Seeking advice on rebuild of a Class-E rig

[W4EWH]

Thanks for reading this. I'm rebuilding a Class-E rig that I just acquired, and I'm looking for advice about the best way to go.

I don't have a lot of information on the unit, other than that it has a Class-H modulator and doesn't have a sequencer. I've attached pictures of the mod and RF decks.

All suggestions welcome, on or off-list, and thanks in advance.

Bill, W1AC

[n1ps]

Looks like one of Steve's designs.  12 pill RF deck.  Who built it?

Why does it need to be rebuilt?

How is it driven?  Sinewave?

p

[KF1Z]

If the Class-H modulator is working ok, then I would leave it alone.

But, if I were to be rebuilding the RF deck,
I would strip it bare. Leaving the heatsink, 2 var caps, and coil in place.
( probably I would leave the TR relay as well

I would convert it to "digital" drive, and only leave 8 of the FETs in place.
( 2 banks of 4 )

Using the schematics:

http://classeradio.com/8_fet.htm

and

http://classeradio.com/vfo_2_band.pdf



Now, the RF deck may indeed WORK the way it is, and if so, you could decide to leave it alone as well.

[steve_qix]

It'll be an interesting project.

You might be able to use the modulator.  That looks like the KD1J board in there.  It needs a modification to include an overload shutdown.  This can be retro-fitted into the design, assuming everything else works that is.

You need a power supply (obviously   )

On the sequencer, a sequencer is an easy project.  Really, every station should be sequenced.  I've been using sequenced transmit/receive for 40+ years and strongly recommend it.  I have a schematic I came up with a few years ago.  If I can find it, I'll share here it on the forum.

The RF amplifier is pretty much a complete rebuild.  You could probably use the same heat sink and drain bus.  The number of MOSFETs in EACH module must be evenly divisible by 2 (4 or 6 per module, but not 5 or 3.  I suggest 4 per module with that modulator you have.  The shunt capacitors need to be replaced with multilayer ceramic capacitors.  The drive needs to be digital drive, which will involve mounting 1 driver IC (an IXDD614 TO-220 package) for every 2 MOSFETs, along with DC bypass capacitors for the drivers, etc.  Not a big deal, but you do need to have sufficient room on the heat sink.  Each drain bus should have 2 540V (or 550V) transzorbs across it.

It's probably a couple of weeks of work, assuming the modulator works.....

The RF amplifier, as is, is a poor implementation and should not be used at all.  It will be unreliable and unstable in its current form.  There are multiple problems from the design to the layout.

Use a modern digital drive implementation, and all will be well !!

[W4EWH]

I've been going through the RF deck that I have, trying to figure out how I'll set up the new layout. I'm trying to understand the existing design so that I don't waste effort duplicating things I'll want to keep from the old design to the new.

AFAICT, the existing deck is set up for direct RF (sine wave) input driving the 11N90 FET's, with two banks of FET's in a Push-Pull configuration. I need some more eyes on this, because I just can't get the "phasing" of the output transformers clear in my head.

The first photo I've attached shows a schematic from classeradio.com (Thanks, Steve!), of a "push pull" Class E amplifier being driven with a sine wave. The second shows the existing output transformers on the RF deck I have.

While the input transformers make sense (the input of one is reversed compared to the other), I have a blind spot when looking at the output transformers: the second picture seems to show that the two outputs aren't wired the way the schematic calls for.

Please take a look at it and tell me if I'm wrong. Thanks.

73,

Bill W1AC

[KF1Z]

The transformers are ok.

What you have to picture, is that you have 2 transformers, stacked on top of each other.

For ease of wiring, one of the transformers is flipped upside down.



Note the Phasing dots on the schematic are flipped as well, which would make it confusing when trying to compare
the physical layout with the schematic!

[n1ps]

Good advise from the pros.  Bruce is this supposed to be a 2 band TX?  Looks like the relay switches in or out the coil in the rear.  Curious about the trimmer in back of the relay.  Be careful tuning it!  Are those pennies? The 'ol 1 cent hole plug maybe

p

[KF1Z]

Peter,
I think the relay is a TX/RX switch
The coil in the back is part of the tuned input circuit.

Looks like a single band

[K1JJ]

Bill,

For pride's sake, when you rebuild it be sure to re-mount the crooked parts at nice right angles....

