The AM Forum

I have quite a few old Sangamo mica capacitor rated at 2500 volts. The are the flat 1/4" inch thick capacitors with two threaded hole or solder lugs on each end. Is twice the plate voltage enough. I think maybe not.
 
I thought that I had some suitable Russian door knobs until I found that NF means pF in Russian.  I have some nice USA made 6KV 1000pF disc capacitors,but not sure they are up to the current requirements, even though it is not great. I had some Sprague and Centralab door knobs but can't find them.

Any ideas on something I can get quick? The Russian stuff takes over a month.

Thanks

Pat
N4LTA

Check those Rusky capacitors with a capacitance meter to be sure.  You might be surprised.  The cyrillic alphabet can be hard to read sometimes.   I''d try a few of those 6kv disk ceramics.  Look at some of the old Johnson transmitters that had disc ceramics in parallel for the plate blockers that are still slaving away half a century later.
If you can find the 850 Centralabs in your stash they will be excellent overkill.  I ran those in a 4-1000A, a pair of them in parallel for decades without any problem.  But a 200 watt pep rig should do OK with the disc ceramics. 
In any case, don't forget to put a fat RF choke from the 50 ohm output connector to ground, just in case of the unlikely event of a breakdown in the blocking cap.  With a thin piece of wire in the HV lead to act as a fuse, it will keep HV off the antenna in case of a Chernoble event.
73 de Norm W1ITT

I checked them tonight and that is the reason for the post. The 10NF was 10pF. I'm glad I checked.

FB on the choke. I have always done that. I think I may have a source for a couple of Russian door knobs. It seems from looking at several typical schematics, the the plate blocking cap needs to be at least double the plate voltage. Something in the back of my mind tells me that may not be enough.

I assumed 1000pF would be about right for 80 meters and higher frequency, several experts on the web say 200-470 pF is plenty. Current carrying capacity is not particularly important because of the high impedance of the tank.


Pat
N4LTA

you can buy caps from Ameritron and probably other Amp makers.

I've seen the disc ceramics take WAY more than you'd think.

5 to 8 hundred pep.

The discs where beefy, but they took it for 20 plus years.

--Shane
KD6VXI

I thought plate blocking cap selection was a fairly simple thing until I read this thread http://amfone.net/Amforum/index.php?topic=27780.0 (http://amfone.net/Amforum/index.php?topic=27780.0) which dissects it into more pieces that I could have ever imagined. ::)

Here is a way to do it on your own by giving an example for a Linear Amp Only:

Expected RF Pout = 50 watts.

DC Pinput = 1.6 X Pout = 80.4 Watts

DC Ip = 80.4 Watts/1200 Vp = 67 mA; Check 0.067A X 1200V = 80.4 Watts

RF Pout = 50 Watts; Conversion efficiency is 50 W/80.4 W = 62.5%

Rplate = Real Part of Plate Impedance = 8250 ohms (using Terman's method)

P = V^2/Rplate; Vrms = sqrt(8250 ohms*50 Watts) = 642 Vrms volts of RF looking into the input of the Pi-L net.

Check Rplate: R = (V rms)^2/50 W = 8243 ohms. Ok, so we differ by 7 ohms, no big thang!



Peak RF voltage is = 1.4X642 Vrms or 899 volts of RF peak V.

Peak DC = 1200 V; Peak RF = 899 volts; Total voltage = 1200V + 899 V = 2100 volts.

So your cap needs to have a voltage rating of at least 2,500 volts.

(A cap's voltage rating is highly dependent upon the expected power output looking into the impedance of the Pi-net and the input impedance of the Pi-net should be = Rplate).

If you need around 40-50 ohms of coupling at the lowest frequency of operation (3.5 MHz).

Ccoupling cap = 1/(6.28 X 45 ohms X 3.5^6 Hz) = 1011 pf so closest cap value is 1000 pF or 0.001 uF@3kv for a bit of headroom.


Phil








 



 

 

Thank Phil

Knowing 40-50 ohms capacitive reactance is the practical stuff that is hard to find. Many different opinions out there.

