HFU HF Underground
Technical Topics => The RF Workbench => Topic started by: digitalmod on January 24, 2018, 2255 UTC
-
:D It would appear some one just is not reading the engineering papers. Class E RF amp was developed at 13 MHz for 200 watts, in 1997 ! Authors use IRF 440 in push pull.
QEX article states class E better above 3 MHz, D below. Also the QEX article states Class E FETS have produced 100 KW transmitters , obviously in tandem in 2001 !
So clearly today in 2018, to ask 200 watts at shortwave a piece of cake,
Ref:
http://www.its.caltech.edu/~mmic/reshpubindex/papers/RFdesign.pdf :-X
-
Note from the paper: They needed water cooling for the MOSFET.
-
Note from the paper: They needed water cooling for the MOSFET.
A minor detail! :D
-
8) I mean is that all you can say, like duh.
The Cal Tech paper, 1997, appears to be somewhat ground breaking in Industry at that time. Therefore the FETS used did require an large drive 10 watts. Pg today is many time less, than ten w. Note the authors state, the high internal temperatures were caused by HIGH drive, ten watt, requirements. Pg = po/pin log 10 = 13 db
The internal structure of the transistors used were NOT designed for high speed switching.
The choice of water cooling was chosen for some reason, not immediately clear, but they state power eff at about 90 percent.
Therefore if YOU disregard drive, 10 watts (quite high), no amp today would require this, the power the water takes out is 20 joules/second or in engineering Parley 20 watts. Water cooling for twenty watts seems an over KILL, but this paper is 20 years OLD. Its interesting this is like a Marconi paper, since Cal Tech would like to believe they were cutting edge.
They suggest such designs may be useable for ham operators. Probably .. but the ground breaking design did have some heat issues to resolve. Also , that ten watts of drive should be added into the twenty watts the E amp cost in switching. So, you're looking at 30 watts of heat. Also Yips, that supply Vdss seems real high. You could hurt yourself!
::)
-
Times and tech have changed, but the same basic rules still apply. Mosfet technologies aside from GaN and SiC have not changed all the radically in the last 20 years. Sine wave drive is an old way of doing things, which is why we see wide adoption of driver IC technology now, fed by TTL level control. You cannot use sine wave drive with the newer SiC devices, due to their extremely limited negative gate swing. Still, at 7 MHz, a 1000pf input cap mosfet with an IC driver, I'm still seeing several watts per device of gate drive input power. The old rules still spply, high power requires high voltage. Some number crunching on a 2.5-5KW solid state SW TX tells me around 550VDC is going to be required for modulation peaks, and that WILL HURT you.
With great power comes great responsibility....
+-RH
-
Yeah, newer devices show huge promise, but I would hate to blow these up at 11 dollars a firecracker.
Was looking at Mouser, SCT 3080KL'
Vdss 1200
Rds 80 mv
Id 31 amps
Pd 165 watts
https://www.mouser.com/ds/2/348/sct3080kl-e-1017519.pdf
-
The Cree version, C2M0080120D is around $16, 2 should handle full boar 1.2KW + Mod with adequate heatsink, 4 should handle around 2.5K. The 1700V devices on paper show promise to just north of 5KW for a quad, but by then your talking serious money @ $90 each.
The smaller brother to the '008 is the C2M016120D. I think these are OK to 500W, but I blew two of them up testing with mod @ 1KW for reasons unknown. Maybe I wasn't driving the gates hard enough, or the output transformer was saturating.
+-RH
-
Red hat.. yes. serious money in a fool's hands where the prototype keeps blowing up. :-\
-
Red hat.. yes. serious money in a fool's hands where the prototype keeps blowing up. :-\
Yup, and in many cases simulation won't help either. When venturing into largely uncharted territory, it's just the cost of doing business.
+-RH
-
You've got to watch those CalTech boys. There's one that lurks here when he's not stalking Belinda Carlisle.
-
Note from the paper: They needed water cooling for the MOSFET.
To be fair, it appears that the use of the water-cooling system is only to achieve max power output and not because the transistor would be damaged with only a heat sink, as the authors mention.
You would not be able to operate this thing without some sort of heatsink because 20 W is far too much for a small package like that to dissipate safely. Reviewing the IRFP440 datasheet says that a reasonable heatsink attached with grease (presumably in still air) would result in ~30-40 degree C die temperature rise, which would not be anywhere close to the damage area the transistor but it would result in an increase in Rds and some other things that are important to power output and power handling. It would still work, but efficiency and power output would suffer. I'm not sure but I have a feeling that water-cooling would not be mandatory here. A reasonable heatsink and a fan to move the air over it might get it pretty close to to his published results.
I would put in some thermal shutdown circuit or VSWR protection or whatever for high(er) VSWR conditions if this were meant for radio and not a CVD plasma machine.
In any case, it was an undergrad thought experiment to see if it could be done. In the real world, I'd probably select a transistor that has a bit more thermal headroom. Technology has moved on a lot 1997 so many options.
-
:D It would appear some one just is not reading the engineering papers. Class E RF amp was developed at 13 MHz for 200 watts, in 1997 ! Authors use IRF 440 in push pull.
QEX article states class E better above 3 MHz, D below. Also the QEX article states Class E FETS have produced 100 KW transmitters , obviously in tandem in 2001 !
So clearly today in 2018, to ask 200 watts at shortwave a piece of cake,
Ref:
http://www.its.caltech.edu/~mmic/reshpubindex/papers/RFdesign.pdf :-X
QEX is a ham rag.... (Its been years since I've seen it). Hams have been using class e for a good while.
Water is a great and cheap way to sink heat.
-
We're all still waiting for you to build something Digiplod.
Waffle waffle, DUH!
How rude.....
Patent goes back to '77 BTW.
Please stop with these pointless posts.
Are you trying to goad someone to build you something...just for you.
Well, wake up bud, it ain't gonna happen.
Str.
p.s. as has been mention in a p.m......
PUT THE BOTTLE DOWN!
-
"PUT THE BOTTLE DOWN! "
HOW RUDE....
No need for personal attacks, Dilly dilly.
rgg