I probably need a switch somewhere between the 2SK4177 (too big) and the ZVN4424 (too little).
It's no big deal, just sort of an interesting aside to an otherwise kinda dull and unprofitable project. Maybe I'll go into the high-voltage business, so it won't be a total loss. I wonder if there is a market for kilovolt-range low-watts pulse generators, below the Pockels Cell turf.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I'd probably try a 40T0100 and see what it did. Looking for minimal xsection, so as not to interfere, however this circuit has pretty respectable voltages being forced. If it ended up doing serious work, with an elevated temperature, then switching to a higher curie temp matl (~35) might do a better job. Hopefully there isn't enough energy involved to raise the issue.
Mr. Larkin says it's not an issue and he's looking at a working breadboard.
You can try increasing the inductor current (lower FET sense resistor), and you can try driving the FET with a pulse-gen Zout lower than 50 ohms (add some 50-phm terminators in parallel).
One application for low-watts pulses is fast piezos. You can make them resp ond much faster than their lowest order longitudinal resonance, by giving t hem a pulse train of variable duty cycle. At the leading edge of the first pulse, the part of the piezo nearest the contacts expands, and it stays the same lingth overall as the acoustic pulse propagates.
You have to keep recirculating the pulses, sort of like a mercury delay lin e.
Just make sure the "interpose circuit" is fast and its prop delay doesn't put you outside of the leding-edge blanking (LEB) window.
For some reason I don't see them much out there in the wild.
Our HV stuff is in fab now. So far they haven't complained about the oddly rounded footprints in the HV sections. Should be here on the bench in a couple of weeks. If afterwards I don't post here anymore you know why ...
Looks like there are DPAK mosfets that will work. I'm waiting for the UPS delivery to make sure. I'm probably more Rds-on limited than capacitance limited, according to Spice.
Our production people say they can pick+place them, no worse than a BGA, but I don't know how we could extract any substantial amount of heat from those things. The EPC folks visited us recently; they have some cool parts, including new lower-current, low-capacitance things.
No GaN parts that I know of are avalanche rated; overvoltage on the drain or the gate kills them pronto.
Our mfg people don't mind rounded pads. They sorta like them.
After many hours breadboarding my 1400 volt supply, I haven't been zapped once. Maybe it's not actually working.
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
On Wed, 09 Mar 2016 11:10:06 -0800, Joerg Gave us:
1200V ain't that much unless you are trying to push tens of milliamps.
The solder has to be round blobs too. Those little corners on SMD resistor and capacitor packages are high gradient points. Any thru-hole stuff in the HV area needs a blob of solder on the protruding leads as well.
I am always scared that the thin grinding disc in the Dremel breaks and chunks fly into me. Ever since dad told me how that almost happened to him. Except it was a giant flywheel chunk off of a motor that launched itself through the roof of one factory building, flew for a while and crashed through the roof of another building where he was.
Make sure the 1206 resistors are untrimmed ones and rated high enough. Otherwise there could be St.Elmo's fire where the laser trim cuts are. It would "work" for a while and eventually a stench develops, then phhssss ... PHUT.
Maybe EDM (the kind of machine tool where you WANT the electrical contacts tp erode) - you have an arc-starting pulse (back it up with a lower-voltage capacitor that can keep the spark going for a millisecond).
Good luck finding any tech info on the electrical machinery in an EDM tool, though: I had to wade through translated Russian tech journals, because info is mainly kept behind the curtains. It's a lot like trying to research precious-metals metallurgy...
Everything is rounded in the HV sections and thru-hole parts are far away. Parts are all de-rated properly. With thru-hole it's sometimes good to cut the leads so short that the solder flows over them and covers any pointy ends. Just difficult to inspect.
How long a test? TVs were infamous for corrosion issues in the focus and grid control resistors, and it was about the high failure rates after months to years of operation. Wise resistor manufacturers used silicone paint to good effect. Wise circuit designs targeted local FIELD strength (volts/mm) limiting, not just voltage. That's why spark plug wires are 8mm; it's not the rubber, it's the air around the outside that has a problem.
A brass washer on a thin wire makes the wire less likely to make ozone. It's very unclear how you can replicate that kind of precaution in surface-mount.
If there is a 10% trim, which sounds like a lot, the voltage across the trim gap must be less than 10% of the overall voltage (via armchair thermodynamics) so must be less than 60 volts. That shouldn't arc.
I can see a high-current, low-ohms trimmed resistor having current crowding and hot spots.
If I get into trouble, I can always buy some high-voltage resistors. Some 1206's are rated for 1500 volts.
Here's the IR image, making 1400 volts into 333K. The fet is a Fairchild FDD7N20. The only hot part is the transformer, about 60C. I should heat sink it some on the real PCB layout.
formatting link
formatting link
Efficiency is about 71%.
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
Laser cuts are very thin. No kidding, it can become a problem. I would at least use something like the Rohm KTR series:
formatting link
Higher than 500V I have only seen for non-trimmed.
Something north of it gets hot as well. 60C for a ferrite isn't so bad as long as it isn't a ferrite that goes into inflationary core losses right above that and then commits suicide.
That is decent. It is surprising that 4005 diodes are ok here. And that you got it all on there without leaving even one fingerprint.
The air alone will do that. Whenever I ride in urban areas with my bicycles a lot the chain turns black quickly. When out in the boonies it doesn't. Which makes me wonder what our lungs take in.
Maybe spray it with clear coat so your grandkids can see your work?
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.