Help identify RF transistor

[Omicron]

I have a Berkeley Nucleonics 6040 pulse generator with the 201E fast rise-time plug-in. The plug-in seems defective as I can't properly set the pulse levels. The output stage uses some sort of RF transistor. See attached picture. They have a code R5 printed on them.

Anyone have an idea of what part this could be?

[fzabkar]

My 1984 Motorola Master Selection Guide shows a package which looks similar to yours, namely CASE 337-02 (u .290" Flange). The only parts with that package are 1.7 - 2.3 GHz Broadband CW transistors which are designed for Class B and C common base amplifier applications.

MRF2001M, MRF2003M, MRF2005M, MRF20010M, MRF20016M

The last 2 digits are the output power, eg MRF20016M has a power rating of 16W.

[Omicron]

The package indeed looks very similar. But the signals don't seem to fit a common base. See attached the input and output signals on the left most transistor which I think is still working correctly. The middle terminal with the screws is connected to ground. The yellow signal is measured on the left terminal and the blue one on the right terminal. The signal is inverted which it shouldn't be for a CB. Also the input signal seems a bit large at 5V if there is a base-emitter diode between the left terminal and ground. Could it be some kind of FET maybe?

[shakalnokturn]

Mitsubishi has some similar looking FET devices in GF-17 encapsulation but these are all >10GHz.
By the marking style I would rather have guessed at Motorola though.

[fzabkar]

This datasheet has the pinout and dimensions for Mitsubishi's GF-17 package:

http://www.mitsubishielectric-mesh.com/products/pdf/mgf2430a.pdf

[Omicron]

Thanks guys, that looks promising. A GaAs FET makes sense, I've read that other fast edge pulse generators also use them. The waveforms seem to fit it as well.

[T3sl4co1l]

It can be base, it's only forward-biasing a volt or so by the looks of it.  JFET and MESFET may also share that behavior though.  The large input signal level does seem to suggest FETs, as it shouldn't be needed for a BJT (but might still be done out of convenience, who knows).

Tim

[picburner]

Agilent (ATF-45101) and Fujitsu (FLX102MB) also produce GaAs-FETs in that package.

[Omicron]

I think the transistors are both still OK. If I force -5V on the gate of the right transistor then the resistance between the source and drain goes from  about 2 ohms to around 20 kilo-ohms. That seems to confirm it is indeed a depletion mode FET.

I think the issue is with the circuit that is supposed to provide the drain voltage for the output transistor. There are two DACs in this device. One sets the source voltage and thus determines the low level of the output pulse. The other sets the drain voltage and so determines the high level. The circuit between the DACs and the transistor seems quite involved though having a bunch of OPAMPS and power transistors in some sort of control loop. I'm still trying to figure it out.

Nobody happens to have the schematic for the Berkeley Nucleonics 201E fast rise-time plugin for their 6040 pulse generator? I asked them but they say they can't provide it anymore due to company policy. I'll put this question in a new thread maybe.

[djidji]

they are for sure some GaAS Fet's and on your picture could not be seen well is there some capacitor in series with that inductor between source of first FET and gate of the second FET?

[Omicron]

I believe it is a capacitor, yes. It's hard to see even in real life. There also seems to be an OPAMP circuit that senses the voltage on the left side of the capacitor and drives the right side.

[djidji]

did you measure the voltage across it? on gates should be something in range -0,35V to -1,2V and more than +3V on sources.

[Omicron]

A small update. Despite my earlier statement it did turn out that the transistor was faulty. It seemed to test OK at DC, but in actual operation it was not turning off properly. I replaced it with an HP ATF45101 and now the pulse generator is operational again. The rise time is a little below spec (230pS vs 180pF), but I'll call it good enough for now.

Thanks for all your help and suggestions!