KF1Z 's new 75 meter class-E

Post your photos here and write-ups you may have

Postby frank carcia » Wed Jun 13, 2007 10:09 am

1:2 will raise your output Z making the LC ratio change to lower C and higher L. You should be able to do 500 watts carrier easy with 12 FETs and around 90% efficiency. You have good drive to the gates.
You should not need to use more than 500 pf series tuning. Be careful with broadcast variable caps they are not designed to operate at high RF currents. Check the ground connection to the rotor for heat damage.
You might consider using the 640 pf with a padder as the loading cap later. I use a 1000 pf Vacuum loading cap with 1000 pf padder.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Wed Jun 13, 2007 11:01 am

Well, I might be able to get 500watts carrier....
But my modulator would be screaming for mercy (as is with the power transformemr I have int it...)

I haven't been brave enough to run it up to 500 watts INPUT yet.....
Maybe I will in a bit.....


As for the loading cap.... very interesting.... I thought most everyone was using the broadcast variables for that purpose....


I checked, and what I have right now is 5.0uH L.... 485pf C, and 2800pf loading C...


I wasn't aware that the loading cap would have to deal with that much current......
Hmmm I have a 1200pf vac variable on the way along with 3 other 650pf....

maybe I will change the transformer to 1:2, and see what happens....
I just got a pile of the Fairrite cores in the mail last week, and some new wire at the home-depot as well....
Wouldn't have to ruin anything in process.....
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Wed Jun 13, 2007 11:43 am

Look at the ground connection between the case and rotor shaft. Brass on spring metal. It was never designed for high RF currents. Yup lots of guys using them but not me. I learned the hard way after burning a few out back in the 90s. I would try the tank as it is now with a 1:2 transformer.
12 FETs should belt out 500 watts easily at 48 volts. I run out of steam at 1200 watts with 22 FETs at 47 volts. The SAT voltage starts getting a bit high.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Wed Jun 13, 2007 12:32 pm

Well now.......

Don't know why I had thought of this before.....

Just for a test, I slid another sheet of Kapton under each heatspreader to lower the Shunt C a bit....

Now my peak voltage across the drains is 140 volts with 40 volts dc in...
Wish I had noted what it was before....


AND, my RF power out went from 250w to 273w ...

Gained approx 6% efficiency...... from 79% to 85%...
(again...crappy measuring device disclaimer..!! but for comparison only!!)



2 questions come to mind...

First... should I try lowering the Shunt C more?
(I'd have to undo a bunch of connections to measure what it actually is....)

Second, how thick can I make this Kapton layer be before the heat-transfer is not good enough?

I can, of course, take the heat-spreaders off, and trimm them, or bevel the edges by a bit instead..... but that's a lot of work!


So, I've got about 3.5 X DC input now... and I'd guess I don't want that higher?


It appears that I can get slightly greater efficiency by raising my current....
I tried going from 40v at 8 amps total to 40v at 9 amps total, and it looked like another 2-3% increase in efficiency....

Not too big a jump... not sure it's worth changing my PWM output filter to 4.45 ohms from 5 ohms imp.
But, something to note.........


Gee, I love it when I make progress!

As opposed to regress.....
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby kf1z » Wed Jun 13, 2007 12:39 pm

frank carcia wrote:Look at the ground connection between the case and rotor shaft. Brass on spring metal. It was never designed for high RF currents. Yup lots of guys using them but not me. I learned the hard way after burning a few out back in the 90s. I would try the tank as it is now with a 1:2 transformer.
12 FETs should belt out 500 watts easily at 48 volts. I run out of steam at 1200 watts with 22 FETs at 47 volts. The SAT voltage starts getting a bit high.



Yep, I can still try that as well.....

I can run mine at 500 watts INPUT, as is....

50 volts at 10 amps.....

after that, I'd probably have to add a third mosfet to the modulator, and get a bigger power toroid......

Of course, I'd rather just run on the conservative side anyway....
Normal operating I don't see myself running more than say 350- 400 ish watts out...
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Wed Jun 13, 2007 1:20 pm

Well if you want to run low power then you could decrease the shunt but a bit extra shunt is good insurance. Sounds like you might try making some more output power to see if you get 3.5 to 1 at higher current with the shunt you have.
I have never used anything thicker than 3 mil so don't know the thermal performance of thicker material.
What was your peak ratio before you decreased the shunt?

I'm talking 500 watts carrier output you should be able to get 90% efficiency with 12 FETs but you will need a bit more voltage say 46 to 48.
I found an amp DC per FET at carrier is pretty conservative. So you could run 12 amps easily. So 500 watts input at 40 volts should be easy.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Wed Jun 13, 2007 1:50 pm

I believe the ratio before was 3.25 to 1....

Unfortunately, I didn't mark it down.......


OK, I just tried going up to 500 watts dc input....
50volts at 10 amps....

The drain voltage ratio didn't change much.... went to 3.6 to 1
(180 volts for 50 v in)

And I was seeing 445 watts RF out...




And yes, could run the higher power with power supply mods... only have about 128vdc available with this transformer.. and would have to add to the PWM modulator.....

Kinda wish I had built a Class-H modulator.... just so output filter impedance wasn't an issue.....

But as I say, I'm happy with 300-500 watts input for now.....

