The downside of the "pre-inductor" is that it tends to destabilize the
amplifier at VHF. So if you use one, you should be sure to check for VHF
parasitics.
My experience is that it's hard to predict how it will work because the
stray inductance and capacitance in the circuit are difficult to predict.
I've made these coils out of a turn or two of no. 10 wire. You get a wide
range of adjustment by squeezing and stretching the coil.
Victor 4X6GP
On Wed, Apr 16, 2025, 22:55 Steve Bookout <steve@nr4m.com> wrote:
> Hi Jim,
>
> Thanks for your useful input. It's got me thinking about a thing, or two.
>
> I think I'm making some incorrect assumptions with the spreadsheet, as
> I'm not getting quite the same #'s as you are. I don't want to make
> some stupid error, that'll bite me in the butt later.
>
> I'm seeing that the larger you make the 'pre-inductor', the larger the
> value of C1 becomes, and I do understand this is the whole reason for
> doing the extra inductor. My question is, what is the down side of 'way
> large' in order to get the C1 value much higher? Q or tuned bandwidth,
> etc?
>
> I had been using a .6 uh inductor. When I changed it to 1.0 uh, the C1
> value was significantly higher, which is sooooo useful. My actual mock
> up that I had done a couple of years ago, that seemed like it wanted to
> tune, was 4 turns, wound something like 1 3/4 dia, with 1/4 inch
> tubing. To make it work better for the mechanical lash up using 2 x 200
> pf doorknobs and the plate choke, it seems like I should make fit my
> physical needs, as a priority. Whether it's .6, .7, .9 or 1.0, I think
> an 'even turn' would work out better for me. I always find something
> like 4 3/4 turns a pain to hook up and get it looking and working
> correctly. Just an OCD thing; I did mechanical prototyping for decades
> and the OCD thing surfaced along the way...
>
> Waiting for the copper tubing to arrive...
>
> Steve NR4M
>
> On 4/16/2025 10:05 AM, jim.thom jim.thom@telus.net wrote:
> > Date: Mon, 14 Apr 2025 13:43:16 -0400
> > From: Steve Bookout <steve@nr4m.com>
> > To: amps@contesting.com
> > Subject: Re: [Amps] 10 meter amp coil
> >
> >
> >
> > **** I've used GM4SEK's on-line tank circuit calculator in the past and
> > have been working with it again for this amp.
> >
> > **** I don't have it all in front of me, as it's all on a computer in
> > the shack next door, but, as I remember, the #'s were .6 or .65 amps @
> > 4KV and a plate load impedance of something like 3600 ohms.
> >
> > **** I entered something like .6 uh for the 'stray inductance between
> > tube and C1' and played with the numbers a bit.? As I recall, with .6 uh
> > for the first 'L' coil,? I was able to get something like 18-20 pf for
> > C1.? That is doable!? The other inductor and C2 aren't an issue.
> >
> > **** The vacuum cap I'm using for C1 is something like 10 to 100 pf @ 15
> > kv; a Jennings cap.? Because I need so little cap on C1, I plan on
> > adding 3 - 50 pf NPO 7.5 KV 'small' ceramic doorknob caps.? That should
> > make the tuning less 'touchy'.? Instead of a tuning range of 10 ish to
> > 100 ish pf, it will give me ~9 to 60 pf.? All my antennas are resonant,
> > low SWR antennas, so the tuning C1 cap may never be touched again, once
> > the proper setting is found.? If I can get it tuned with a reasonable Q,
> > I'm not worried about frequency excursions.? On my 160 meter amp, the C1
> > vac variable doesn't have a turns counter, or any sort of scale, just
> > some really old masking tape with some felt tip tick marks.? Only been
> > like that for 40 years.
> >
> >> The pre-coil and main coil can be one coil, with the C1 cap connected
> to
> >> the junction point. OR you can use 2 x separate coils. Each of the 2
> x
> >> coils can be made from tubing or strap, in any combo u want. IE:
> >> coil-coil. Strap-coil, coil-strap, strap-strap.
> > ****? Trying to do separate coils
> >> I design those networks in software all the time. You need a LCR meter
> to
> >> tweak the coils / taps though. Lemme know if u need help on this. I
> need
> >> to know aprx loaded B+ and also loaded plate current.
> > **** I have an 'almost all digital' meter (AAD) which works well.? I
> > also have an AIM 4170, which I have used for years and have learned to
> > rely on, but may ask for your wisdom.
> >> This is no place for 1/4" tubing, too small a diam, and coil will run
> > hot
> >> on 10m.
> >> The way I deal with heat is...... don't generate it in the 1st
> place...or
> >> at least minimize it.
> > **** Amen!
> >> Since it's only used for CW, you can reduce the idle current by a huge
> >> amount, just more bias with a simple string of diodes. You can't
> reduce
> >> the idle down to zero, or you will get key clicks. Increase bias V,
> till
> >> idle current is reduced to aprx 20-40 ma. No clicks, tank eff is
> > improved,
> >> and minimal idle power. I can work that out for 4 kv if u like.
> > **** That would be great.? My thoughts were to get it to about -20 volts
> > bias.? I know that -22 volts is into class C, but don't know how far
> > into it, it is.? I have a good stash of 1n4007 which would be fine, but
> > also have some others as well.
