Question about protecting amplifier input

[K1JJ]

Hola,

I built up one of those MRF-150 power mosfet linear amplifiers using 4 FETS.  It has 48V on it and puts out an easy 600W pep with 10w pep drive.

It needs only about 10-15w pep of drive MAXIMUM. I want to use one of my riceboxes to drive it, but the ricebox is capable of 200w output - this would fry the FET gates. I now turn back the drive power, but there will be a time when I make a mistake and wipe out the whole FET string. I need to use a few different rigs as drivers, so this is prone to errors.

I understand that using a fuse or latching relay to protect the gate input is too slow, as the FET junctions can be destroyed in 1/2 microsecond!

Frank/GFZ suggested using a "high power clipper."  I'm not sure if this includes using transorbs or something else.

Before I investigate further, does anyone have suggestions to solve this problem with a high speed protection device or circuit for the gates?  The amplifier works on 160-10M.


Bottom line is  I need to limit RF power to the FET gates to a maximum of about 15 watts - when available power can be as high as 200w by error. The FET amplifier's input impedance is 50 ohms.

Thanks.

Tom, K1JJ

[WA1GFZ]

3 dB power pad with a clipper after it. You can build a clipper with high speed diodes and power transorbs. Transorbs alone have too much junction C and RF will cook  in time. You use high speed diodes to isolate the transorb then bias the junction of the diode transorb to just below the transorb fire voltage. The diode is reverse biased until the RF level gets one diode drop above the bias voltage then it sees the junction C go higher and the transorb conducts. You have to build one for the positive peak and another for the negative peak.
This is very fast. fuses relays way too slow. I've also seen pin diodes used as clippers but a 200 watt pin is hard to come by. MaCom used to make them years ago.

[WB2EMS]

I had thought about this for a similar situation. My 706 has a bug in that in AM mode, when you key it down it initially spikes up to full output before it settles down to the 25 watt carrier level. I had thought about building an RF powered ALC circuit that would sample and rectify some of the RF and create an ALC feedback voltage to hold down the output so I don't toast the grids on my Titan amp. I never got around to bread boarding it to see how fast the ALC responded,  but might before winter's out. Wouldn't depend on me remembering to turn the output down and wouldn't depend on a battery that might go dead at the wrong time.

Don't know if that's a useful thought or not. If you're swapping exciters you might need to adjust the ALC voltage for each different unit, so that might not make it a good solution for this specific case. The other question would be whether the ALC would respond fast enough to keep the output from rising high enough to damage anything.
 

[n2bc]

A 12dB attenuator... run the exciter at full power - they are usually much cleaner at max power rather than reduced power.   

A T-pad  would be   a 30 ohm 100W at the input, to a 30 ohm 10W to the output and a 30 ohm 60W at the junction of the other two to ground.   90 ohm 30W resistors in a TO-220 style package are about $7 each, smaller wattage jobs cheaper for the other values.

I've built a couple similar attenuators with very good results.

Another source:   ePay for a 10dB Bird atten and build a much smaller wattage pad for it's output.

I have a Harris military amp that only needs 150mW drive... much safer to pad the heck out of the input and let the exciter run it it's 'sweetspot'.  Additional benefit: the exciter is always looking at a 50 ohm resistive load.

73, Bill  N2BC

[w3jn]

I had thought about this for a similar situation. My 706 has a bug in that in AM mode, when you key it down it initially spikes up to full output before it settles down to the 25 watt carrier level.

I noticed this with my FT1000MP.  What was happening was that the "click" from the PTT switch in the hand mike was causing the RF spike.  Using the footswitch and a different mike solved the problem.

[K1JJ]

OK on the pad attenuator, Bill.

Yes, that would be a good idea if I were to run full power from the driver. But I want to idle the FT-1000D up to class A and use it at about 20W maximum out for the cleanest signal possible. I think by running it in class A, 20W, it will produce a cleaner signal than class B at full power. (I hope - this is still to be determined)  This wud require a 3db pad.   However, there is a chance I could still generate a 200w  power spike and blow the fets.

