Ive seen it proposed that you connect to the valve anode .via a high
voltage cap ...never tried . looked a bit risky with 3/4 kv of rf ..
but i suppose that would prevent the 'suck out' from the pi tank
cct .. the rx would be looking through the cct , as long as the vlave
was biased off .. not too much noise ?

G ..


On Aug 23, 1:15*am, "COLIN LAMB" wrote:
I vaguely recall that some T/R electronic switches added the bandswitch to
eliminate suckout.

73, *Colin *K7FM

On Wed, 27 Aug 2008, Graham wrote:

Ive seen it proposed that you connect to the valve anode .via a high
voltage cap ...never tried . looked a bit risky with 3/4 kv of rf ..
but i suppose that would prevent the 'suck out' from the pi tank
cct .. the rx would be looking through the cct , as long as the vlave
was biased off .. not too much noise ?

G ..

I thought that was the common form, certainly in homebrew articles
that seems to predominate. Yet, the commercial ones are likely to
be standalone, so one doesn't have to fuss with the transmitter.

I've never used one, and never given it much thought. I thought
the suckout issue was when it was connected to the plate of the
transmitter, but that's more likely because the suckout issue
was a common issue with TR switches, and I was more familiar with
connecting it to the plate of the transmitter's output tube.

One thing, it certainly was not uncommon to see articles where
the TR switch included a stage of amplification and a tuned circuit
of some sort.

Also, when connecting to the plate, they'd use a very small value
capacitor, and one article I just checked said to use an even smaller
value for a higher power transmitter. That's going to cause some sort
of loss, indeed that article mentioned one might need to add a tuned
circuit to the TR switch to compensate for the loss.

MIchael VE2BVW

On Aug 27, 5:19*pm, Michael Black wrote:
On Wed, 27 Aug 2008, Graham wrote:
Ive seen it proposed that you connect to the valve anode .via a high
voltage cap ...never tried . looked a bit risky with 3/4 kv of rf ..
but i suppose that would prevent the 'suck out' from the pi tank
cct .. the rx would be looking through the cct , as long as the vlave
was biased off .. not too much noise ?

G ..

I thought that was the common form, certainly in homebrew articles
that seems to predominate. *Yet, the commercial ones are likely to
be standalone, so one doesn't have to fuss with the transmitter.

I've never used one, and never given it much thought. *I thought
the suckout issue was when it was connected to the plate of the
transmitter, but that's more likely because the suckout issue
was a common issue with TR switches, and I was more familiar with
connecting it to the plate of the transmitter's output tube.

One thing, it certainly was not uncommon to see articles where
the TR switch included a stage of amplification and a tuned circuit
of some sort.

Also, when connecting to the plate, they'd use a very small value
capacitor, and one article I just checked said to use an even smaller
value for a higher power transmitter. *That's going to cause some sort
of loss, indeed that article mentioned one might need to add a tuned
circuit to the TR switch to compensate for the loss.

* * MIchael *VE2BVW

Yes , that was as far as I went , made a dual gate fet t/r switch, but
it needed a tuned cct , did work on the low z side but was down in
gain ...

G ..

"k3hvg" wrote in message
. ..
Bob Spooner wrote:
"Richard Knoppow" wrote in message
...
Are we talking about RF here?

Yes. I was triggering the scope on the transmit RF and looking at the RF
on
the output of the receive port across 51 ohms. 73, Bob AD3K



That about +16dBm. Shouldn't do any real harm. Healthy signal but
non-destructive.

Electronic switches might not have the isolation you would prefer. With a
single band you have the luxury of hanging a stub on the RX port to short
the input during TX Otherwise, something should do that fast enough. Since
there is 100w pep and I assume you measure 4v pp that is 16/51= .314 w pep
so:

10 log 100/.314 = 25 dB isolation. Not real swift. I would want better
than 35 dB Isolation. Even more for QSK. I wouldn't want to wait for the
RX to recover. I have equipment that burns out the input protection above
..25 w input (ave.)

You might not have trouble with a tube RF amp but...

"JB" wrote in message
news:fuAtk.914$UX.763@trnddc03...

Electronic switches might not have the isolation you would prefer. With a
single band you have the luxury of hanging a stub on the RX port to short
the input during TX Otherwise, something should do that fast enough.
Since
there is 100w pep and I assume you measure 4v pp that is 16/51= .314 w pep
so:

10 log 100/.314 = 25 dB isolation. Not real swift. I would want better
than 35 dB Isolation. Even more for QSK. I wouldn't want to wait for the
RX to recover. I have equipment that burns out the input protection above
.25 w input (ave.)

You might not have trouble with a tube RF amp but...

The problem with a stub is that you still have to disconnect it somehow
during receive, right? I want to work full break in without listening to a
mechanical relay. From my tests, I've found that the T-R switch attenuation
isn't linear - higher power levels turn the switch off faster and produce
more isolation than low power levels, as would be expected with an active
device that used a sample of the output signal to drive a grid negative for
switching. My old Valiant II had an output of 275W for CW and isolation was
no problem for the HQ-170A receiver. I could hear my transmitted signal and
tell whether it was clean or not. Of course, the AGC on the 170A might be
better than the older generation NC-125 and NC-173 I'm planning to use. The
T-R switch is rated for up to 4KW, so it was just loafing even with 275W.
73, Bob AD3K