I have the MRF454 (datasheet attached) and the metal base which you screw down to the heatsink, is this the emitter?

 

I have the MRF454 (datasheet attached) and the metal base which you screw down to the heatsink, is this the emitter?

Your ohmmeter is your friend.
Tabs 1 and 3 are the emitter. Tab 4 is the collector. Check continuity between the tabs and the screw holes.

 

Thank you

 

I'd expect it to be completely isolated from all pins.

 

Actually we are having a problem right now. When the bottom of the transistor package touches the heat sink the base voltage just drops to 0V and we get no output. When we take the screws off and lift the transistors slightly we get output (mind you we are using a very weak input RF signal).

We are extremely frustrated because the heatsink is grounded and despite the fact that the base metal of the transistor not connected to any of the pins when it touches the heatsink through the mounting screws or direct contact (maybe our rubber is conducting, who knows) but somehow it is shorting the base pin of the transistor.

 

Actually we are having a problem right now. When the bottom of the transistor package touches the heat sink the base voltage just drops to 0V and we get no output. When we take the screws off and lift the transistors slightly we get output (mind you we are using a very weak input RF signal).

We are extremely frustrated because the heatsink is grounded and despite the fact that the base metal of the transistor not connected to any of the pins when it touches the heatsink through the mounting screws or direct contact (maybe our rubber is conducting, who knows) but somehow it is shorting the base pin of the transistor.

Post your schematic.

 

Here it is:

 

Are you using the same PC layout as the app note?

 

For the construction of the kit we used this file the company we bought the kit off of provided.

As far as I'm concerned, the app note and the other file the company provided us are the exact same pcb board with the exception of using a few alternative components in place of what is noted on the app note since those components are no longer available (ie. the voltage regulator, the pass transistor, etc).

I don't think those are the issues. I just don't understand why when we bolt down the transistors to the heat sink the base short circuits. The metallic packaging is not connected to any of the pins (we tested via a continuity check using a multimeter). Somehow, isolating the metallic base from the heatsink allows the transistors to operate but now it's not fully connected to the heatsink which is why they're smoking up fast. Instead of the heatsink paste, we used 3M silicon rubber that is better in terms of thermal conductivity than heatsink paste and also provides isolation so the metallic packaging doesn't touch the heatsink but now the issue lies with the screws. The screws we're using are also metallic and its unavoidable since it bolts it down to the heatsink and its going to touch. We haven't tried nylon screws doh, although we're speculating those might melt easily when the 140W amplifier is at rated power.

What baffles me in this situation is why the metallic packaging can't be connected to the heatsink. It's clearly not connected to any of the pins so I don't even know why this is an issue in the first place. Jeez....

 

Are you using the same PC layout as the app note?

There are more details in his first thread about the project, here;
http://forum.allaboutcircuits.com/showthread.php?t=82725

Maybe moderators might want to merge the 2 threads?

I think it could be a capacitance issue, (just guessing) from grounding the metal cases of the RF transistors. RF circuits can be really fussy as to construction details!

 

Are you setting up the bias for class A, AB, or B operation? Class B (and AB, to a lesser extent) will have a threshold. If you are applying a small RF signal to the base, grounding the case may introduce enough capacitance to put the input signal below the threshold.

 

We set the bias voltage based on how much idling current we wanted in the collector. The manual says to set the idling current to at least 100mA per transistor or 200mA for both at the least. We set it to 250mA for both transistors in total by varying the potentiometer and this yielded a base bias voltage of 0.68V which is what the manual also predicted would be optimal. This voltage gets reduced to 0V after we observe the output of the amplifier reaching 20W and then they start smoking up and the base voltage all the sudden reduces to 0V. I just hope the transistors are not damaged due to the smoking....