My guess was 1000pF but the power out is 200 watts from the design schematic,.   Anyway - the plate current at 1400 volts in the schematic is specified ay 240 mA. That is 336 watts input. I am running it at 1200 volts so I am assuming something like 180 watts.

running your calcs for 200 watts:

320 watts in
267 mA Ip
R plate Terman - 4494 ohms

948 V rms  x 1.414 =1340 peak

V total 1200 +1340  = 2540 - need at least 3KV

40 ohms capacitive reactance

1033pF - 1000pf will probably do as I calculated for 40 ohms

I need a 1000pf 3kV doorknob


Fun - Like being back at NC State in 1970 in EE101 with a slide rule - HP 35s were only for the rich boys - $595 was three semesters tuition

Pat
N4LTA

I think Rplate is a bit high, but here is what I get:

For an expected Pout = 210 Watts

Vp = 1400,

Ip = 0.240

Pin = 336 Watts

Rplate = 2.8k

Eff = 0.625

Cptune = 131 pF  (3.7 Mhz)

L = 13.5 uH  (3.7 Mhz)

Cpload = 665 pF  (3.7 Mhz)

LPchoke = 1.1 mH

Vrms across 2.8k  = sqrt(2.8kX210W) =767 volts; V peak = 1074 volts.

Vtotal = 1400 Vdc + 1074V RFpeak = 2.5 kV.


Phil 

 

Me too on the plate voltage. That's why I dropped it to 1200.

These are beefy looking tubes - it will be interesting to see how they perform.

Thanks for all of your help Phil. I greatly appreciate it.

Pat
N4LTA

I advise staying away from those mica caps so common in the day, I had some trouble with using them as replacements in transmitters I repaired. Where they failed most was in repairing the switchable ceramic pillar mounted loading cap in the Valiant 2, they couldn't handle the current in only a 100W AM/CW transmitter. BTW Timtron and I agree, they're the worst POS E.F. Johnson made. Things went downhill fast after the Viking 2.

There are the sure fire caps I've seen a dozen times, once cheap and dirty 500pF 20KV TV HV doorknob caps found in so many old B&W TVs. I used to raid the junk pile behind the local TV repair shop (Remember the TV repairman?) and had boxes of them, only now the price has gone up by $40. https://www.surplussales.com/Capacitors/Trans_Coup_Caps/cap_trans.html They were used in many Amateur transmitters, they can handle the current which isn't *too* much in the high impedance plate output circuit. If you're worried about current you can halve it in each cap by putting 2 in parallel, .001uF isn't too much in most cases.

Ah for the good old days, what once cost nothing is now made of gold. <sigh>

Russia.

Just bought 10 150pf doorknobs (with leads).  Ratings in the multi kilovolt range.

Working fb as a padder on the output of a pair of 250B tubes.

Cost me like 12 bucks plus shipping from New Jersey.  Got them in a couple days.

Since you're using Euro tubes may increase the cool factor.  Maybe not.

--Shane
KD6VXI

Shane,

Where did you find them?

thanks

Pat

Ebay.  From ICA Manufacturing.


https://rover.ebay.com/rover/0/0/0?mpre=https%3A%2F%2Fwww.ebay.com%2Fulk%2Fitm%2F202485469874

Hope the url comes through.

12.99 plus shipping.

I also has a website, but doesn't have these caps listed.

However, they have a huge selection of cases for amps, TX, rx, whatever.  Aluminum, predrilled with pem nuts and a divider to seperate driven from plate areas, etc.

I purchase quite a bit from them and have used them for years.


https://www.icamanufacturing.com

--Shane
KD6VXI

I bought some Russian doorknobs - 1000pF 30KV - thinking they were RF caps. Nope, laser caps. They worked FB is a 100W and 300W linear amps.

Thanks

Fatboy has some good stuff. I ended up finding a 1000pF Sprague Vishay one at Mouser for more than I wanted to pay, but only $10 more than I could find anything used that was not a risk to be junk. I need 1000pF and they are not easy to find. I also picked up some Russian ones on Ebay but they won't be here for a month or so.

Pat
N4LTA

Are you able to use those 2 470pf doorknobs I sent?

John

They would work in parallel but the size of two of them is a problem. I thought 470pF would be enough according to some published data, but it turns out I need closer to 1000pF.

I will return them if you need them.  Otherwise I may use them on a future larger build that I have in mind.

Pat
N4LTA

Hold on to them. I don’t need them.