The Inexpensive, simple, MOSFET A.M. linear project

[W1RKW]

Steve, what is the tune up procedure?  Tune for max outpoot?

[KA8WTK]

I have a couple of heat sinks that might be good for this project. They are 10.250" long, 3.125" tall and the long fins are 2" long. I do not know what they were from, but they each had 2 pair of 2SA1094/2SC2564 (all shorted out) on them. In the photo I have 6 W45NM50FDs sitting there to get an idea of spacing. I will need to drill holes for the new MOSFETs.

Bill KA8WTK
 

[W1RKW]

Anyone building this amplifier, if you need some silicone transistor insulators, PM me.  I have a boat load and will only use 10 for my project. Left overs are available at $0.25 each.  Send an SASE with QTY and funds and I'll promptly return.

[WBear2GCR]

While it does not appear to be an issue at present, any amplifier that runs at ~50% efficiency and
does 200 watts will produce a LOT of heat.

The heat must be removed efficiently or one can run into thermal failures.

Definitely the advantage of multiple devices is to reduce the thermal load per device.
Also, afaik generally speaking multiple devices in parallel exhibit lower distortion than a
single device when used in an output stage... So that's good!

One issue is that the heat inside the device and the heat on the heatsink or even on
the surface of the package are not the same. The heatsink's temperature rise lags
very far behind the device's.

In addition, the lower the operating temperature the greater the effective SOA.

One issue is the thermal mass of the heatsink vs. the radiation possible & transfer to
air possible from the surface area, including the fins. The thermal mass will sink heat
until it becomes saturated, then there is no place for the heat to go, so the device
temperature suddenly goes up quickly - unless the surface can shed the heat at
a rate equal or greater than the BTUs being generated by the devices.

If one has sufficient thermal mass, and keydown is short, probably the thermal load
cycling will average the temp rise within a reasonable range. But we're talking 200-400watts
of RF and an equivalent amount of just heat?

The heatsinks shown in this thread will not likely be suitable for convection cooling.
They may or may not be sufficient for "blown" operation.

I think that maybe if I was prototyping an amp like this, I would run it keydown for
more than 30 minutes and note the temperature rise of the heatsink and the devices.

I'm sure that many have seen how in many RF amp projects using one or two high power
devices they use a copper heat spreader? It may not be necessary here - but the key
to knowing if it is will be the rate of temperature rise in the junction (actually the bonding
wires tend to blow first) of the device vs. the rate of rise through the transfer to the
heatsink.

IF all the devices' mounting surfaces are electrically connected there would be a big thermal
benefit to mounting them all to one copper plate, without insulation, and insulate the plate
to the heatsink. This permits the rapid spread of the device's heat to the copper plate which
has both a high thermal mass and a superior thermal conductivity, permitting a large surface
area to be available to transfer through the lossy insulator layer to the heatsink. Otherwise
the first thing each device hits is the silicone insulator or mica...

FYI, companies like Wakefield (they make heatsinks) have calculators and pdfs on heatsink
considerations and design, convective and blown.

Just thinking out loud...

                           _-_-

PS. would really like to do that IMD test...
PPS. Steve, did your FFT show 2x the carrier freq and 3x the carrier freq?

[steve_qix]

Steve, what is the tune up procedure?  Tune for max outpoot?

I don't have a great tune-up procedure yet.  I tune for maximum positive peaks, given a certain power input.  However, I think there may be a better output transformer ratio (maybe 1:2), that may make tuning easier.  Will be experimenting !!

[ka1tdq]

I would just like to add that I didn’t learn how to tune class E stuff until I built one. Reading and theory is good, but until you turn some knobs, you’ll never know. Now it’s a piece of cake for me.

Jon

[KA8WTK]

I was talking to a friend about needing a heatsink larger than what I had around. (shown in previous post.) He offered this one. I am going to take him up om it.

[Opcom]

Someone murdered a power supply!

[steve_qix]

Heat:  There's been a lot of discussion about heat.  All linear amplifiers, particularly when used for A.M. generate a *LOT* of heat !!!  You must use a SERIOUS heat sink.  The heat sink I am using is big, and it has a good, fast fan blowing on it, and it gets HOT.

I have a temperature measuring device attached to the heat sink.   I have it set such that when the heat sink reaches 109 degrees F, the relay in the temperature sensor will close, and this will activate the fan.  The sensor costs about $7, and a link can be found here:  https://www.mpja.com/Digital-Temperature-Controller-FAHRENHEIT/productinfo/32764+MP/

So, you really do need a big heat sink that will actually get rid of hundreds of watts of heat.  If you're running the amp at 200 watts carrier out (600 watts input), 400 watts of heat will be generated.  That is actually quite a lot of heat.  Even with a good fan, the big heat sink I'm using will reach 115 degrees.  It stabilizes at this level when the fan comes on.

[WBear2GCR]

Needing more information - they've changed this, and sold the company - but this design guide will
help with the magnitude and nature of the issues. You can skip most of the math, look at the pics
and the graphs - note that they show the losses as resistances.

https://www.digikey.com/en/pdf/w/wakefield-thermal-solutions/introduction-thermal-management

https://www.digikey.com/en/pdf/w/wakefield-thermal-solutions/heat-sink-thermal-resistance

                   _-_-bear

[WU2D]

Dip the plate! In this case the plate is 1 inch aluminum, and the dip is a liquid nitrogen cryo loop.

[w4bfs]

check out May 2021 Q street magazine ... liquid heat exchange .....  its making me moist 

[KB2WIG]

https://www.google.com/search?q=wlw+transmitter+cooling+pond&client=ms-android-samsung-rev2&prmd=sivn&source=lnms&tbm=isch&sa=X&ved=2ahUKEwjSjeDQlvnvAhXkQ0EAHay0AYUQ_AUoAnoECAIQAg&biw=798&bih=340&dpr=2.63#imgrc=NTnN88eal6Cp8M


The old ways are the best ways.

KLC

[Opcom]

These are pretty nice. No relay output but very inexpensive, 3 1/2 digit, and reads 0-1999 deg F (1200 C). Uses a 'type K' TC which the user supplies.

https://www.bgmicro.com/TST1201.aspx