> >
> > **** Do have a separate question for you.? I'm just ignorant on the
> > application specifics.? DC blocking cap between the plate choke and the
> > tank circuit.? When is capacitive reactance TOO small?? I have several
> > 200 pf, 15 KV doorknobs.? If I put two in parallel for current,? that
> > gives me 400 pf for a cap reactance of about 14 ohms.? Is that good
> > enough at 28 Mhz??? 'X' amount of RF current thru 14 ohms gives "ZZ'
> > watts of heat.? Don't want heat. Then, again, I have an 'old school'
> > antique, boat anchor ceramic/mica HUGE transmitting cap (like 2 x 3 x 3
> > inch high) .01 mf @ 5000 volts.? That gives me less than an ohm of cap
> > reactance at 28 MHz and eliminates the need for a ceramic or fiberglass
> > standoff to support the 2 x 200 pf caps..? This big cap is its' own
> > standoff.? It's labeled for several amp of RF current at some low freq.?
> > My 160 amp has used an identical one for a blocking cap for 4 decades,
> > but there is a world of difference between 160 and 10.? Not being
> > familiar with any specs, or history associated with this type of
> > ceramic/mica cap, it may work fine,? or maybe the worst thing I could
> > do.? May have been from some commercial AM transmitter from 80 years ago.
> >
> > ?Comments?
> >
> > Steve NR4M
> > ## use .8 to 1.0 uh for the pre coil. Then it's 20/30 pf for the C1
> > tune cap. Then u don't have to add the 7.5 kv doorknob caps. It will
> tune
> > just fine.
> >
> > ## Make sure the pre coil and main coil are BOTH wound CW...... or BOTH
> > wound CCW
> >
> > ## Use something bigger, like 1N5408 or 6A10 / PM-600 for bias diodes.
> > Per constant current charts, with 4 kv, you need 16 vdc of bias, for 50
> ma
> > of idle current.
> > That will require 22 x diodes. Leave room for more if needed, like
> another
> > 3 x diodes.
> >
> > ## Use 2 x 200 pf @ 15 kv HT-57 doorknobs for plate block cap assy.
> > Those HEC/ centralab 857 caps are rated for 15 amps CCS on any freq >
> 5.11
> > mhz. So a pair is good for 30 amps, which is ample. Those are 35 kva
> > rated caps. Current through the block cap assy is aprx 4.2 amps. 4.2
> amps
> > X 14 ohms xc = 59 vrms = 84 V peak. That 84 V peak AC Vdrop across
> the
> > cap plus the 4 kv B+ ( what HEC calls DC bias is what we call B+) =
> 4084
> > V peak. HEC sez the peak V across the cap plus any 'dc bias' can not
> > exceed the stamped HV rating of the cap.
> >
> > ## The tubes anode to grid C is 10 pf = 566 ohms of xc @ 28 mhz. You
> take
> > the 4 kv B+ and multiply it by .6 to obtain the RF RMS voltage ( =
> 2400
> > vrms) . 2400 vrms / 455 ohms = 4.2 amps. Think of the tubes anode to
> > grid C as being a major portion of the total C1 required tune C. Tube C
> > and C1 are in parallel. Problem is u have the plate block cap between
> em.
> > That 4.2 amps also has to flow through the block cap assy.
> >
> > This is why block caps have to handle more current on 10m...vs 160m. XC
> > from the tube C is sky high ( and 16 times greater ) on 160m. RF current
> > is minimal, and minimal through the block assy. However, typ we want a
> > low enough XC on the block assy, such that the xc of the block assy is no
> > more than 5% of the plate load Z.
> >
> > ## The block cap has to be big enough in value to meet the 5% of plate
> load
> > Z rule.
> > At the high freq end, we have to worry about current. In your case,
> > it's a monoband 10m amp, so cap value is a non issue, esp with it's 14
> > ohms of xc vs the 3400-3600 ohm plate load Z of the tube.
> >
> > ## Don't worry about the Xc of the plate block cap assy ever being too
> low.
> > ## any current through the plate block assy is watt less VA power for
> the
> > most part, but u can't exceed the RF current rating of the cap.
> >
> > ## strap the pair of doorknob caps such that the current is divided
> 50-50
> > between them. This implies that the distance in = distance out. IE:
> 2
> > x caps in parallel, with their main axis N-S. One strap comes from
> the
> > W side ( tube side) .....and cover's the tops of both caps. Output
> strap
> > comes from the E side..and covers the bottom of both caps.
> >
> > ## Don't use that giant mica / broadcast cap. They handle max rated
> > current at low freqs like 1 mhz....... and less on higher freqs like 10m.
> > Ceramic caps are the opposite.
> > It would probably work, but is not optimum. The HT-57 cap in the 200 pf
> > value is the 'go-to' cap in the HT-57 series, since it handles the most
> > current.
> >
> > Jim VE7RF
> > _______________________________________________
> > Amps mailing list
> > Amps@contesting.com
> > http://lists.contesting.com/mailman/listinfo/amps
>
> _______________________________________________
> Amps mailing list
> Amps@contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
>
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|