I've built a little RF amp (for 20 meters) using a single MTP3055V
MOSFET.
Class AB, binocular core RF transformer in the output. I drive it
with
a homebrew 20 meter DSB rig.

Here is the problem: The amp works fine IF I take the low pass filter
out of the circuit. I'm using a groundplane antenna with a T network
tuner.
I can adjust it to make it look like 50 ohms. Without the low pass
filter,
I get a very nice clean output (looks good on a scope, and SWR is low,
so I don't think it is putting out spurs or harmonics).

But if I include the LP filter in the circuit, the amp takes off,
turning
into an 18 Mhz oscillator!

I checked the filter with a sig generator. The cutoff freq is where
it is
supposed to be.

I've had this kind of problem before, with other amps.

Any suggestions?

73 de Bill M0HBR CU2JL N2CQR
http://planeta.clix.pt/n2cqr

Unlike gain, stability is a DC-to-daylight phenomenon.
Meaning...the LPF presents the "Right" impedance to the amp output in
the passband, but consider: the LPF is a gross mismatch in the
stopband. Evidently the output of the amp is only conditionally
stable above the LPF cutoff, and the stopband impedance of the LPF
lies in the unstable region.
This kind of thing can be quite common when not considering stability
across all possible frequencies that the amp has a gain1.


On 26 Sep 2004 12:11:07 -0700, (Bill N2CQR
MOHBR) wrote:

I've built a little RF amp (for 20 meters) using a single MTP3055V
MOSFET.
Class AB, binocular core RF transformer in the output. I drive it
with
a homebrew 20 meter DSB rig.

Here is the problem: The amp works fine IF I take the low pass filter
out of the circuit. I'm using a groundplane antenna with a T network
tuner.
I can adjust it to make it look like 50 ohms. Without the low pass
filter,
I get a very nice clean output (looks good on a scope, and SWR is low,
so I don't think it is putting out spurs or harmonics).

But if I include the LP filter in the circuit, the amp takes off,
turning
into an 18 Mhz oscillator!

I checked the filter with a sig generator. The cutoff freq is where
it is
supposed to be.

I've had this kind of problem before, with other amps.

Any suggestions?

73 de Bill M0HBR CU2JL N2CQR
http://planeta.clix.pt/n2cqr

On 26 Sep 2004 12:11:07 -0700, (Bill N2CQR
MOHBR) wrote:

Here is the problem: The amp works fine IF I take the low pass filter
out of the circuit. I'm using a groundplane antenna with a T network
tuner.
I can adjust it to make it look like 50 ohms. Without the low pass
filter,
I get a very nice clean output (looks good on a scope, and SWR is low,
so I don't think it is putting out spurs or harmonics).

Suggestion, try a low pass to the antenna and a highpass to a 50ohm
load. then the amp[ sees 50 ohms for all loads and also the high pass
section will absorb harmonics. The pow pass and the highpass share a
common port (the one looking toward the amp). This is documented in
EMDRF and other places.

I'd also look at the layout and maybe try loading the drain choke with
a resistor in the 40-100ohm range. The choke may be resonant at some
nearby frequency.

One of the problems with MOS power FETs is there is substantial drain
to gate capacitance and if the gate is not loade or the drain enough
results and feedback can be problematic. You can also try loading the
gate to ground with a resistor in series with a cap (try 50 ohms and
..01uf) or a series combo of 100ohms and .01uf from drain to gate to
add negative feedback.

Allison

Just a 50 cent plug for the yahoo group rfamplifiers. You might join the
group and ask your question over there. More than one RF Amp Engineer
hangs around to bounce your questions off.

good luck
cheers
skipp

just go to the main www.yahoo.com page, type in yahoo groups. Then type in
rfamplifiers and join.


: Bill N2CQR MOHBR wrote:
: I've built a little RF amp (for 20 meters) using a single MTP3055V
: MOSFET.
: Class AB, binocular core RF transformer in the output. I drive it
: with
: a homebrew 20 meter DSB rig.

: Here is the problem: The amp works fine IF I take the low pass filter
: out of the circuit. I'm using a groundplane antenna with a T network
: tuner.
: I can adjust it to make it look like 50 ohms. Without the low pass
: filter,
: I get a very nice clean output (looks good on a scope, and SWR is low,
: so I don't think it is putting out spurs or harmonics).

: But if I include the LP filter in the circuit, the amp takes off,
: turning
: into an 18 Mhz oscillator!

: I checked the filter with a sig generator. The cutoff freq is where
: it is
: supposed to be.

: I've had this kind of problem before, with other amps.

: Any suggestions?

: 73 de Bill M0HBR CU2JL N2CQR
: http://planeta.clix.pt/n2cqr

Bill N2CQR MOHBR wrote:

But if I include the LP filter in the circuit, the amp takes off,
turning
into an 18 Mhz oscillator!

I have found that the best way to build a stable amplifier is
to try and build an oscillator. To build a guaranteed to start
first time, every time oscillator, just try to build a stable
amplifier :-)

Follow good HF/VHF practice with the circuit layout.

Negative feedback can sometimes help (or make the problem worse).

Pay careful attention to power supply decoupling. The very
high LF gain of RF transistors can cause parasitic oscillation
at the resonant frequency of an L/C or R/C decoupling circuit.

Try the usual parasitic suppression tricks. A low value resistor
or ferrite bead in series with the FET gate. A ferrite bead
in series with the FET drain. A resistor in parallel with the
drain load inductor or transformer, as suggested by Allison.

Sometimes it can be easier and faster to start again with
a different layout than to debug a 'bad' circuit.


73, Ed. EI9GQ.



--
Remove 'X' to reply via e-mail.
Linux 2.6.7