Hi,
I have a few ideas in response to the comments from Peter and Dave (below):
In RF power amplifiers "properly designed" for HF or VHF, the loading
control may have range to match a variety of load Z to the plate of the
tube. Designed so that the output C is centered in it's range with 50 + j0
Ohms on the output. So, it isn't so important to know if 2000 Ohms is
presented to the tube when 50 Ohms is on the output, as it is to know the Q
and to sweep around and verify that other resonances aren't present in a
way that the amplifier might support feedback and turn into an oscilattor
with parasites. If you don't hit the magic plate impedance at resonance,
then the efficiency may suffer, the amplifier may distort, etc.
I guess that's just my approach, and it has worked. The calculations of
course, must be made for the Pi. And the capacitors measured with a C meter
for verification, along with the value of L. I think that's what you guys
have said also.
I use the network analyzer way, but use it to indicate resonances, and to
indicate the transmission loss through the network. The Hewlett Packards
have a High Z GaAsFET probe which can be calibrated out (sort of) and hung
on the plate of the tube with little effect. It's not cheap, the probe is a
few thousand US dollars I think, in a fancy wooden box. And extremely
static sensitive.
Then drive the output connector with the source 50 Ohm port on the network
analyzer.
Yes, it is not as accurate to use the nework analyzer to display a high
impedance. Use the |Z| meter.
John
K5PRO
To: <amps@contesting.com>
>Date: Fri, 4 Dec 1998 10:09:04 -0000
>From: Peter Chadwick <Peter_Chadwick@mitel.com>
>Subject: [AMPS] Tank coils etc
>
>Doug, KF4KL says:
>
>>Another question. I have access to a vector impedance meter. I know
>>what the plate impedance will be for the tubes I've chosen. Can I
>>terminate the tank output with 50 ohms and use the vector Z meter to
>>select and verify the taps on the tank coil? It's an easy thing to do,
>>but in practice will it result in the correct tap selections?
>
>The Q of the tank will be set by the input capacitance of the Pi or Pi-L
>network. The correct capacitance has a reactance of RL/Q, where RL is the
>plate load, and Q is between 10 and 15. I'd suggest setting the capacitance
>at the ends of the Pi to the right value before setting coil taps.
>
>A point to watch is the variety of vector impedance meter. The network
>analyser approach uses directional couplers to measure amplitude and phase
>of the reflection in a 50 ohm system. If you measure a 2000 ohm load, quite
>large changes in the load have relatively small changes in return loss, so
>the accuracy goes to hell in a handbasket. Thus depending on the way in
>which the box works, it might be a waste of time. If you use a vector
>voltmeter, that's a different approach - and it needs a signal source.
To: <amps@contesting.com>
>Date: Sat, 05 Dec 1998 11:33:11 -0600
>From: Dave <dhaupt@bewellnet.com>
>Subject: [AMPS] Tank coils and vector impedance meters
a signal source.
>Peter's comments are correct regarding network analyzers used to measure
>impedance. A network analyzer measures the ratio between reflected and
>forward power, and therefore is optimized for measuring how close to 50
>ohms a given widget may be. Far from 50 ohms, and the accuracy
>suffers. To be within a useful amount, the general rule is that you
>should not attempt to measure an impedance who's VSWR with respect to 50
>ohms is more than 10:1. 2000 ohms is much more than 10:1.
>
>Fortunately, most instruments which are labelled "vector impedance
>meter" are not based on directional couplers and ratio measurements.
>The HP 4815A, available sometimes surplus, although I have yet to find
>one for a pleasant price, actually has both a voltage and current
>sensing transformer in the probe head. That is how most such impedance
>measuring instruments worked until the early 1980s when they switched to
>an auto-balancing bridge configuration. In either case, a true
>impedance meter can provide accurate results over a wide range of
>impedances.
>
>Peter also makes good points on the values of the network's capacitors.
>Merely demonstrating that you have achieved the correct load at the
>operating frequency is necessary, but insufficient. You also must
>demonstrate that you have the proper Q value in the network, and
>assessing the capacitor values is a reliable way to do so.
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|