HFU HF Underground
Technical Topics => The RF Workbench => Topic started by: digitalmod on January 13, 2018, 1502 UTC
-
It amazes me how this rather simple designs are kept so secret. I suppose it the cult of the elite Ham Radio AM-er?
But, equations are straight forward. Parts? Well, would be nice if somebody could cave brew a 200 watt one either kit or assembles, because these kittens are actually safer to use and yes, simpler in design than old tube type rigs.
The huge variable caps can be replaced by door knobs or other high HV caps. At 41 meters, a 150 pf 3 to 5 kv door knob would do, and make design so much more compact. Many of these garage ham rigs look like they are spread over a whole table.
I post this picture of a design that is proven (WA1QIX). 400 plus watts.
I would guess most pirate broadcasters would NOT want quite so much smoke as that attracts radio calvary.
-
I've seen his stuff before. There is too much circuit strays to work on 40 meters, and one reason all his stuff is built for 80 meters. The output transformer alone is 4-6 times the size it needs to be. Also with class E, you usually need an adjustable output network to get the waveforms to where they need to be.
Again, 8 fets to get 400W? really?
+-RH
-
That design, particularly the drivers and FETs are 10 years in the past. GaN on silicon and even SiC have greatly moved things on...
11N90's need Watts to drive then, you can drive GaN with 5V and 50mW easily, the drivers are SOIC8 and could fit four on my thumbnail.
You don't need great big doorknob caps either, there's plenty of 'modern' COG devices at 1KV.
I could go on.
I have a design ready...
Look to the future, Not the past!
Str.
-
His website hasn't been updated now for nearly a decade. I get it, people want to stick with what they know works. Change and the learning curve of new technology are things we all have to deal with. That said, I'm still working on SiC stuff for the moment. I just havent had the time to move the design past the first two prototypes. Here is an early build from mid last year. I had it running on 75 meters, but the efficiency was not so good as rf drive duty cycle control had yet to be implemented. I was getting 500W carrier out of it fully modulated. The deck is currently tuned for the high end of the MW band for testing. The deck blew up when rf drive was interrupted, something that was fixed in version 2.0 with a drive loss protection circuit.
https://drive.google.com/open?id=1lbmY5WaVqBtR1BtVfAC54SDbAbE0HhyT (https://drive.google.com/open?id=1lbmY5WaVqBtR1BtVfAC54SDbAbE0HhyT)
+-RH
-
Ground the gates of your FETs with a 1K res, same as LuLu and any FET amp that will stop them blowing up with no drive!
Drivers, use NCP81074's.
Forget the rest!
Str.
-
When the input to the phase splitter loses clock, one side of the bridge will stay on (as designed) until the current through it destroys the device. I later fixed this with a retriggerable one-shot set for 1.5 times the normal rf drive period. The outputs from the drive splitter with duty cycle control are AND'ed with the drive status line. If drive is lost, the fets are tuned off. I haven't lost a fet yet since this was implemented.
Stupid stuff happens no matter how good the protection system is, but this should take care of the majority of failures.
As for the drivers you mentioned, I'll take a look. For long term reliability I prefer something I can tie straight to a heatsink, tough to do with SMD power devices without exotic board materials.
+-RH
-
I've heard the WA1QIX transmitter on the air a number of times. You can fuss about how he built it or overbuilt it, but the thing sounds great and does the job it was designed for. When he keys the mic, the band goes dead silent.
-
Just look at the pictures. No pride in the workmanship. Sloppy may work fine for 160 and 80, but it won't work on 40.
+-RH
-
Yes, this stuff does work but the lack of a PCB and all those clips wires makes me cringe!
That and everything is so MASSIVE, so not ness.
Anyways things have moved on considerably and there's much, much better devices and drivers around now.
Whilst I wont be publishing a schematic for mine just yet anyone looking around the 'other' forums can pick on what's current and copy that.
That's ALL I have done.
The NCP drivers are great and in a SOIC8 package not too small and will drive ALL the SiC and GaN you'll ever need!
Str.
-
Stupid stuff happens no matter how good the protection system is, but this should take care of the majority of failures.
I find that I spend as much or more time on protection circuitry and thinking through fault issues than I do on the intended function.
-
8) Red Hat: Man I agree. The stuff is maybe working, but its a GD junk heap. These people, one in particular K1KBW sounds very dim wit, yet he did a workman like job on his and my presumption its close to legal ham limit 375 watts of carrier.
