultrafast temperature controller,
which I call a thermal Faraday shield. ?? (Details forthcoming when the
patent gets filed.)
 
Cheers
 
Phil Hobbs

 
I use the 2037 and 2038 in the output stage of our DDG
 
https://highlandtechnology.com/Product/P500
 
which was a lot of work.