high power solid state amplifier schematics

[kb3ouk]

does anyone have any schematics for a 100 to 300 watt solid state rf amplifier? in particular, class c solid state amps that could be modulated by something like modified heising  with a hi-fi audio amp. i've found plenty of class e amps but nothing else. the reason i'm looking at class c is because i could drive the amp with my icom if i were to build some kind of transmitter. class e needs to be driven with a digital input so couldn't be driven directly by a transmitter, right?

[Pete, WA2CWA]

Try here:
http://www.communication-concepts.com/
Also has Application Notes: http://www.communication-concepts.com/application_notes.htm

[w3jn]

No reason at all you couldn't drive Class E with your Icom.  Not sure why you'd want to though...

The much discussed "digital gate drive" is somewhat new to the Class E arena.  Several years ago most guys were driving them with riceboxes.  Switching transients, accidental overpowering, etc., fried many a MOSFET back in the day, which was the push to go a better route.

It's easy to build a VFO and the matching gate drive.  If you don't wanna build a VFO then try the N3ZI DDS VFO.  Easy and inexpensive.  http://www.pongrance.com/ddfc.html

[kb3ouk]

the reason i would want to is to use it as an amp if i want to run CW, then be able to drive it with a carrier and modulate the amp  for AM.

[Opcom]

Some bipolars are designed for class C, some are not and can be damaged over time. How MOSFETS will do I don't know. But go for it! 20 million old CB AM transmitters can't be wrong, so scale up!

[w3jn]

the reason i would want to is to use it as an amp if i want to run CW, then be able to drive it with a carrier and modulate the amp  for AM.

Like I said, not sure why you'd want to.... use a 100-300W amp on a radio that's already 100W for CW.  The few dB you gain going from 100W to 300W is negligible.

No reason you couldn't "plate modulate" a class E xmitter using a transformer, given enough current carrying capability on its secondary.

Why not build a Class C rig from the VFO up, if you wanna do Class C?  You only need a couple extra stages - a VFO and a driver.  Then, you don't have to worry too much about the input VSWR of the final, you can design the network between the driver and final to match each other without having to worry about overdriving the final input, switching transients, etc.  You'll learn a lot about impedance matching, VFO stability, etc with a project like that.

Many riceboxes have nasty spikes when first keyed, due to a number of factors.  The ALC circuit sometimes has too slow of a time constant and you get overshoot (ie full power) on keyup.  I had a particularly bad one on my FT-1000MP when using a hand mike.  The "click" from the PTT switch caused a big audio spike which in turn wasn't suppressed by the ALC due to a long time constant.  Using a desk mike cured that one.

[steve_qix]

Drive wise - at least with MOSFETs, there's really no significant difference between class C and Class E.  The difference will be the output network and of course what the waveform looks like at the output.

You can absolutely drive class E MOSFETs with your rice box directly - you just have to make sure you have the rice box output set correctly.  You a TVS (transzorb) across the gates of any MOSFETs you use to protect them from accidental overload or transient spikes.  You will need some type of matching network.

But, why bother with all this complexity in the driver when an IXDD614 driver will do it all, at least up to 40 meters.  Above 40 meters, I still use analog drive.

Using a class C, D, E or F  RF amplifier for CW will generate key clicks unless you key the power supply also, and give it a reasonable rise time.  This is very easily accomplished using a series MOSFET and some type of R-C or other timing circuit in the gate to control the turn-on time.  Of course the same MOSFET can be used as a series modulator.

[w3jn]

Is there ANYTHING a MOSFET can't do? 

Steve - not sure why you'd necessarily get key clicks with Class C, D, and E... is that something peculiar to solid state amps?  I confess to never having used a class C or D solid state amp.

[K1JJ]

Shelby,

There are many variations of class E rig design complexity. Maybe the basic six fet 375 watt analog modulated rig would suit you.  You are guaranteed success without re-inventing the wheel.  The class E guys here and on the air won't let you fail...

Here's a thread that describes building a 24 MOSFET PDM class E rig from scratch. It's buried in the archives. This rig could have been driven by a ricebox, but instead uses the newer VFO/digital drive. It is very stable and I've never blown up the 24 pil RF deck to date. I use it as my main 75M AM rig these days:

http://amfone.net/Amforum/index.php?topic=23354.0

I just realized the rig was built in April, 2010 -  a little over a year ago. It seems so much longer back that Rico Suave started dancing in his crib.

