The AM Forum

I posted about this   amp in "filament choke construction" earlier.  There is a picture of the underside of the chassis there.  I have the amp working now and I am pleased with the results so far.  C1 is an Eimac 10-60 pf vacuum cap 20KV.  L1 is tapped at 1.1 uhy  C2 is a Comet 5-500 pf 5KV vacuum cap.  The plate by pass and blocker caps are russian 15 KV.  I got some good advice from gfz Frank on using a small inductance between C1 and the blocking cap.  The tuning of the tank smoothed out and got broader when that small inductance was added.  Frank also told me about swamping the fil choke with 47K resistors to kill the Q of any VHF gremlins in there.  The testing so far has produced the following results.  I'm at 3300 V and .250 ma driving with a FT102 at 50 watts out.  That's 825W input and the Bird says 600W out  That's 73%.  I did not expect this high efficiency.  I'm just tickling this thing right now.  My power supply is capable of much more,  so I will be working the plate voltage up to 5 KV.+
   I built this amp in an open rack with sorta modular construction and it makes it a lot easier to build and work on..  When you make something like this, you know you will have to have it on and off the bench countless times, hook and un hook the wires, so I  tried to make that process easier.  I will report again as the voltage and current go up.  I am looking for a couple Eimac  glass chimneys .  I have a line on a couple metal chimneys for 4x1.  They have a small peep hole so you can see the plate but how much fun is that?

pics

4X1's: the Venus de Milo of RF glasswork. What a shame it would be to hide those beauties away in metal chimneys.

I like your vertically mounted "sub-chassis"—very nice layout; easy to work on, which is inevitable, as you say.

I'd be wary of metal chimneys at 10 meters.  More capacitance from plate to ground is not what's called for.  Down in the broadcast band it probably makes little difference.  I can't say that I've ever seen a metal 4-1000 chimney.  I'll look in my goodie boxes as it's possible that I might have a spare glass chimney.  It's the clips that are hard to come by lately.
At 5 kv that rig will make a joyful noise unto the world.
73 de Norm W1ITT

The metal cimmneys made me take notice too.  In addition to the comments already made, there's another glaring disadvantage...  a glass tube with its elements inside a vacuum depend on radiation cooling to get rid of the heat. A metal chimney is like a big reflector pushing the heat back in.  Yes, some heated air finds its way out with the pressured air, but a 4X1 with a glass chimney can really heat up surrounding RF components mounted nearby, proving infrared heat goes through the glass as part of the overall cooling design.   IE, don't put your finger on one of those metal, black receiving tubes... :-)

BTW, the best add-on mod I ever did for my 4X1 rig (four tubes, modulator and RF finals) was to cut two large 8" X 10"  windows for viewing - and use REAL glass.  A small hole is a peep show while a large rectangular viewing window is like being in the front row of a Pink Floyd concert.   I'll see if I can find a pic....

Use mirror clips to hold the glass in place. The glass gives easy access to the front components.   The RF deck on top sits on a swivel bearing and can turn 180 degrees for EZ access.  Your back will love it.

T

I don't know what application metal chimneys may have had. Was the 4x1 used in pulse or radar or xray circuits?  I was asking the boys around here about 4x1 chimneys, and this pair of metal  ones came up.  I havent seen them yet, but my friend will bring them when he comes by later this year.  I plan on using glass.

This might be late in the process, but if you are building a single band 2x4k for 10 meters, the best possible circuit might be a push-pull setup.
The advantage is that the ouput capacitances are in series and cross neutralization is very easy.

Also, the center tap is cold, so the requirements for an RF choke disappear.
Make it all from 3/8”  soft copper tube and use link coupling with a coupling link made of a piece of the inside of some RG-213, which you could put in a teflon tubing for good measure….
If you center the link, it will be in  voltage minimal, so less likely to arc also.
A real old school solution to higher frequency operation for efficiency and stability…

Single band amps always the best for performance.

73, Mike


There is a nice circuit in this publication that is adapatable to 2x4ks

https://www.smcelectronics.com/DOWNLOADS/1951-POWER%20TETRODES.PDF

I built a pair of 4-1000s recently.  Used a big Basler xformer out of a 5kw fm rig I got from Sam, KF7WWW.

One of the most underrated transformers EVER.  Guy I built the 4x1s for used it on a pair of 3000s!

Anyway, that one ran 6kv loaded on the 4x1s.  A couple amps of plate current was easy.

It was fun to tame down, I ended up swamping the input as you did and put a couple 1k ohm 2 watt carbons from each fils pin to ground.

Another thing I did was deconstruct the sockets.  Bullah shittah design.  Mounted the sockets to the bottom of the chassis and with the pins exiting the sockets at a right angle for the screen and grid pins, the distance to the chassis was halfed!  This is mo bettah! The filament pins remained the same.

--Shane
WP2ASS / ex KD6VXI

Nice design change on socket remount, Shane! Eimac just suggested using screen lead inductance be used as a series resonant wave trap to stabilize them at vhf.
They put the self resonant frequency at 25mHz. I have wondered if they calculated that with the tubes in the sockets?
Those cast aluminum abominations have a lot of inductance in the long bolt-like leads.

I never considered a push pull..  Good idea.
   I had  many new Eimac 500 cast sockets  I cut them down considerably for less  vhf gremlins and better air flow from a pressurized chassis.  Not ideal, and they still have the long bolts on the bottom , but I have used them in several amps including this one.  

