Update: Experimental 4 Silicon-Carbide MOSFET 1kW Transmitter on the air

[steve_qix]

Over the weekend, I constructed and got a 1kW 4 MOSFET transmitter on the air using Silicon-Carbide FETs.

The foundation heatsink is 6 x 7.5 inches.  There is a LOW speed fan on the heatsink, since the fins are pointing down.

Efficiency is amazing (do not have hard measurements yet, will do that soon).  I used my existing modulator which I use with the 24 MOSFET rig I use most of the time.   This rig runs the same power as the 24 MOSFET RF amplifier, so the same modulator can be used.

Each module consists of a single Cree FET driven by a single IXDD414 driver.  This is a single ended push pull implementation.  The shunt capacitor is 2000pF.. This may be incorrect, as the capacitors I am using are have insufficient current capabilities.  I need to get some more multilayer ceramic capacitors.

More later!

Front View:


Side View:


Closeup:

[VE3ELQ]

Awesome Steve.  One small step for Cree, one giant leap for Hamkind.
Will be on 80M tonight and say hello.
73s Nigel

[W1DAN]

hi steve

thanks for the report and photos. fun to see these new devices sing for their supper.

wonder about direct analog drive like i do as well as linear service and practical frequency limit.

73,
dan

[steve_qix]

Hi Dan, hope your arm is healing (for anyone who doesn't know, Dan fell and broke his arm quite badly - will require (I believe) titanium to correct).....

Anyway, I don't think analog drive will work with these devices.  The gate is not bidirectional with respect to the maximum gate-to-source voltage, and they can handle more voltage in the positive direction than they can take in the negative direction.  Usually analog drive is a symmetrical waveform, so the negative going voltage will be a problem.

But, so far I am reasonably impressed with the performance.  I did have to adjust the duty cycle UPWARDS (I'm at almost 50%) as opposed to using standard MOSFETs, where a smaller duty cycle actually works better.  Most likely due to the shorter rise, fall and turn off delay times of the silicon-carbide devices.

[W1DAN]

steve

thanks for the info. interesting that you had to change the duty cycle. does the output tuning feel different? they also have an SiC diode that should be looked into.

yes, i will be bionic. typing with my left hand. need to re-solder a pl259 connector to get back on their with my e-rig. that'll have to wait... ;-0

dan

[KA2DZT]

Steve, only you can make something that looks that bad actually work.  Hopefully this is not the final assembly model.

Fred

[w1vtp]

steve

thanks for the info. interesting that you had to change the duty cycle. does the output tuning feel different? they also have an SiC diode that should be looked into.

yes, i will be bionic. typing with my left hand. need to re-solder a pl259 connector to get back on their with my e-rig. that'll have to wait... ;-0

dan

Dan what about Dragon Speak?  Would that help?  http://www.nuance.com/dragon/index.htm

Al

[n1ps]

Steve, only you can make something that looks that bad actually work.  Hopefully this is not the final assembly model.

Fred

Ugly but technically beautiful    

Peter

[steve_qix]

Steve, only you can make something that looks that bad actually work.  Hopefully this is not the final assembly model.

Fred

Guilty as charged     I think this is one of the neatest prototypes I think I've ever done!!!  There are almost no clip leads at all 

[N2DTS]

4 KW pep and nothing flashes over?
Amazing, and small.

[steve_qix]

So far, so good.  I used the rig all evening and had numerous QSOs.  The modulation is different on this RF amplifier as opposed to the normal RF amplifier that I use with this modulator, and I need to investigate why.  Both positive and negative peaks are affected, so I think it's a capacitance (load on PWM filter) difference.  So that's the next series of measurements.  But, overall things are working very nicely.

[N1BCG]

I did have to adjust the duty cycle UPWARDS (I'm at almost 50%) as opposed to using standard MOSFETs, where a smaller duty cycle actually works better.  Most likely due to the shorter rise, fall and turn off delay times of the silicon-carbide devices.

