Amps
[Top] [All Lists]

Re: [Amps] Pi-L values

To: <zl1aih@ihug.co.nz>, <amps@contesting.com>
Subject: Re: [Amps] Pi-L values
From: "Tom W8JI" <w8ji@w8ji.com>
Date: Sat, 2 Sep 2006 08:45:22 -0400
List-post: <mailto:amps@contesting.com>
> In the Pi-L network, C1 and C2 (vacuum variables) are 
> coupled by a
> toothed belt - motor driven by a servo amp.   The manual 
> describes
> C1 and C2 maintaining a capacitance ratio of 2.5:1, L1 has 
> taps
> selected by the bandswitch while L2 is 17uH variable to 
> some
> minimum inductance - the continuously variable part of L2 
> is a
> 'pancake' inductor, with minimal lead length.   L2 is 
> motor driven by
> another servo amp.

How do you know Sunair intends it to behave like a Pi-L and 
not a double L? By the capacitance ratio?

>>From Bill Orr's handbook (23rd Edition), the Pi-L table C1 
>>and C2
> values show a ratio of between 4.5 and 5:1 for an anode 
> impedance
> of 2000 ohms - close enough for 2500v @ 0.65A.

All that is is just an approximation for some value of tank 
Q and transition impedance between the Pi and L that Orr or 
someone picked. There isn't anything cast in stone that says 
those values have to be used.

> Test conditions :-
> Amplifier running - no HV.   1900 ohm resistor from anode 
> to ground.
> MFJ analyser connected to output side of L2.
> C1/C2 and L2 adjusted manually band by band for minimum 
> SWR
> reading - all bands less than 1.2:1 except for 10 metres. 
> The highest
> frequency that will yield less than 1.5:1 is 25.5 mHz.

My guess, and it's only a guess, is the stray C from the 
anode is preventing the network from tracking above 25 MHz.

> Question - Did anyone on the the list ever use the 
> GSL-1900A amp on
> 10 metres?

Not me.

> Also, I may have missed something along the way, but why 
> do the
> amp manufacturers use 2.5:1 ratio when Bill Orr's 
> (Eimac's) tables
> show an average of 4.75:1?

Because any value can be used so long as a minimum Q 
requirement is met. There is also nothing cast in stone that 
says the network must behave like a pi-L, or the first 
section must have a phase shift more than ~100 degrees. It 
could be in double L mode. Many commercial amps do that to 
obtain more bandwidth before retuning is necessary.

Many of us tend to think in terms of "my amp won't work 
unless I follow this published tank formula". In reality the 
"formulas" we follow exactly are just approximations someone 
decided to use. They aren't rules.

> Question - Observing the serious decrease in power output 
> with
> increase in frequency, should I be looking at the C1/C2 
> ratio as a
> likely problem?

I'd look at stray C on the plate tuning cap as being a 
contributor. One would think they compensated with some 
fixed C at another appropriate spot in the tank, like across 
the "loading" capacitor.

> Disregarding the difference in recommended ratios, it's 
> possible that
> some previous owner has "played around" with the toothed 
> belt.
> The manual calls for a Vector Impedance Meter to set the 
> capacitors -
> I don't have access to one.    Maybe there's a 
> work-around?

The method you are using with the 259 should work, although 
it may take a bit more fooling around. I assume you are 
connecting it so stray impedances are minimized.

73 Tom 


_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps

<Prev in Thread] Current Thread [Next in Thread>