What tube would you consider for this amp project?

[ke7trp]

I have a rack cabinet that contains, three power supplys, variacs and metering.  I want to build an RF deck for it to use as an SSB and AM linear.

The Screen supply has its own variac, meters and output connector.  0 to 800 volts as tested.
The HV supply uses a 3650-0-3650 transformer rated at about an amp.  It has two big chokes and two big oil caps. There is plate current and plate voltage meters. Tested to 4K volts with variac.
The last supply is a bias supply. I dont know its rating but its adjustable.

This thing is ready for an RF deck.  What tube or tubes would you consider for max output power with these supplys?

It seems that the 4-1000, 3-1000, 1500 ect will not be a good choice for me as the voltage is only 3 to 4K on the HV.

I want to run 20, 40, 80 and 160 with this.  Plan to use a Transiever on a shelf so the entire thing rolls around. Probably a 1000D or FT 101.

I thought of three or four, 3-500s as I have plenty of tubes and sockets. However, I am worried this will produce an incredible amount of heat.  I thought of a couple 4-400s but then I would be dealing with very low Drive requirements.

I have never used the Russian tubes so maybe there is one out there that is perfect for this?

What is a simple, cheap and effective RF deck for this thing?

C

[The Slab Bacon]

A pair of 4-1000s would be happy as a clam on 4Kv!! Not to mention look pretty kool through a peek-a-boo window!!

Lose the screen supply and biass supplies, run them grounded grid. keeps everything nice and simple and leaves room in the rack for otha stuff!

Just my $.02 worth. I love those tubes and used broadcash pulls are still cheap!

[KM1H]

A 4x1 with 3500V will do 1500W out in GG and 100W drive. A 3 holer of 4-400's will do a bit more and give you 200W more of Pd headroom for AM. Either tube can be run GG or swamped grid driven. With the 4-400 you can get away with the ceramic socket with a small fan blowing on the pins and a bigger one on top blowing on the glass. Saves the hassle of pressurizing and chimneys.

Carl

[WA1GFZ]

What they said I agree. My old 4-1000A rig did 1500 watts using the pulls that are Frank's spares, (sept for the brown one) Multiple tubes in parallel makes life interesting because layout is much more critical

[ke7trp]

Hmm.. I am showing 6K to 7K is needed on the 4-1000 for full output.   3 to 4k is fine??? 

C

[WA1GFZ]

you need to grow some BA hair before you start messing with 6KV on a 4-1000A sonny. And get a good blower. Above 5KV is a whole new ball game.

[ke7trp]

Yeah.. I got plenty of help.. = You guys.

Nobody likes the multiple 3-500 idea?  I happen to have a box O tubes and sockets are cheap. 

C

[W2PFY]

Above 5KV is a whole new ball game

I was amassed on how fast you can blow holes right through 833'A at 3.5 KV. I was going to run them at 4 KV. The arc over's were also frighting at 3.5 KV so i cannot imagine what life must be like above 5 KV.

In my Westinghouse MW-2 I had some arc over's to ground. The rectifier uses three phase and I have a 16 KVA plate transformer on there. It would sound as loud as a 12 gauge shotgun going off which gave me pause to think about what it would be like to get your body across it.

I guess the message here is to BE CAREFUL.

[W2VW]

Pair 4-1000 cathode driven with G1 and G2 potentials both used.

[K1JJ]

Hmm.. I am showing 6K to 7K is needed on the 4-1000 for full output.   3 to 4k is fine???  

C

According to Chuck/ K1KW, an ex-Varian engineer, keep the HV below 5500V maximum on a 4-1000A in linear. I can't recall exactly what it's called, but something to do with electrons re-radiating off the plate or something similar causing non-linearity. Besides, 5500V max (or less) works FB in my experience and need not be increased.

It's just a matter of getting a good tank impedance match for good efficency at whatever voltage you run.

Cooling will be very important at those voltage/power levels, so use a chimney....  

T

[WA1GFZ]

I ran 5500 one time on 10 meters. The plate color approached the color of the sun. Made some real juice though. 

[ke7trp]

I used to have a 4-1000 amp build by someone in Cali.  It was very well built with Vac caps and big Dahl supply.  This amp ran 6600 volts on the plate. It ran hot but sure made the power.

C

[Opcom]

a 3CX3000?

but if you got the 3-500Z's, use them but put them in chimneys and close together. The blower comment is on target no matter what you do, if you intend to push that plate supply at all.

BA hair? Is that the sticky gray dust bunny things in the bottom of old transmitters?

[WA1GFZ]

nothing wrong with 3-500s and 3500 volts should make them sing. Just need to adjust the bias so resting current isn't too high.

[ke7trp]

I wanted the 3cx3000 tube orginaly.  But was told that it really needs more then 3K volts and alot more current.  I will look into this more. Thanks for the ideas. Keep em comming.. 


So far, I like the 4-1000 idea at G1 G2 ground.

