"Telstar Electronics" wrote in message
oups.com...
Frank Gilliland wrote:
Ok. So what does that have to do with the difference between a PCB
mounted amp and a chassis-wired amp?

I think the problem here is always one of repeatability. Anyone can
fiddle around for a time impedance matching and make one amp work...
but when you plan to make a number of units... having wires running all
around, and putting components in different physical locations
unit-to-unit makes performance unpredictable. There is another problem
that shouldn't be overlooked about the wiring method... that is the
problem of manufacturability. The PCB version is usually much easier
and faster to build... not a insignificant advantage when building
numerous units.

www.telstar-electronics.com

Point to point wiring does not have to mean the birds nest constuction you
describe. RF wiring could well be stamped out metal pieces mounted on Teflon
or ceramic. very repeatable, very stable. It could also mean the isolted pad
constuction where smalll sqares of PC board are glued to a substrate. This
has proven reliable, obviously an improvement over PC board material could
be found.and better mounting methods than glue.

There are two big problems I see with most amp designs. One is biasing, even
the cheapest Boomer would be much improved if it had an adequae bias
circuit. The other is proper setup. I doubt if someone mostly interested in
how well they can make the power meter swing will pay much attention to
setup instuctions. Attenuators on the front end would be a joke. How would
establish proper drive for a given setup with only a cheap power/VSWR meter.
FET amps really dont have a stability problem if properly designed. The idea
of using a switching power supply to take the voltage up to 50 volts or more
is a good one that has already been proven in audio amps. It is impossible
to make an amp idiotproof, just when you thingk you have along comes a
better idiot.

Jimmie D wrote:
Point to point wiring does not have to mean the birds nest constuction you
describe. RF wiring could well be stamped out metal pieces mounted on Teflon
or ceramic. very repeatable, very stable. It could also mean the isolted pad
constuction where smalll sqares of PC board are glued to a substrate. This
has proven reliable, obviously an improvement over PC board material could
be found.and better mounting methods than glue.

Yes, I agree with you. The problem is that these amp builders are not
using the point-to-point methods you describe. From what I can see...
they are gluing pieces of copper clad board (via double sticky tape) to
a non-etched copper-clad board. They have no regard for characteristic
impedance of trace widths
(http://www.pcb123.com/help/calculators/microstrip.html). I also am not
really crazy about the longevity of double-sticky tape... LOL


There are two big problems I see with most amp designs. One is biasing, even
the cheapest Boomer would be much improved if it had an adequae bias
circuit. The other is proper setup. I doubt if someone mostly interested in
how well they can make the power meter swing will pay much attention to
setup instuctions. Attenuators on the front end would be a joke. How would
establish proper drive for a given setup with only a cheap power/VSWR meter.


Agreed, my new amplifier design
(http://www.telstar-electronics.com/S...202879ABTC.pdf) now has
thermal tracking class AB biasing. I also don't use any front-end
attenuators. Like you say, proper setup is critical to any good
station. The operator needs to pay attention to the drive requirements
of the amplifier... and should adjust the driver accordingly.

www.telstar-electronics.com

On a sunny day (24 Sep 2006 07:20:57 -0700) it happened "Telstar Electronics"
wrote in
. com:

Jimmie D wrote:
Point to point wiring does not have to mean the birds nest constuction you
describe. RF wiring could well be stamped out metal pieces mounted on Teflon
or ceramic. very repeatable, very stable. It could also mean the isolted pad
constuction where smalll sqares of PC board are glued to a substrate. This
has proven reliable, obviously an improvement over PC board material could
be found.and better mounting methods than glue.

Yes, I agree with you. The problem is that these amp builders are not
using the point-to-point methods you describe. From what I can see...
they are gluing pieces of copper clad board (via double sticky tape) to
a non-etched copper-clad board. They have no regard for characteristic
impedance of trace widths
(http://www.pcb123.com/help/calculators/microstrip.html).

At 27MHz 50Ohm this is not critical.


I also am not
really crazy about the longevity of double-sticky tape... LOL

Maybe it is epoxy glue, how would you know.

Agreed, my new amplifier design
(http://www.telstar-electronics.com/S...202879ABTC.pdf) now has
thermal tracking class AB biasing.

Yes that is good.

I also don't use any front-end
attenuators. Like you say, proper setup is critical to any good
station. The operator needs to pay attention to the drive requirements
of the amplifier... and should adjust the driver accordingly.

There are many sets on the market, somebody will drive it with too much,
and somebody else will drive it with too little.

Tough world out there.
;-)

www.telstar-electronics.com

Jan Panteltje wrote:
At 27MHz 50Ohm this is not critical.

Huh???? Since when is impedance matching at RF frequencies not
important? Then maybe you could tell us at what frequency it is
important... LOL

www.telstar-electronics.com

On a sunny day (24 Sep 2006 10:03:00 -0700) it happened "Telstar Electronics"
wrote in
. com:

Jan Panteltje wrote:
At 27MHz 50Ohm this is not critical.

