EMRFD Message Archive 6525
Message Date From Subject 6525 2011-08-01 14:15:24 w9ifz paralleling RF output transistors try to enter this before but made the mistake of using the enter key to many times.
I want to parallel two 2N7000 transistors in the output stage. I could not find a reference in emrfd but expect its there some place.
My question is --- the circuit is common emitter. Should I add a low value resistor 1 to 3 ohms in the emitter to ground line or each or should I add an additional base resistor so I am driving each through its own bias resistor.
thanks, Stan AK0B
6526 2011-08-01 14:20:45 Chris Trask Re: paralleling RF output transistors >many times.
> try to enter this before but made the mistake of using the enter key to
> I want to parallel two 2N7000 transistors in the output stage. I could
> a reference in emrfd but expect its there some place.value
> My question is --- the circuit is common emitter. Should I add a low
> resistor 1 to 3 ohms in the emitter to ground line or each or should I addan
> additional base resistor so I am driving each through its own biasresistor.
>Either method will work, though the first one is preferred.
N7ZWY / WDX3HLB
Senior Member IEEE
6527 2011-08-01 14:43:27 w9ifz Re: paralleling RF output transistors thanks Chris
6528 2011-08-01 23:35:53 Johan H. Bodin Re: paralleling RF output transistors Stan,
2N7000 is a MOS-FET. The gate-source turn-on voltage can differ more between
parts than it does with bipolar transistors. If you are building a linear
amplifier I'd suggest that you use separate source-to-ground resistors, as you
suggested, and also arrange for individually adjustable bias voltages for the
gates. If you are building a hard switching amplifier with stiff squarewave
drive, class E for example, you may get away by simply paralleling the transistors.
6529 2011-08-02 03:55:43 Tim Re: paralleling RF output transistors Look at e.g. KD1JV's PSK transceiver which uses 3 BS170's in parallel in the output to get 3 watts out:
He is doing Class C PSK and uses logic gates for the drive.
Also look at Dan Tayloe's NC2030 which uses (again) 3 BS170's in parallel, class E, and gets 5 watts out. Using logic gate drive for 2N7000 or even IRF510 (I parallel more gates for the IRF510) is incredibly easy. The NC2030 has a clever SWR protection circuit.
A BS170 is probably a little superior to a 2N7000 spec-wise but I note that several major manufacturers call the exact same part both a BS170 and a 2N7000. (e.g. Vishay).
For better linearity, resistors in the sources help immensely.
Some worry about matching but I've found BS170's/2N7000's from the same tape strip to be very close matches already. There's not really much reason to buy floor-sweeping BS170's/2N7000's because name-brand distributor parts are so cheap to begin with.
As to emitter (source in this case) resistor values, if you wanted to use it to enforce current-sharing in the linear region the value should probably be in the low tens of ohms for a 2N7000. A ballpark current at the 1 watt per device level would be 100mA, and you want maybe 30% of Vcc drop across the resistor at that current level. But as you saw above KD1JV side-steps this by using logic-gate drive to deeply saturate the parts.
In terms of a different way of doubling up on jellybeans, fig 12.73 of EMRFD shows two 2N3904's in push-pull as a driver. The idea works almost exactly the same with 2N7000's except you can cut the emitter resistor by a factor of a several and the 2N7000 bias voltage should be tweaked lower depending
6530 2011-08-02 08:32:22 w9ifz Re: paralleling RF output transistors Thanks fellows - appreciate the references. looks like i need to do more experimenting on the bench. i am lucky i have a roll of 2n7000 that i acquired at an excellent price so can afford to waste a few.