¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io

Re: High Temperature in Toroid of the PI-L network (Alpha 78)


 

¿ªÔÆÌåÓý

Jeez, I love your sob stories, Jim !!! What a read. You should write a small book ¨C a little like the ¡°Care and Feeding¡± thing from Eimac. All of us amp geeks would love it.

?

From: [email protected] <[email protected]> On Behalf Of Jim VE7RF
Sent: Thursday, March 7, 2024 12:53 AM
To: [email protected]
Subject: Re: [ham-amplifiers] High Temperature in Toroid of the PI-L network (Alpha 78)

?

On a side note, on these hb amps,? I design the PI net in software...but I also use a LOT lower loaded Q.? (I gave up on the PI-L...major pita.? The PI-L config uses a LOT more L1 coil between the caps...and a LOT smaller C1 + C2 caps.? That results in L1 coil values that are absurdly too big on lower bands.? It also results in impractical extremely low values of C1 on upper HF.? ?It also means the extra L2 coil, plus tapping the L2 coil...pita.? ?It also means much higher voltage across the C2 load cap, since it's then at a 200-400 ohm point.? It also means when using a lower loaded tank Q, the C1 + C2 values decrease even more...and the L1 coil increases even more).??

I have done a lot of experimenting over the years....(PI net) esp on hb amps that used a roller coil.? Like trying both the calculated L1 value for a 'normal' tank Q of 12.? (input Q of 10..and an output Q of 2)..... then progressively using MORE L1...( and progressively using LESS? C1 + C2) and the results always astound me.? With the lower Q? (amp re-tuned each time for max PO), the PO kept increasing.? Anode diss kept decreasing.? The amp is far more broad banded ( doesn't require constant re-tuning, esp on the lower bands), and way less heat in the tank coil.....and? way less current though both caps...and the tank coil (roller or bandswitched)..and roller contacts ( weak point in any roller)...and less current in bandswitch contacts.?

The tradeoff being... slightly less harmonic reduction.? IMO, that's a good trade off.? As noted before, monoband ants are not resonant on their even harmonics..... and even with a? lower tank? Q, harmonic suppression is still, greater? on the 3rd-4th-5th harmonics...vs 2nd harmonic.? And on loaded 40m yagis, mine does not resonate on 21 mhz ( it resonates on 24 mhz...and will screw up a nearby 12m ant).??

?

I recently experimented with the adjustable hb manually tuned input, and VARYING the tuned input Q.? ? ? Without any tuned input (or the un-un gong show they used in the henry 3/8 k ultra) harmonic? suppression is degraded.? ?Eff is reduced, and IMD increases a bunch...(and that's with a tube that already has a 50 ohm input Z).? ? ? With the PI? tuned input (and the typ low Q), amplifier harmonic suppression increases a bunch ( as Clark noted when he installed the new relay switched PI tuned inputs on his 8K).? ?Typ a low Q is used, so the tuned input is tweaked for the middle of each band...then left alone.?

With the new 0 - 4.4 uh roller coil (8 ga solid wire) and the pair of broadcast caps, I can vary the tuned input Q between extremes...on any band.? What I found was, it's a trade off.? With a? slightly higher tuned input Q, (more C1+C2..and less L)? everything improves, BUT on the lower bands, it's not broad enough to cover the entire band.... but almost.? ?Being manually tuned, it's a non issue.? ?IE: run into the shop to tweak the vac? tune+ load caps? from 3900 khz...down to 3504 khz ( using the handy dandy cheat sheets), it only takes a few more seconds to tweak the C1 + C2 broadcast caps on the manually tuned, tuned input. ( same cheat sheet. Cheat sheet divided up for each band, with 2 x columns... Vac C1+C2 caps...and tuned input? broadcast C1 +C2 caps.? ?Then everything done at 3 x freqs per band.? Bottom-middle-top.?

?

On the bandswitched, manually tuned, tuned inputs... they are configured for a slightly? higher tuned input Q.? Same deal, just tweak the C1/C2 broadcast caps? via cheat sheet.? ? Get the tuned input Q too high, and the resulting? current through the tapped? coil? /? roller contacts? /? bandswitch contacts becomes too great.....esp on upper HF....and esp when? more drive is used...like 500-700 watts on the bigger tubes.?

?

On these 50 ohm input 3CX-3000A7's....and also the 3CX-6000A7's,? I designed the tuned inputs with an overall network Q of ...'4'.? ?( input Q =2.? Output Q = 2).? This results in the best compromise.? ?The C1 + C2 values are tabulated below.?

160m -? 3600 pf
80m - 1800 pf
40m - 900 pf
20m - 450 pf
15m - 300 pf
10m - 225 pf.?

?

As you can see from above, the C1 + C2 values are a lot higher vs what you normally find in typ amplifiers.....with their fixed fixed values, tweaked for the middle of each band.?

?

I also put watt meters on both sides of the manually tuned, tuned inputs, and measured PI vs PO. (that came from John, K5PRO).? ?On my 3x3 amp, with it's manually tuned, tuned input, complete with bandswitched? 4 uh coil ( made from 17 turns of? 6 ga solid cu, with a 1.5" ID).? ?Pin = PO.... except on 15m.? On 15m, the Pin was 200 watts, but the PO was only 160 watts !? ?I had the Q too high on 15m..... and that 40 watts lost was dumped into the remaining 2 turns used on the 4 uh coil.? I could easily feel the heat on those last 2 x turns.? ?40 watts is a lotta watts for 2 turns of 6 ga wire.? ?My fix was to increase the inductance on 15m by re-tapping the coil? by just 1/4 turn. ..which then required slightly less C1 + C2.? ?PO increased to 195 watts on 21.225 mhz..... good enough.? ? It's not always easy to do exact coil taps.? IE: the? 1/4" wide interconnecting copper strap may have to go over the top of the coil (or beneath it) to get to the opposite side....( 9 o'clock)? ?A longer strap also has more XL...and also has to be factored into the equation, along with the total path through the bandswitch used. IE: from C1 to coil, then through the various taps, back to the bandswitch, then to C2 cap.? ? IE: bandswitch contacts face the coil at the .....3 o'clock? position. On upper HF, each turn is a lot of XL, so fractional turns are used, like top of coil...and bottom of coil? ?(12 o'clock and 6 o'clock).

So I use a lower Q on the multi? KW? output PI network, but a? slightly high Q on the tuned input.? Then u get ur cake ...and eat it too.? Good imd, flat input swr across each band, excellent harmonic reduction,? easily handles up to 1 kw CCS CXR of drive.?

I wasn't about to attempt to construct 6-9 fixed? fixed tuned inputs, and either bandswitched or relay switched, that would handle 500-800 watts? CCS of drive.? ?Ok, that's my sob story for this week.?

Join [email protected] to automatically receive all group messages.