Yes, CW operation should be no problem once the time constants are smoothed out. Just like any rig, the CW note chracteristics must be tailored. Even some new riceboxes had terrible clicks until modified.

Good luck.

T

[WA1GFZ]

You would be better off with FETs. Bipolars present a weird input impedance to a rice box final when operated in calss C. It would be pretty hard to broadband.
FETs a BB transformer and swamping resistor and you are done. Add transorbe for safety.

[W1RKW]

You would be better off with FETs. Bipolars present a weird input impedance to a rice box final when operated in calss C. It would be pretty hard to broadband.
FETs a BB transformer and swamping resistor and you are done. Add transorbe for safety.

Agree with the bipolar statement having gone down that road.

[steve_qix]

Is there ANYTHING a MOSFET can't do? 

Steve - not sure why you'd necessarily get key clicks with Class C, D, and E... is that something peculiar to solid state amps?  I confess to never having used a class C or D solid state amp.

Well, I know from the tube days that you couldn't just key the drive to a tube class C final - major key clicks.  The same goes for class C, D or E solid state amps.  You can't just key the drive - same problem - key clicks.  This is because the shape of the keying waveform is not preserved in class C, D, E or F due to the fact that the amplifiers are not linear (nor should they be!!).

"Overshaping" of the input waveform is NOT recommended with MOSFETs because they have so much gain.  A parasitic can result when the MOSFETs are operating in the linear region.  And if you are using digital drive (which is a better way of driving MOSFETs) you can't even overshape the input.

So, what is generally done is to key the drive and key the high voltage using a series element (a MOSFET or MOSFETs in parallel) - with the input to the series element being a good CW rise and fall time.  The series element is most easily created using a source follower and some type of pull up with controlled rise and fall times.  Easy to do with simple R/C circuits.

[Opcom]

Is there ANYTHING a MOSFET can't do? 

None have lived long plugged into a 6L6 socket.. but maybe I was doing it wrong.

[kb3ouk]

well, i believe i found a simple schematic using an IRF840, its designed for 40 meters but could be modified for 75 meters. now, the IRF840 is rated for 125 watts of dissipation and max 500 volts and 8 amps. now for AM service, the safe max voltages would be at the most 250 volts and 4 amps, which allows for double voltage and double current, or should it be more like 125 volts and 2 amps, allows for 250 watts input, should put out abut 150 watts.

[WA1GFZ]

I run 14 IRF840s in push pull parallel on 160 since 1996. It is PDM modulated with an 80 to 90 volt power supply. It will do almost 2KW PEP on 160. It started out class D until I discovered class e by mistake.

[kb3ouk]

they seem like they are very durable. what's the maximun power you could get out of a single one on AM?

[steve_qix]

well, i believe i found a simple schematic using an IRF840, its designed for 40 meters but could be modified for 75 meters. now, the IRF840 is rated for 125 watts of dissipation and max 500 volts and 8 amps. now for AM service, the safe max voltages would be at the most 250 volts and 4 amps, which allows for double voltage and double current, or should it be more like 125 volts and 2 amps, allows for 250 watts input, should put out abut 150 watts.

Do you want a safety factor?  As an example, all of the rigs shown on the class E website have a 100% safety factor built into their design.  So, all devices are run at 50% of their breakdown voltage and current at the highest peak, maximum modulation operation of the RF amplifier.

If you want to have the same safety factor, at 200% positive modulation, the IRF840 can be run at 25VDC at about 1.2 amperes DC power input, carrier.  A single FQA11N90 can be run at 45 VDC at about 1.3 amperes.  These are conservative class E numbers using my design criteria (100% safety factor), at 200% positive modulation.  If you run it this way, and put a 300 volt TransZorb across the drain to ground, the device should last forever.

If you limit the modulation to, say, 150% (in other words, you build a modulator that is absolutely incapable of more than 150% positive modulation), the voltage can be increased to 30 volts and the current to about 1.4 amperes and still maintain the same safety factor.

Just as an academic example, if your modulator were capable of putting out no more than 100% positive (don't do it, but it's an example),
you could increase the voltage to 37 volts and the current to 1.75 amperes.

