I've found that 33 ohms in the gate always works. That may slow down a big part, but a bead will slow it down too.
I like gaasfets; their feedback capacitance is so low that they are, in my experience, always stable if the source is at rf ground.
John
Yeah but let's be honest, the trend is quite clear :-(
Then as engineers we have to put the thinking caps on and figure out the future. Which IMHO will not be hydrogen or ethanol, at least not longterm. We need to do better than that but we are a country of inventors and tinkerers so we'll get it done.
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
No, but for the top :-)
Do you or others think BJTs like the BUX85 are here to stay? Pretty soon I may have to select some HV parts myself. Don't know about the required noise specs yet but could be tight.
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
It's already cheaper to make gasoline from coal than from oil, or to make ethanol.
But the leftist weenies will block it.
Fortunately natural selection will enter in, we'll have a major depression and the leftist weenies, with no one to support them, will die off.
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: "skypeanalog" | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
America: Land of the Free, Because of the Brave
Well, it takes food off the table, not good.
I am not in favor of it either. The world doesn't have the ag resources for it and likely never will.
Wishful thinking?
:-)
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
I used to rely on small gate resistors to solve this problem, but I've run into too many places where it did not, or maybe I should say, there are many MOSFETs that aren't so easily tamed. The ferrite bead I use is Fair-Rite 2673000101, but many others should work. Usually one bead each on the gate and source is sufficient. I have certainly seen many instances where one bead alone (on either the gate or the source) was not sufficient.
One more point, these are high-voltage MOSFETs we're talking about, and when they're seeing more than 20 to 25V drain-source. I haven't seen the problem with say, 55V parts.
Question: What are your favorite HV-FETs du jour? I mean the range of
800V plus. Finally I got the SPICE files from Infineon for their SPD06N80 and its brethren. It's a bit like pulling teeth but this time I received an answer within a week. Still waiting for another one from a couple years ago ;-)While we are at it, any hot leads on BJT? Do you think there is a chance the BUX85 might stick around another 10 years or so?
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
You may have to be more selective, or work with parts that are likely to remain in use.
Think lamp ballast, if you want parts that are dirt cheap.
Might be worth seeing what you can do with (multiple) 20W parts, currently residing in energy efficient substitute lamp assemblies.
You can already get the transformer materials for pennies.
RL
Maybe we need to start questioning the concept of individually shrouding ourselves in tons of plastic and metal for locomotion.
robert
I have about 30 parts in my inventory that I cycle among, depending on the task. Most of them have 1000 to 1200V ratings. I usually use as small a part as possible, to keep the capacitance down. The IRFBF20 is one part I use, having purchased 100 some years back (small-die high- voltage MOSFETs all-too-often become discontinued, and lifetime buys are one tool to combat that problem).
Last year we made 160 amplifiers, each of which had six 1.2kV MOSFETs (in series), I used Fairchild's FQD2N100 in that design. Right now I'm working with an elegant 1.5kV part by ST, the STP4N150.
But that's not to say I don't have a good collection of 600 and 800V MOSFETs. There certainly are plenty of nice ones to chose from in that region. Some would make good choices for fast 600-volt 50-ohm pulse generators, but in those designs I often end up with 12A 1kV parts.
Sorry, I'm not much of a fan of large, slow, low-beta high-voltage BJTs. I know they can be very useful in fluorescent ballast designs, etc. I suppose the days of color-TV HOTs are coming to an end, and obscure companies, like Wing Shing Computer Components in Hong Kong, making the 1.5kV 2SD870, etc., will have to find something else to make. I see they're selling NiMH battery packs and bluetooth headsets!
Thanks Win - I'll take another look and let you know what I find.
Yesterday we were testing the first article of a new 8 channel arbitrary waveform generator. It appeared to have low-frequency channel-channel crosstalk through the power supplies of the output amps; each channel uses a pair of THS3201's. The crosstalk behavior was truly bizarre, between strange combinations of channels, dependent on loading and cable lengths and such. Young Engineer spent some optimization of a couple of Cuk/Sepic switchers (useful detour) but things just got stranger. Old Engineer starting poking things with fingers, and it was obvious that something was happening that the 250 MHz scope couldn't see. So we cranled up the spectrum analyzer and, lo, a rainforest full of birdies around 700 MHz. A crosscheck with the trusty 7104 (1 GHz analog scope) showed multiple oscillations. We have invented the 16-opamp distributed RF oscillator.
John
Thanks, but those are a bit slow and hardly have any gain.
Yes, transformer materials have become cheap.
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
Thanks, I look into the ST part. I've also checked IRFGB etc. but availability is one of the concerns since HV designs often have very long production lifetimes. And lifetime component buys are very frowned upon by the CFO. Also by the VP of Sales (what if we land a windfall account and then we don't have enough stuff to fill the order?).
Using small parts is smart, I also like to go as high in Rdson as I can without the dissipation increasing too much.
That's exactly what I am afraid will happen. H-flybacks going lalaland and only the slow ballast BJTs with poor current gain remain.
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
A client once blew the EMC. So I flew out there. After having to go to their hardware store and place chicken wire over the Agilent analyzers noisy LCD (!) I found a small stand of redwoods near 260MHz. Came right out of a SMPS module. Not el cheapo but one of the major brands. Called them up. "Oh really? We'll check and call you right back". Ring, ring ... ring, ring. "Well, ahm, what can I say? S..t!" They had never noticed that the feddback stuff was singing.
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
Late at night, by candle light, Jim Thompson penned this immortal opus:
Never miss a chance to jump on your hobby horse, do you? And you have the gall to say you don't start this sort of shit, only reply to it.
- YD.
-- Remove HAT if replying by mail.
Same ft, slower turn-on, faster turn-off.
You'd likely see that gain is similar at similar collector currents and voltages.
I had thought you were doing some dickey hv linear that needed low output capacitance. None of these parts has much in the way of SOA.
RL
Surely it wasn't the LCD - but the backlight or something else behind it?
LCD's mean big panel holes that aren't part of an effective screen.
I had to use mesh screen over a COTS combined touch-panel and lcd display. The noise source was actually a crystal oscillator module buried deep in the guts of the beast.
RL
Well, yeah, it probably was some stuff from the innards carried onto the LCD traces. But I didn't have time for that and didn't want to do Agilent's job on my client's time so I just bought some spark screen or whatever it was during the lunch break. That "fixed" it.
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
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.