C

[w4bfs]

FWIW  the 1978 ARRL handbook showed a 1kW input linear amp with 50 Ohm terminated grid driven ... can use triodes ,tetrodes or pentodes .... not having to bandswitch an input circuit is attractive .... the article pictured a 833 in it ...hmmmm

[The Slab Bacon]

So far, I like the 4-1000 idea at G1 G2 ground.

Clark,
        If you can find the odd ball filament transformer (7.5V@21A), and the socket. Grounded grid 4x1s are a very friendly tube. Friendly and non finicky.

I've been running the same one in my big TX for almost 10 years now!

[K1JJ]

http://amfone.net/Amforum/index.php?topic=18634.0


Clark:


The 4X1 tetrode is an anomaly when placed in hi-mu GG triode service. It is cleaner than most other tetrodes in GG.  That's why the triode-connected 4X1 modulators work so well. Same for 813's.


The above thread contains a rare letter from the Eimac boys spelling out the suitability of a 4X1 in GG. Great info when running the G1/G2 at ground - info not found anywhere else.

The recommended grid current is revealing as well as the suggested output power at various plate voltages. 3KV is the minimum plate voltage due to grid current exceeding tube parameters.


In general, I've found a GG 4X1 in linear service too be reasonably clean compared to other specialized triodes designed for linear service.

T

[ke7trp]

Interesting. This makes the construction so much easier.  4K on the plates. This thing should do 3K out easy with into the dummy??

C

[K1JJ]

Interesting. This makes the construction so much easier.  4K on the plates. This thing should do 3K out easy with into the dummy??

C

Well, for a single tube if you run the grid currents within Eimac specs to be clean, it says 4KV gives 1870w pep output, so you wud need a pair as you intended..

When I had Quadzilla going (R.I.P.)  I found about 2KW per tube into a dummy load was the best I cud do at 4KV. These were pulls and I drove the tubes to slightly under saturation.  Driving them harder would make the IMD specs unacceptable, in my opinion.

A pair is not a bad idea in the end...


And as Jay/W1VD says - reduce the full power output by 3db (1/2 power) and the 3rd IMD spec will clean up roughly ~9db better. That's quite significant.

T

[ke7trp]

Ok sounds good.  Now I need to stumble on two good tubes and sockets and start looking at the network.

C

[WD5JKO]

*3650-0-3650 transformer rated at about an amp*

That will make about 3250 volts with a choke input filter. My vote is an 8877 tube, and use one of those 'medical pulls' available for a reasonable amount. You can adapt an old 4D32 tube socket, and make your own chimney.

I also like the trio of 4-400's...

Jim
WD5JKO

[ke7trp]

I think it an amp.. Its an electro engineering job. Same one used in Gates or 707 for plate transformer.

I have two 1500 tubes.  I can go that route.. One or two of them.

C

[KM1H]

I have to laugh at that Eimac sheet!  If you underload it at 5KV it will take off like a rocket. Load it untill it wont put out more and then continue loading until the power drops about 50-100W. That is the point of best IMD.  If a 4X1 can handle 2500W out in AM Class C then about 2200W in SSB wont bother it since the PD ratings are CCS and quite conservative.  Same with a 4-400 or 4-500.

OTOH a 3-500Z sucks for several reasons. Its a SSB tube and rated as such. At 3500V key down it lights up brilliantly. Use it as an AM linear at that voltage and you are asking for fireworks. Older Eimacs often wont pass a 14KV Hi-Pot and neither will most all Chinkadink versions. Some years back I tried several tubes in an AL-82 as an AM linear before I built a Hi-Pot tester. They all worked fine in my LK-500ZC at 2500V. In the AL-82 (tested 1 tube at a time on 80M) they gave light shows, fried parasitic suppressors, blown fuses which resulted ime losing 2 tubes. Its the last time I tried that dumbass stunt.

The 3CX3000A7 is a fantastic tube and almost impossible to kill, its a favorite with technically challenged CBers and still used in many BC stations. The Chinese version is highly acclaimed.

The 8877 is a great tube as long as you tread lightly and have full protection circuitry built in otherwise it very easily can be a display piece. Its not for AM at any serious voltage unless you have a 3CPX1500A7 version.

My tube of choice would be the YC-156. A 3CX15000B7 internals and a 3CX5000 anode. It will easly blow past 1500W with 100W drive and really wakes up at 300W. Anything from 3500-5500V works just fine. Modest filament requirements but it really wants a 6-8 minute warmup so you turn it on once and then off at bed time. No socket, it bolts right to the chassis; stable as a rock, no parasitic suppressor even needed. Its a medical tube and IMD in the -50dB area is attainable; pulls are readily available around $300 and will be the last amp and tube you will need. You will be Loud and Clean.
http://www.g8wrb.org/yc156/

Carl

[Opcom]

I have to laugh at that Eimac sheet!  If you underload it at 5KV it will take off like a rocket.

Carl

So at higher voltages, the gain in GG becomes such that it will go nuts, or, the stress caused by underloading with drive applied will blow it? Interesting that do not mention that but I suppose they expect designers to know it.