As long as it's playing nice!!

Oh, by the way...
THANKS FRANK for your input, on my output!

still haven't built that window comparator circuit....
Soon though.... I want to see what difference cutting the drive down to under 50% on time does....
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Wed Jun 13, 2007 3:18 pm

PDM filter output Z is not real critical. You can move around a bit. I have simulated my filter set for 2 ohms and didn't see anything until I got to 3 ohms where I got peaking in the response. Don't worry about that now a couple DB in response change can be fixed later when things are functional.
I'm at about 175 volts peak at 47 volts power. with about 88% efficiency so sounds like you are close.
If it looks stable then move on and try modulating it at low power and work your way up.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Wed Jun 13, 2007 3:27 pm

Ok Frank thanks......

Oh, I've been on the air with it....



I'll wind up a new xformer tomorrow, probably just use #10 wire 'cause it's quick and easy....
Besides, I don't think I can fit 3 pcs of this #8 through those cores!


I'll report my findings......
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Wed Jun 13, 2007 3:40 pm

#10 will be fine I use #10 teflon in my rig.
Start at a low voltage. I will be interested in what you get. Take some good notes of where you are right now at a couple different power levels so you have good data to compare performance. I found 1:2 worked better than 1:3 and 1:4 but never tried 1:1 because of my low output Z.

I never tried the comparator thing but got some samples. I'm in the middle of building up my HPSDR stuff.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

What a difference a day (and a 1:2 transformer) makes!

Postby kf1z » Thu Jun 14, 2007 8:10 am

Ok,
As suggested, I wound a 1:2 outout transformer, and did some prelim testing....


at 3.885mhz, with 40 volts dc, and 8 amps....

I now have about 138v peak on the drains...

Wonderfull looking waveform....(was pretty before too)

And, about 294 watts RF OUT.....for the 320 watts dc in.

That's about 91%.....

A 7%+ gain from what I had yesterday.



Sorry, I still haven't taken things apart enough to tell what my Shunt C is...

This is with :
5.4uH inductor
484pf series C
1160pf loading


I'll do some more tests, at different frequencies later on .....
But, sure looks like the 1:2 ratio output transformer really made a significant improvement!
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Thu Jun 14, 2007 8:30 am

Very Good!
Now your output circuit more matches the output Z of the final.
Your Q was too high for the 1:1 transformer.
Now run it for a while and see if the inductor is heating up.
That is my present problem because I only have silver plated 1/8 copper tubing. It runs hot at 1 KW. I need to got to 3/8 to cool it down.
My inductor value is still a bit high. My final is 1/2 the output Z of yours so this confirms why I am only at 87 to 88%. My efficiency goes into the 90s when I reduce power and run at a higher output Z.
Now increase your power and see what happens. Sounds like the shunt value is good at this power level. When you increase the power at the same voltage you will see a point when waveform starts looking funky meaning you have reduced the output Z to a point where you need more shunt C. You want to stop there and decide if you want to run more power or back off to a nice waveform. The nice thing about the spreader shunt C is you don't have all that RF current in a dork knob heating it up and making the value drift down.
Would you send me the size of your heat spreaders and thickness of the kapton? I should have used bigger spreaders so had to add dork knobs in parallel. fc
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Thu Jun 14, 2007 8:54 am

Frank,
The heatspreaders are 2"x6" ea.

I think there's 6 mil of Kapton....
BUT, I don't have my micrometer any more to confirm....
AND, the spreaders are torqued right down pretty tight....

I bought it as 2 mil Kapton.... (ebay)
I honestly don't know if it even IS Kapton.... never worked with it before.



After about 15 minutes , the inductor is very cold...

I run out of series C at 3.7mhz.... still good eff. at 4. mhz

So I guess now I need just a touch more L (so I can work down to 3.6mhz)

I'v got a 1200pf vac variable on the way....
Looks like that can find a home as my loading cap....
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Thu Jun 14, 2007 9:10 am

See if you can find a 100 or 200 pf NPO dork knob and throw it in parallel with the 500 pf series or go to the 640 pf unit. More L may degrade efficiency. I'm running 4 by 5 inches with 3 mil kapton but there is a 1/4 hole under each FET with a void where the 1/2 inch copper tubing primary is silver soldered to the plate. Sounds like I have slightly more surface area but I had to add 1000 pf in parallel to hit 1 KW out and be clean.
kapton looks like butter scotch color. Without the additional 1000 pf 1 kw was really pushing it. When we tested the rig at Steve's I didn't have the additional shunt C and it was barely stable at 1 KW.
Sounds like your shunt value is good for the power you plan to run.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Thu Jun 14, 2007 9:18 am

Ok then....

I guess there's a good chance that this IS Kapton....
It is butterscotch in color anyway....

The big question is the actual thicknes.....


Right now, I've got to figure out where the RF is getting to the rig...??

When I'm on the dummy load.... the two modules are balanced in current....

When I go to the antenna.... one module drops in current by 1/2 amp (out of 4)

My guess is that RF from the ANT tuner is getting to the gate drive line of that module......


Always SOME little thing!!!
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

PreviousNext

Return to Your Photos and Articles

Who is online

Users browsing this forum: No registered users and 1 guest

cron