Will the QIX digital driver need to be modified to accommodate these FETs?

[steve_qix]

Yes!  I am using that exact board in all of my setups.  I just adjusted the duty cycle to be closer to, but not at 50%.  Looked at gate waveforms of the two phases on a dual trace scope, and adjusted the duty cycle so there was no overlap and very little gap between the two streams.

[KF1Z]

Neato-keen....

Can I get the full part number?
Can't quite see it in the pix.

[steve_qix]

Neato-keen....

Can I get the full part number?
Can't quite see it in the pix.

Sure - I forgot to include it.  I'm using the C2M0040120D device.  There are other devices which may be more appropriate for different power levels.  Check the Cree web site.

[KA0HCP]

Steve, rig sounded great in Maryland last night.  Just clear as a bell.  Big signal.  bill

[Mike/W8BAC]

Nice transmitter Steve. It is amazing you can get efficiency that high at power outputs like that. I suppose you will have a portable rig to take to the island this summer?

As far as the clip leads go, it looks like you substituted many for Kurney Split Nuts. Those are great. I keep a verity of sizes here as well. Have fun Steve! If you can get a picture of the transmitter and modulator lashed together, Please click away.

Mike

[steve_qix]

Hi Mike,

Thanks!   It was an interesting thing to put together.

At the Island I already have a transmitter of similar power, but I think it might be neat to use these MOSFETs up there, although if it ain't broke 

The advantage would be higher efficiency for sure, so I might do it.

The picture does show the connection to the modulator, but it's not obvious.  The modulator connection is that gray 4 conductor cable, and the heavy green wire (that's the ground).  Each module connects to one of the 4 wires in the interconnecting cable, and there is a current meter on each of those lines.  I use the same modulator for other similar RF decks like the 80/160 meter deck and the 40 meter deck.

Might be on with it this evening... depends on what else is going on!

Regards, Steve

[Detroit47]

Steve

You sounded good in Michigan this morning. A great big signal and nice audio. The new TX is a winner.

John N8QPC

[steve_qix]

Steve

You sounded good in Michigan this morning. A great big signal and nice audio. The new TX is a winner.

John N8QPC

Thanks John!  You were doing quite well yourself.  Always nice to meet some new folks on 75 meter A.M.

Regards,  Steve

[Mike/W8BAC]

Hi Steve,

What I am hoping for is a picture or two of the entire modulator your using so I can get an idea of the component size and weight. I would like to get an idea of the size rack cabinet needed to house the transmitter and modulator.

Mike

[steve_qix]

This is the modulator I use at Rattlesnake Island.  It puts out the same amount of voltage and current as the modulator in Townsend, however this one uses cores for 2 out of 3 of the inductors.  This is in a standard 19 inch x 7.5 inch high x 16 inch deep cabinet.  The power transformer is external.

The Townsend modulator is much larger due to air core coils.  I have a picture of it somewhere.  It is not mounted in a regular cabinet.  It is simply a collection of parts, many of which are just sitting there, unattached.

This one is compact and light weight:

Top View:



Front View:

[Opcom]

This is all very impressive. Nice looking modulator there too.
The annoyed words of the semiconductor engineer at a previous employer still echo in the distance "You can't use switching MOSFETs for RF!"

[Rob K2CU]

Quite a rig Steve! Kudo's

For the caps, have you looked into silvered mica designed for high current IGBT snubber applications?

[w1vtp]

This is my PWM (PDM).  It's for a 400 watt carrier but is very similar to a 1000 watt carrier unit except the power transformer and filtering is built to handle the increase in power.  I believe the same control boards are the same and the heat sink on the back would be bigger

First pic is the whole transmitter - top to bottom = VFO / digital drive, Modulator / PS and RF deck

Second pic is the modular without the shelving section showing the power supply comprising two power transformer which secondaries are wired in series to get the proper PS voltage

Third pic is the shelving section comprising the pulse width modulation circuitry and protection / efficiency circuitry

gl, Al