Have fun with your project!

T

[KF1Z]

I have a zip file with the schematics and board layout for that modulator if you need them

[W4EWH]

Bill,

For pride's sake, when you rebuild it be sure to re-mount the crooked parts at nice right angles....

Good idea: I'll attend to that right after I give the Drain bypass caps a return path that isn't going through the heat sink, two or three screw-heads, and ghod knows what else, when I have a valid electrical bond between the heat sink and the backplane, when there's more than .0001 inch of clearance between the FET Drain leads and the backplane, and after I divine the number of ground loops between the modulator deck and the RF deck.

It's right up there on my priority list!

73,

Bill, W1AC

[W4EWH]

I have a zip file with the schematics and board layout for that modulator if you need them

My hero! Thank you!!

I'll PM you my address.

There's a special place in heaven for you.

Bill, W1AC

[steve_qix]

There are a number of model class E rigs that should be used as guides.  The greatest number of digital drive model rigs were probably built by Bruce KF1Z and Wayne WA1SSJ.  There are other model rigs out there as well.  I mention Bruce and Wayne because they have built a good number of these model-type rigs for others as well as for themselves, and pretty much have it down to a science.  The rigs built by me work well, but are not particularly neat and therefore aren't necessarily good models.

We need to get some good, close up pictures of one of these model rigs, and I will then put them on the class E web site.

Following the schematics on the class E site, and using a layout like the model rigs use will pretty much guarantee success!

[KF1Z]

I started to with "George's" rig, got a lot of the modulator, but only a couple of the RF deck.
Was looking last night on disks to see if I could find more....

http://greenmountainradio.com/k8rhh.html

I'll keep looking....

And, since my original rig, with the copper heat-spreader shunt caps is in need of replacement/rebuilding
maybe soon I can build another and take lots of pics on the way..  :-)

[n1ps]

Nice pics!  3 modulator fets....looks like lots of pills coming for the RF deck. I like the capacitor bank PC board.  What is that case?  Also where do you get meters now?  One of the guys around here is about to start an E rig.
p

[KF1Z]

The case was from Par-metal.com  ( It is owned by the same people that run Antek.... that do toroid transformers)

The meters I have a bunch of that I bought from overseas.
I bought a bunch, thinking I'd be building a bunch of rigs, then it never happened.

[W4EWH]

I've been going cross-eyed selecting components on the Digi-Key site. The last item on my shopping list is Sil-pad: it's the insulation material used to mount the 11N90's and 614 chips on the heat sink.

If you've used Sil-pad, please give me a clue as to what I'll need.

I figure that -

• A 450 watt Class E RF deck that puts out 385 watts has about 70 watts of heat to dissipate, counting the drivers and incidentals.
• In a push-pull configuration, that's ~33 watts per bank of 11N90's.
• The Sil-pad specs list thermal performances of several degrees centigrade per watt, which scare me.
• I need to know which Sil-pad material is 'best' for this application.

Thanks in advance.

[KF1Z]

BER178-ND  digikey sil-pad

[steve_qix]

I've been going cross-eyed selecting components on the Digi-Key site. The last item on my shopping list is Sil-pad: it's the insulation material used to mount the 11N90's and 614 chips on the heat sink.

If you've used Sil-pad, please give me a clue as to what I'll need.

I figure that -

• A 450 watt Class E RF deck that puts out 385 watts has about 70 watts of heat to dissipate, counting the drivers and incidentals.
• In a push-pull configuration, that's ~33 watts per bank of 11N90's.
• The Sil-pad specs list thermal performances of several degrees centigrade per watt, which scare me.
• I need to know which Sil-pad material is 'best' for this application.

Thanks in advance.

If all is working properly, a 450 watt power input class E RF amplifier on 75 meters will put out around 405 watts of RF.  On 160, it will be even better.

And of course, all of the power used by the drivers is converted to heat.  For 8 MOSFETs on 75 meters it's probably around 40 watts.

Each individual MOSFET isn't dissipating a lot of heat - less than 6 watts in the drain and probably 2.5 to 3 watts in the gate (all drive power is ultimately converted to heat).  But, you do need to have a reasonable thermal connection to the heat sink regardless.  The SIL pads Bruce recommended work very well.  The drivers should be screwed directly to the heat sink without any sil pads or other insulating devices.