Thanks for the reply. The EB63 is in a box right now so I'm playing with the
other one. The toroid with a resistor seems to help it but a piece of wire
all but eliminated it. I don't know the properties of the core but it has
the same dimensions and was wound with #20 wire. This amp uses 2 GE 104P1
transistors that I had around here. I think it originally had TRW
transistors and an 80 watt output. Is it OK to vary the resistor and number
of turns to achieve best results?

Thanks,
Chris
wrote in message
...
| On Tue, 07 Sep 2004 17:14:00 GMT, "Chris"
| wrote:
|
| The driving radio is an RCI2950. The radio output goes directly into the
| input transformer. There is a VSWR of 1.2:1 between the radio and the
amp. I
| also tried adding various caps on the power lines to no avail. As
mentioned
| earlier, I have another amp with MRF455's with no problems. The only big
| difference I see in the circuits is thatr the good amp has an RC network
| between the input transformer and the bases.
|
| Chris,
|
| That amp was designed around MRF454 not 455 there is a difference
| and the 455 has a tad more gain and different match. I might add the
| MRF454 is a 80W device and the 455 is only 60w as well. The
| differences are not always apparent and the output match and reverse
| transfer are very different. Pull the datasheets for both off the
| 'net and look at them.
|
| The design requires a matched pair. The bias seems to be best
| at around 200-300ma per transistor. Any asymetery in grounding
| (that includes mounting the PCB to the heatsink) can also cause
| problems. The transformers MUST be wound as specified using
| materials as specified.
|
| If you going to try and tame the beast. Try this. From collector to
| base on each transistor: series connect, 12ohms (2w non inductive),
| .33uH molded choke and capacitor pair (10uf and .1uf in parallel).
| the leads for these must be as short as possible (none or nearly so).
| This may work. If not:
|
| Feedback version two: FT50-43 torid with 10t #26 from collector to
| collector, pass through the center a 4.7ohm 1/2watt resistor and
| connect the leads to the bases. If the leads are backward it will
| oscillate in that c ase exchange the leads for the 10turn winding
| between the collectors. If this doesnt work used MRF454s.
|
| If feedback works expect about 1-2db of gain reduction and a
| corosponding increase in required drive for the same output power.
| However, do NOT exceed 5W input power of the transistors will cook.
|
| Also if you scope the power leads at the junction of the output
| transformer and see anything going on there in the below 5mhz range
| you bypassing on the board is wrong or grounding ineffective.
|
| One OBTW: input and output connectors, case and grounds. I've seen
| one that when on the bench open worked as predicted. In the case
| built by the owner it would take off. Seems the case and the SO239s
| mounted to it were RF wise not at the same ground as the groundplane
| of the PCB and created a untidy feedback path.
|
| I've built "EB63", it was featured as an app note from Moto some 20+
| years ago and it was a stable design using the MRF454s. I had little
| trouble with it as a base design for a 18-35mhz commercial
| application.
|
|
| Allison