The reason for the seemingly low voltages is due to the fact that the RF peaks will far exceed the DC under many conditions in all classes of operation.  In class E, the normal RF peaks are approximately 3.6 X the DC under normal conditions.  Class D is a bit less - again, under ideal conditions.  But, bad loads, mis-tuning, etc. can cause even larger spikes, depending on the circuit.

If you use the FQA11N90, the voltages shown above can be increased by approximately 80%, and the current increased by approximately
30%.

Either way, MOSFETs are very inexpensive and small, so simply use more of them to get higher power.

Everyone, to a person, who I have talked with on the air and who is running a modulated solid state transmitter, and who has eliminated or significantly reduced the safety factor blows devices.  The importance of the safety factor and using a conservative design cannot be overstated.

[WA1GFZ]

The FQA11N90 blows the doors off a IRF840. much better FET.
Yup I run 25 - 30  volts at carrier.
Class E waveform swings about 3.5 times drain voltage at turn off so you need plenty of safety factor in case of antenna issues.

[W1ATR]

If you don't wanna build a VFO then try the N3ZI DDS VFO.  Easy and inexpensive.  http://www.pongrance.com/ddfc.html

Hey John, did ya ever have any thoughts on using one of those to replace the veefo in the 750? Maybe a more custom job with an LED display that fits the window instead of the LCD it comes with? Kicking the output up enough for the driver wouldn't be too hard and maybe even bypassing the multiplier for direct dialing the whole HF band would be another thought. My veefo is a little quirky and takes days and clear 6ft work bench to take apart. My wavetek works fine for the task, but I was thinking more of something internal and just as stable.

Sry for the highjack. You can PM me if needed.
J 

[kb3ouk]

i did some searching on here and found this,

http://amfone.net/Amforum/index.php?topic=11017.0
http://www.w1vd.com/375wattclassD.html

i can't believe how simple the schematic is, now my question is, can i use an analog modulator on this, and what kind of vfo should i use? analog or digital, the schematic says digital so i guess thats what it needs, but i want to be sure, i'm really thinking about building a rig that is more reliable than what i have at the moment.

[w3jn]

Hey John, did ya ever have any thoughts on using one of those to replace the veefo in the 750? Maybe a more custom job with an LED display that fits the window instead of the LCD it comes with? Kicking the output up enough for the driver wouldn't be too hard and maybe even bypassing the multiplier for direct dialing the whole HF band would be another thought. My veefo is a little quirky and takes days and clear 6ft work bench to take apart. My wavetek works fine for the task, but I was thinking more of something internal and just as stable.

Sry for the highjack. You can PM me if needed.
J 

I use a HP3325A synthesizer on the GPT-750 - easy keyboard entry!

[kb3ouk]

for the class d transmitter i posted, i would assume that a low pass filter is needed, or isn't it? i'm guessing it is.

[steve_qix]

for the class d transmitter i posted, i would assume that a low pass filter is needed, or isn't it? i'm guessing it is.

A low pass filter is absolutely required, and the filter must be property designed, implemented and terminated in its characteristic impedance (usually, 50 ohms).  If you use an antenna tuner, you should be able to guarantee 50 ohms for a load.  If so, class D will work if you build it correctly. 

If you don't have an antenna tuner, or can't guarantee 50 ohms resiistive, use class E because it will pretty much work into any reasonable load from 25 ohms to well over 100 ohms.... and you can have some Z in there as well, and it will still work correctly.

[kb3ouk]

if i build it it will be running into an antenna cut for 75 meters that usually has a 1.5:1 swr, is that good enough? and i figured the filter was needed, i just wasn't sure as to why it wasn't mentioned on the website  (except for showing it in the pictures and i few breif mentions) or in the schematic.

[kb3ouk]

http://www.w1vd.com/375WclassD.pdf

ok, i think i answer one of my own questions. i wasn't loooking at the schematic right, the LPF is L1, L2, and C1 in the schematic, right? in that case, if i build the lpf seperate from the rest of the transmitter, then all i have to do to change from 160 to 75, if i want to, is change the lpf. looking around on jay's site, i found a 160/80 meter vfo and the class d transmitter, and i already have an idea how to modulate it, the only thing i haven't figured out yet is what kind of power supply to use for the rf deck. the schematic shows the drivers as ussing 12 volts, but i can't find any mention of what voltage the 11N90's  run on.