1KV 1ma 1W. Anything low powered will have efficiency issues
At 100mW it's a different animal.
I was suggesting a repurposed CCFL transformer for use as an inductor in a buck reg configuration, because they're cheap, available and wound for high voltage with low distributed capacitance. The normal drive winding addresses 8-20V bulh control. Backwards, like I suggested.
Lowest frequency to deal with other caps in the viscinity, as the original CCFL will have sensibly done.
RL
Mfr Part # Price Mfr Drain to Source Voltage (Vdss) Current
- Continuous Drain (Id) @ 25°C Drive Voltage (Max Rds On, Min Rds On) Rds On (Max) @ Id, Vgs Vgs(th) (Max) @ Id Gate Charge (Qg) (Max) @ Vgs Vgs (Max) Input Capacitance (Ciss) (Max) @ Vds IXTP02N120P 4.31 Littelfuse Inc. 1200 V 200mA (Tc) 10V
75Ohm @ 100mA, 10V 4V @ 100µA 4.7 nC @ 10 V ±20V 104 pF @ 25 V IXTA06N120P-TRL 7.57 Littelfuse Inc. 1200 V 600mA (Tc) 10V 34Ohm @ 300mA, 10V 4V @ 50µA 13.3 nC @ 10 V ±30V 236 pF @ 25 V STF8NK100Z 8.5 STMicro 1000 V 6.5A (Tc) 10V 1.85Ohm @ 3.15A, 10V 4.5V @ 100µA 102 nC @ 10 V ±30V 2180 pF @ 25 V IXTY02N120P 3.07 Littelfuse Inc. 1200 V 200mA (Tc) 10V 75Ohm @ 500mA, 10V 4V @ 100µA 4.7 nC @ 10 V ±20V 104 pF @ 25 V IXTY01N100 5.22 Littelfuse Inc. 1000 V 100mA (Tc) 10V 80Ohm @ 100mA, 10V 4.5V @ 25µA 6.9 nC @ 10 V ±20V 54 pF @ 25 V STF5NK100Z 6.65 STMicro 1000 V 3.5A (Tc) 10V 3.7Ohm @ 1.75A, 10V 4.5V @ 100µA 59 nC @ 10 V ±30V 1154 pF @ 25 V IXTA06N120P 7.63 Littelfuse Inc. 1200 V 600mA (Tc) 10V 32Ohm @ 300mA, 10V 4.5V @ 50µA 13.3 nC @ 10 V ±20V 270 pF @ 25 V IXTP08N100P 3.69 Littelfuse Inc. 1000 V 800mA (Tc) 10V 20Ohm @ 500mA, 10V 4V @ 50µA 11.3 nC @ 10 V ±20V 240 pF @ 25 V STF5N105K5 5.03 STMicro 1050 V 3A (Tc) 10V 3.5Ohm @ 1.5A, 10V 5V @ 100µA 12.5 nC @ 10 V ±30V 210 pF @ 100 V IXTP05N100M 6.49 Littelfuse Inc. 1000 V 700mA (Tc) 10V 17Ohm @ 375mA, 10V 4.5V @ 25µA 7.8 nC @ 10 V ±30V 260 pF @ 25 V STFW2N105K5 4.81 STMicro 1050 V 2A (Tc) 10V 8Ohm @ 750mA, 10V 5V @ 100µA 10 nC @ 10 V 30V 115 pF @ 100 V STF3NK100Z 5.6 STMicro 1000 V 2.5A (Tc) 10V 6Ohm @ 1.25A, 10V 4.5V @ 50µA 18 nC @ 10 V ±30V 601 pF @ 25 V IXTA08N120P 6.49 Littelfuse Inc. 1200 V 800mA (Tc) 10V 25Ohm @ 500mA, 10V 4.5V @ 50µA 14 nC @ 10 V ±20V 333 pF @ 25 V IXTP06N120P 7.6 Littelfuse Inc. 1200 V 600mA (Tc) 10V 32Ohm @ 500mA, 10V 4.5V @ 50µA 13.3 nC @ 10 V ±20V 270 pF @ 25 V IXFP5N100PM 12.68 Littelfuse Inc. 1000 V 2.3A (Tc) 10V 2.8Ohm @ 2.5A, 10V 6V @ 250µA 33.4 nC @ 10 V ±30V 1830 pF @ 25 V DI2A2N100D1K 2.99 Diotec Semi 1000 V 2.2A (Tc) 10V 6.8Ohm @ 1.5A, 10V 4V @ 250µA 25 nC @ 10 V ±25V 510 pF @ 25 V NDFP03N150CG 3.53787 onsemi 1500 V 2.5A (Ta) 10V 10.5Ohm @ 1A, 10V - 34 nC @ 10 V ±30V 650 pF @ 30 V 2SK3746 5.54675 onsemi 1500 V 2A (Ta) 10V 13Ohm @ 1A, 10V - 37.5 nC @ 10 V ±20V 380 pF @ 30 V 2SK2225-80-E#T2 15.85 Renesas 1500 V 2A (Ta) 15V 12Ohm @ 1A, 15V 4V @ 1mA - ±20V 990 pF @ 10 V IXTA05N100HV 6.54 Littelfuse Inc. 1000 V 750mA (Tc) 10V 17Ohm @ 375mA, 10V 4.5V @ 250µA 7.8 nC @ 10 V ±30V 260 pF @ 25 V
Some people have ideas, and like to play with them as a kind of team sport.