Yet I prefer pro gear, as yourself, I like neat pro job and somebody with RF engineering smarts, not amateurs.
8)
-
Just read thru this post again...
The Drivers don't need a heatsink as the Qg of any 'Modern' FET is <10nC.
Forget those silly bolt on things as they're way too slow!
Your gate protection sounds interesting but a resistor is all that's needed!
Str.
-
;D Streachy .. I assume you are pulling ahead on a real transmitter, the 200 watt PWM. We certainly aware reports of your progress. The Wee box is working and the spectrum from same very good.
Of course free broadcasters are always at risk from Gestapo .. but the junk designs can get on some sensitive frequencies unintended and that is much more a serious issue as power levels advance beyond the flea power level. :-X
Some of the non techs are screaming drift, but heck a crystal needs to come up to constant temp to stabilize and that delta in frequency will be proportional to frequency multiplied by and doublers etc. The old 10 watt Gates FM little toy transmitters for those old enough to recall were HUGE and had xtal heaters always ON.
The frequency was pull by what was a strange motor device.. used the crystal mixer and developed a DC off set if the THING was straying. Electronics has come so far since 1967 its unrecognizable and the EE curriculum is like something from a mathematical space story.
-
I've been involved in electronic production for over 30 years (I'm in my 50's) and have work with companies small and large but always with a radio based product.
I'm self taught I guess but have been taught by the best too and really want to provide something that's really up there with the latest tech.
I'm not 'nostalgic' in any way but guess some folk are and dont really get it.
I'm going to provide a very efficient RF generator, built in DDS etc so just supply Volts and an Antenna and youve got a thumping carrier.
2 choices of mod, PWM or with half the Volts a mod tranny and amp.
Some folk can supply there own modj I guess as mod tranny and simple 100W audio amp will be fine and will make it cheaper.
I still dont expect many takers despite all the interest, mind you I was surprised at how many LuLu units I sold, around 70 in total + all the kits and PCB's I sold initially.
Absolutely No interest from Hams, so dont bother with them (am one myself!) they always think they could built it better cheap with some crap out of there junk box, another pet hate of mine as no-one else ever has the parts apart from the junk box author!
ALL the stuff I use is readily available from China via eB and the funky bits from DiGiKey.
I'm still musing over SiC or GaN and have PCB's for both, still need time to fully test them out.
The design is so simple it a joke, however making it stable is no easy matter!
I've got things running fairly well at 100W but really want to get to 200W and beyond.
Cheers fer now.
Stretchy
-
Good for you Streachy, I am an older retired from academia. EE degree only first level mate. BSEE
Now the mathematics is just fine and its a natural science without substance. People as yourself flesh out the equations and find what the hell caused the bloody thing to explode last time. LOL
In US there no issue with your generator as its a ham AM transmitter. The DDS is a superb idea. Xtals fine, but your know what happens to pirate ships that are always in the same port? I await further details as I am certain many of our comrades do.
Good Work Old Boy 8)
BTW, you do good work.. and your products can be custom also and maybe a bit of a kit. I am attempting to sell all my HR junk and get into design. At present not even a scope , those Siglents are catching my eye.
-
40% duty cycle only?
Hmm..
rgg
-
Just read thru this post again...
The Drivers don't need a heatsink as the Qg of any 'Modern' FET is <10nC.
Forget those silly bolt on things as they're way too slow!
Your gate protection sounds interesting but a resistor is all that's needed!
Str.
Apparently we're back to our breakdown in communication. A resistor is not going to keep the fets alive when the driver has latched them ON because of a drive signal loss to the input of the splitter arrangement, which is based on the circuitry from a Nautel NX series AM transmitter. The drive splitter produces push-pull drive required by the fets in the bridge.
BTW I've tried chip type drivers before and all of them burn up at 7 MHz, the power dissipation is just too high when driving 1000pF gates, and please spare me the rhetoric about the merits of GaN as they are cost prohibitive and not suitable for the power levels I'm heading towards. My only complaint about the SiC devices outside of their cost is that they want +20V/-5V drive for best switching perfomance.
Gaga, the 40% figure come from the speed limitations of the driver/fet combination. The 40% allows the fet to switch off before the other device begins to conduct, preventing a condition known as 'shoot through' in which for a short time there is a dead short across the power supply rail and the devices can be destroyed from excessive current flow.
+-RH