Huh???? Since when is impedance matching at RF frequencies not
important? Then maybe you could tell us at what frequency it is
important... LOL

You really have no experience now have you?
First the little PCBs just form some extra capacitance to ground.
There is no transmission line.
And you may need that anyways, why else solder those caps over your
output transformer.
Second, for a transmission line L and C are important, and L is close
to zero here, and f is very low.
The 'length' of the track is very very short in their design.
You can hardly speak of a transmission line at all at 27MHz.
Not a significant part of the wavelength anyways.
'Transmission line' properly matched only implies no refections
and some known impedance after all.
And where is this 'line'?

You mentioned that calculator, but did you ever use it?
If so what did you enter for the constants?

"Jan Panteltje" wrote in message
...
On a sunny day (24 Sep 2006 10:03:00 -0700) it happened "Telstar
Electronics"
wrote in
. com:

Jan Panteltje wrote:
At 27MHz 50Ohm this is not critical.

Huh???? Since when is impedance matching at RF frequencies not
important? Then maybe you could tell us at what frequency it is
important... LOL

You really have no experience now have you?
First the little PCBs just form some extra capacitance to ground.
There is no transmission line.
And you may need that anyways, why else solder those caps over your
output transformer.
Second, for a transmission line L and C are important, and L is close
to zero here, and f is very low.
The 'length' of the track is very very short in their design.
You can hardly speak of a transmission line at all at 27MHz.
Not a significant part of the wavelength anyways.
'Transmission line' properly matched only implies no refections
and some known impedance after all.
And where is this 'line'?

You mentioned that calculator, but did you ever use it?
If so what did you enter for the constants?

Obviously you kneel before the SWR gods. You dont have to have a transmision
line to have a matching problem. Matching means to have the output impedance
of a driving stage equal to the input impedance of a driven stage so maximum
power is transfered with minimum distortion. Nothing was said of
transmission lines or SWR.

On a sunny day (Sun, 24 Sep 2006 16:50:55 -0400) it happened "Jimmie D"
wrote in
:


"Jan Panteltje" wrote in message
...
On a sunny day (24 Sep 2006 10:03:00 -0700) it happened "Telstar
Electronics"
wrote in
. com:

Jan Panteltje wrote:
At 27MHz 50Ohm this is not critical.

Huh???? Since when is impedance matching at RF frequencies not
important? Then maybe you could tell us at what frequency it is
important... LOL

You really have no experience now have you?
First the little PCBs just form some extra capacitance to ground.
There is no transmission line.
And you may need that anyways, why else solder those caps over your
output transformer.
Second, for a transmission line L and C are important, and L is close
to zero here, and f is very low.
The 'length' of the track is very very short in their design.
You can hardly speak of a transmission line at all at 27MHz.
Not a significant part of the wavelength anyways.
'Transmission line' properly matched only implies no refections
and some known impedance after all.
And where is this 'line'?

You mentioned that calculator, but did you ever use it?
If so what did you enter for the constants?

Obviously you kneel before the SWR gods. You dont have to have a transmision
line to have a matching problem. Matching means to have the output impedance
of a driving stage equal to the input impedance of a driven stage so maximum
power is transfered with minimum distortion. Nothing was said of
transmission lines or SWR.

Oh I am not questioning 'impedance matching', but some points.
1) He refers to that calculator, that also asks for a track length and
load capacitance. In case of distribited loads, and there are of course.
So we need his numbers.

2) The impedance is complex, not simple 'Ohm', and in this case nearly
100% a capacitance, basically the same (leaving out the extra value of the
tape or glue), as between a top and bottom PCB layer, you take the surface
area, and get a capacitance, it is almost a square 'pad' they use (see
pictures) not a stripline! (and so no transmission line with separate
in and outputs either).

So we ask him again: What values did you enter, and what dielectric constant
for the PCB material and 'tape'.

Yes SWR must match but really that is matching of the antenna to the
transmitter, we are here internally, at the most, if there was one!!!!,
we were here in the C of an CLC Pi filter, and impedance matching.
And that is linked to SWR for a known 50 Ohm load in this specific case,
if the antenna is also 50Ohm.
So, any Pi filter in the output could 'match' the 50 Ohm antenna to any
other (higher or lower) transmitter output impedance.
If the input C (the PCB pad cap) changes, the L and output C of the Pi filter
will need to be adjusted.
He does not provide a diagram, but question arises if he HAS a Pi filter
to reduce harmonics.

How ever you slice it, teh few extra Pf from those pads wll make little
difference, maybe he even has a double sided PCB with the same problem
of capacitance to the backplane.
It is all just a way to frabricate as many posts pushing his amp as possible,
right or not makes no difference, it is not some open source project, secret
diagram, just a marketing troll.
No problem with me, but I'd love to hear wha the entered in that calculator,
at least if you recommend something you should have used it once.....

So, now for the next episode of 'the better amp'.
Grin

Sorry it is late, need to get up in a few hours,

Jimmie D wrote:
You dont have to have a transmision
line to have a matching problem. Matching means to have the output impedance
of a driving stage equal to the input impedance of a driven stage so maximum
power is transfered with minimum distortion. Nothing was said of
transmission lines or SWR.

Right on Jim... impedance matching is always important for te reason
you stated.

www.telstar-electronics.com