Some people don't have ideas and want to club them to death at birth.
The STN0214 is a cute little 1200 volt SOT-223 for 70 cents at 100.
1400 volts in real life.Roger about using search engines to, well, search for things.
Some people have enough ideas that they can choose the ones that they play with. This does develop the skill of picking the less good ideas early, and ditching them before you waste too much time on them. I've had enough goods ideas that I've now got three patents to my name.
I've also had a great many more less good ideas.
I've worked with people who are less discriminating about the ideas that they will run with. Even people who have several patents can become remarkably attached to very silly ideas, and rigorous quality control is necessary to keep them from wasting loads of money. C.A.G.LeMay at EMI Central Research comes to mind. I wasn't able to stop him from wasting money, but I did manage to avoid working on his daft project.
John Larkin doesn't seem to have all that many ideas, and he does feel hurt when one of his occasional inspirations turns out to be less inspired than he likes to think.
If I had thought that he might be helpful, I'd have been more diplomatic.
I need at least 1.7kV. A 1400V part isn't interesting.
I found the Infineon INWH170R450M1 part for which I'm now looking for a Spice model with exactly that kind of search string.
I've been jumping through Infineon's support hoops for a couple of days now, and they haven't yet even admitted that they have a Spice model
They do have an LT Spice model library that works.
John Larkin hasn't paid any attention to the 1.7KV lower limit on the drain to source voltage. His cavalier attitude to absolute maximun ratings is well known, and pretty silly.
I did find the IMWH170R450M1 with a google search. It clearly didn't pick up parts that I would have expected it to.
There was a 4.5VkV $200 part that it should have found, but any part that expensive probably isn't going to be off the shelf or widely advertised, so there is an excuse.
<snipped badly formatted list>
The highest voltage part in there was good for 1.5kV, so not good enough.
Teaching grandmothers to suck eggs.
The latest version of LTSpice seems to want you to use Notepad on the .asc file to modify the transistor symbol. I didn't used to have to do that.
The 1ma referred to the current-carrying constraint on the switching transistor, not the power source. The Baxandall configuration relies on reactive current circulating inside the tank circuit maintain the output voltage, and some of that can flow through the switches.
Got a part number?
Not a good idea. You really want to run a inverter as fast as the core material (and the switches) will allow. This isn't a CCFL application so the stray capacitances are less of a problem.
The piezoelectric transformer is an interesting idea.
Neither is all that efficient.
I can't see how that could work. Charging up lots of capacitor is series, and discharging them in parallel is one mode of current multiplication, but about the only kind of switch that would work would be a reed relay, and they are slow and don't last long when cycled fast.
Dry reeds are good for 10 million closures, mercury-wetted reeds for about 100 million, and neither is all that cheap or compact.
If you used a three-core mains lead with one of the cores connected to Neutral at the plug, one to Live at the plug and one floating, the floating lead would take up a potential somewhere between the other two, the stray capacitances between the leads acting as a capacitive divider.
The Live and Neutral leads are left unconnected at the equipment end and the floating lead is connected to a rectifier system which takes it's earth reference from the equipment being supplied. It would be approximately a constant-current supply, so a capacitor and zener diode would be sufficient to stabilise it at the required low voltage.
The mains lead is acting as a high voltage capacitor but is a lot cheaper and more reliable.
If you need a solid earth on the output, use a four-core cable.
Centuries ago they'd put two transistors in series in the TV horizontal deflection department because a single one didn't have enough Vce max. And a drive transformer with two secondary windings for the base current in each.
The Supertex depletion fets are often used as current limiters. Turns out that a high voltage series string works; the lower current parts get the most voltage and avalanche. No harm done.
I miss the big ole television flyback driver bipolar transistors. They had some nice undocumented features.
A single BU105 was often used. I salvaged some from old televisions on# a scrapheap. John
A ccfl transformer is ideal for the HV step-down application, and dirt cheap is a side benefit.
They often have several windings, which helps build oscillators. More details might involve using a search engine.
I'd expect that the tranny and one transistor and a few passives would make a basic step-down converter without low-side logic to power up. Parts cost could get below $3 for a regulated 1K to 3.3 supply, $2 in quantity.
I can right click any file and edit it with anything I like here. Ususally Notepad++ If search engines are a challenge then I can understand that "open with" might also be.
Bill, the Baxandal cct is great in lower voltage input applications, but the one thing you don't want to do with HV sources is ADD voltage stress.
Hence recommendation for buck-type topologies.
If the Baxabdal could be reconfigured as a half-bridge, it might fly here.
Still don't know what the actual 3V3 power demand is, but if the source is not limited, as you suggest somewhere else, then safe limiting will be an issue.
RL
Others may be refrain from running to your assistance in anticipation of your gracious response.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required