I also want to use a FT-102 that will run at 10W, but can also generate spikes at 150w. 

So, maybe the 20 watt power clipper is the way to go for this application. Frank's idea sounds good, but requires some parts and adjustments, whereas the 200W pin diodes sound quite good, if they are available cheaply.

Thanks for the ideas. I'll keep thinking and looking until I decide on something.

BTW, here's the spec sheet on Macom Power Pin diodes: (DownEast Microwave in NJ carries them)
http://www.downeastmicrowave.com/PDF/ma4p.PDF

Do I simply use 50 ohms at 20 watts to determine the clamp voltage? 

From the low sustained power ratings I see, I assume that if they are banged with 200w for a few seconds, they are toast, short out, but protect the circuit?

73,

Tom, K1JJ

[W2VW]

Is there a way to drive the sources instead of the gates without launching the FETs?

[K1JJ]

Is there a way to drive the sources instead of the gates without launching the FETs?

Hi Dave -

I suspect that wud work, just like a GG tube amp.

Though, the module is already built and works FB. I just wanna add a little box in line with the input coax to limit power spikes.

Frank sent me a circuit that we're discussing. I'll post whatever I decide on and build up.

BTW, what's your opinion on eBay prices for used ham gear? Are they falling, getting thin or holding up?  Are you seeing real bargains hitting the mkt?

I notice prices of some parts are kinda holding firm. I'm still looking for a spare power tube for my linear.  The Chinese new parts bargains still amaze me.

T

[KB1ORY]

Hi JJ and the group....

 It is interesting that the overdrive problem was brought up today because
I just 'toasted' the input tuning circuit and the finals in my 600 watt MFJ amp....
 My Kenwood TS-850 power control circuit is poor at best, it puts out approximately
four times the carrier power on 80m than 160m and when I went down to 160m I
forgot to readjust the carrier.... Y' all know what happened after that !!!!

Old Edd KB1ORY

[W2VW]

Ebay prices for old gear fell as the price of gas went up. They are still way off. Good time to stock up.

[K1JJ]

Hi JJ and the group....

 It is interesting that the overdrive problem was brought up today because
I just 'toasted' the input tuning circuit and the finals in my 600 watt MFJ amp....
 My Kenwood TS-850 power control circuit is poor at best, it puts out approximately
four times the carrier power on 80m than 160m and when I went down to 160m I
forgot to readjust the carrier.... Y' all know what happened after that !!!!

Old Edd KB1ORY

Edd,

Boy, that sux.  It goes out in an instant.

Don't feel bad. I know a guy who lost four 8877's in one mistake when he drove it with a 3-500Z linear amp... :-)  That was expensive.

It seems your's is a fairly common problem.  I'll try some stuff and see what I come up with using Frank/GFZ's guidance.
I'll post anything that works well.

Dave:  I think the price declines have just begun.   There's been no liquidation panic yet. When you hear them complaining that no one wants ham gear anymore... that's the time to buy... :-).

T

[n2bc]

Sometimes no amount of protective gizmos can offset one's creative stupidity. 

I had re-arranged the shack one weekend. Stripped out everything, cleaned the dust bunnies, moved the tonnage back in and cabled it all back up.

I was happily yacking on my KWS-1 the next morning when I noticed a curl of smoke coming out of my TenTec amplifier - which was not even turned on.

Yep, I had cabled the Ten Tec on the output of the 'station' selector switch so the KWS-1 was running through the Ten Tec.  Seems the QSK RF sniffer on the Ten Tec input didn't care for the KWS-1's 600 watts.  Quite the charred mess.

As I advance in age (and creative stupidity), I find that simpler is better....  The opportunity for destruction increases as the square of the complexity.  One station is more than enough!

Nite all,  73,  Bill  N2BC

[WA1GFZ]

This is not a big deal. Try protecting an input form 20 or 30 KW of induced lightning.