Great information, very helpful! This is so much better than trying to figure
this out
while soldering taps on the coil, cutting tubing, etc.
I am also wondering whether I will need parasitic suppressors at all. Many 813
amplifiers
I've seen (GE SB handbook, ARRL HB's, etc.) do not use them. I plan to mount
the sockets
slightly below the chassis and bring bypass leads to the top of the chassis
with very
short leads to the same ground point used for the tuning capacitor. Also I am
going to use
parallel 0.01 and 0.001 uf for the bypasses. This should improve stability.
On 4/7/2010 5:30 AM, Jim Thomson wrote:
> Date: Mon, 05 Apr 2010 23:03:33 -0700
> From: Vic K2VCO<vic@rakefet.com>
>
> pi-net spreadsheet question
>
> I'm using the GM3SEK pi-net design
> spreadsheet<http://www.ifwtech.co.uk/g3sek/>. It
> requires that you enter the inductance and resistance used in the parasitic
> suppressor.
> What I'm wondering is how to do this for two tubes in parallel, where there
> are two
> suppressors.
>
> Do you enter the value of one coil and resistor, or do you treat it as though
> these
> components are in parallel?
>
> ## excellent question.. and I have been using his pi sheet for yrs now...
> but only
> 1 x tube amps... or 2 x GS35bs... where no suppressor is used at all.
>
> ## I'd say to parallel the values of the 2 x suppressors. If it's a GG amp,
> the grids
> are in parallel at one end. The plate block cap is where the 2 x
> suppressors come
> together at the other end.
>
>
>
>
> This makes a big difference at 28 and 21 MHz.
>
> I am finding this exercise very interesting, playing with the value of "lead
> inductance".
> I'm trying to add a small "l network" before the main pi network which will
> cancel out
> some of the large amount of output capacity of a pair of 813's so they will
> be reasonably
> efficient on 10 and 15 meters. The parasitic suppressors become part of this,
> too.
>
> ## that feature is the BEST part of the spread sheet ! Start with just
> the paralleled
> suppressors, NO lead uh. The tune cap may end up neg value.. or very low..
> like
> just a few pf. [say for a Q=14-18] Take what ever value it spits out for
> the main coil..
> and then divide by 2. EG: spits out say .. '.8 uh' .8/2 = .4uh
> Ok, then make the
> lead uh .4 uh. Main coil value may well change. Juggle the lead Uh
> value, till BOTH the
> lead uh and main coil are identical.
>
> ## Ok, now you can probably DEcrease the loaded Q. This all depends on Min
> C u have
> and whether a vac cap is used, etc.
> The idea here is to obtain the lowest Q you can, and still not have the
> tune cap totally
> unmeshed.. and keep the main coil and 'lead uh' the same value.
>
> ## I call the result a ' L-PI' and it works very well ! On my
> 160-15m [ no 10+12m]
> 3CX-3000A7 amp, sure, it would tune 15m.. with out the extra .66 uh b4
> the main
> pi-net.. but loaded Q was WAY too high for my liking. Also, my 15 M
> position is also
> used on 17m. So loaded Q on 17M is through the roof. The 3x3 tube has
> 24 pf of
> anode to grid C... and rises to 33 pf.. when tube in the socket. Min C of
> the vac cap is
> 12 pf... so I have 45 pf.. with the vac cap unmeshed !
>
> ## with .66uh b4 the pi -net..... plate load Z drops way down. Now I can
> get the loaded Q
> on 15m down to just 8....... which rises to Q=12 when on 17M.
>
> ## In this case.. a .66 uh coil was made from 7 x turns of 3/8" od
> tubing... on a 1.5" ID.
> The .66 uh coil is at right angles to main tank coil. I THINK the
> harmonic suppression increases
> having the 2 x coils at right angles. Main tank coil was 3" OD.. also
> made of 3/8" OD tubing.
>
> ## so the extra coil doesn't have to be the same diam as main coil. In my
> case, I wanted it
> long and skinny, to bridge the gap between Plate block cap assy.. and vac
> tune cap.
>
> ## on another amp, the vac tune cap was tapped one turn into the main
> coil. In this case, it's
> all one coil. So both coil orientation schemes have been tried, and both
> work very well.
>
> ## On older desk top amps, where loaded Q is sky high on 10M... adding a
> tiny bit of uh between
> plate block cap and C1 cap does wonders. The extra uh can also go anywhere
> between the anode
> and C1 cap. Then, using the stock value 10m tank coil... the loaded Q
> drops like a rock, way less
> circulating current, more eff, and main tank coil and the poor band switch
> don't get cooked! But the
> extra L coil has to be same OD as main tank coil.. or bigger, if main 10m
> tank coil is marginal to start
> with.
>
> ## I designed a 6m 8877 tank coil setup for a fellow 2 yrs ago. It used
> a .5 uh coil... wound with
> 1/2" wide flat strap... and tapped dead center in the middle of the coil..
> with the tune cap. The strap
> coil is wound like a suppressor.. with the advantage that there is zip C
> between turns.. since they
> look like knife edges facing each other. 3/8" wide strap is the eq of
> 1/4" tubing, since the
> circumference of both is aprx the same. 1/2" wide strap = .318" OD tubing,
> etc. He ends up
> with .25uh on either side of the tune cap. This works very well. Two
> separate .25 uh coils
> can also be wound... and at right angles, or anything from in line to 90 degs
> between em.
>
> ## That extra L coil gets installed on all my amps, since I discovered
> the spread sheet, 10-12
> yrs ago. b4 that.. the maths by hand was tedious.
>
> later....... Jim VE7RF
>
--
Vic, K2VCO
Fresno CA
http://www.qsl.net/k2vco/
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|