Is there a non-plastic socket for the 4-1000 that has lower capacitance and respects the base cooling requirements? I remember something about the socket contacts should not be too fixed or stiff because it can stress the tube pins.

One feature of the long bolt cast sockets is:  the those long pin receptacles grip the pins very tight but they do float in the insulating base.   I have looked at the plastic Chinese sockets and they seem pretty weak.   There are some other sockets out there but I haven't tried them.

[/quote]
I talked to a ham today that said that back in the day he had some "special" 4-1000 tubes that were pulls from a linear accelerator.  They had filaments that were twice that of a normal 4-1000.    He used them in an amplifier.
"
[/quote]

Raytheon used a series of 4-1000A like tubes that I have picked up several samples of over the years. These were designed to be pulse modulators for Radar applications. The tubes that I have were proprietary, made by Eimac for Ratheon, and have a part number of 284 and 294. They are indeed 4-1000As on steroids with very rugged graphite plate structure and 7.5VAC 25A rather than the standard 21A filaments. I also have seen 4PR-1000 tubes which are the standard Eimac version of a pulse modulator 4-1000A. IIRC the 4PR-1000A has the rugged graphite plate but standard 21A filament current.

Shielding the amplifier...or shielding me from the amplifier....
  I have put the amp chassis in a relay rack. 1/8 inch aluminum panel for the front and bottom.  24 ga galv for the sides and back.  Panels all fastened in typical relay rack fashion with machine screws.  I have 2, 5 inch dia tube viewing windows made from the round front glass (plastic) and aluminum bezel from old Eico O-scopes.  I shielded these with bronze fly screen between the plastic and the alum panel.  the bronze screen is bonded to the panel.  No shielding on the panel mounted 4 inch bakelite meter.  I am wondering if this is sufficient protection from RF for the operator at hi power 29 Mhz.  Is there a simple way to test?  I haven't got this all together yet and I will post some pictures soon.   

Great job Steve.

The amp is working with 5KV plate voltage.  I'm very pleased with the performance.  10 Meters has been open and I have been using the amp on AM and SSB with good reports.  I have found several voltage settings that give me slightly over 70% efficiency.  55-60% at most plate voltages.  I have built and painted the rack panels and enclosed the amp. The power supply is in a separate 20 inch Bud rack on casters.
    Driving with a FT102 25 watts AM, the amp will forward modulate a 400 W carrier.  This amp loves voltage and loves drive.  The testing continues...

Steve, FB on the testing.

Did you try the "class C"  AM linear mode yet?  For AM and CW,  add a string of 1A  diodes to the cathode CT until the idle is cut off.   There's a good thread about doing this on the forum. It makes a big difference in plate color and less heat.  I'll bet that whatever efficency you now have can be increased another 10% at least. (on AM or CW)

I noticed the Winchester 30-30 brass on the windows; nice touch!   ;D

http://amfone.net/Amforum/index.php?topic=32635.0


T

Thanks Tom.... I read that thread and remain confused.....  Where would you recommend I start with biasing my amp?  Straight diodes?  Zeners?  What values?  I have my ideas but I would like to know what worked best for your 2x4.  Right now I have a 10K ohm wire wound switched in the ctr tap to reduce idle current.

Hi Steve,

I would use a rotary switch, maybe at least five positions.   Start with a big string of diodes to go gradually from class AB to simulated class C.  Each diode is about 0.6V drop, so figure how much bias you need for each mode.  The diode string goes in the fil xfmr CT lead with the anodes facing towards the tube cathode pin.    You can start with a big string and short them out as needed. Use clip leads to find the right spots at first.  Zeners work just as well, or a combination of zeners and diodes; the diodes will give you better fine selection.

The more bias, the more RF drive is needed on AM.  You will find a sweet spot balance between drive and efficiency.    The reduction in heat difference on AM can be dramatic, especially visual plate color.  If you are going to run it hard on AM, this is the best way I know.   You will find the spectrum is just as clean running AB or "C."  Read the thread results of a few of the tests guys made on AM.  (this will not work on ssb, just AM and CW)

http://amfone.net/Amforum/index.php?topic=32635.0

Disadvantage:  The ~20-30V diode bias is subtracted from the overall plate voltage.   But, the constant 30V bias subtracted from 5KV plate voltage is very little difference. (4970V)

T

The amount of bias needed will depend on plate voltage used.

MO volts = MO bias!

And you need a scope to keep everything clean.



--Shane
WP2ASS / ex KD6VXI

.......I built up a string of identical 2A  1kv diodes.  started with 10 then 20 then 30 in the ctr tap to gnd filament trans.  I tested at 2 drive levels and kept them the same for all the testing.  10 watts and 50 watts. The plate voltage was set and never changed.  Only the bias was changed by adding more diodes.  The amp  tune and load were retuned for  each drive change.  I did see some improvement at a specific bias range.  I didn't  see anything earth shaking however.  Just a few percent improvement... This testing was in the 5KV range.   I will test in the 3-4 KV range because that's where I got the best efficiency before. 
   TRP and VXI   Did you figure any efficiency percentage numbers for your testing?   I have mine....

Steve,

Just to explain it more clearly: It will require at least twice as much RF drive when biased into cutoff as in AB1/2.  So expect to need maybe 20 watts drive in class AB and maybe 40 watts needed in class C. That extra RF drive is not wasted, however; it is added as feedthru power just like any GG linear.

Did you bias it well below cutoff on AM?  The tube color reduction should be quite noticable if it is working as we expect.  Go WAY into cutoff and look for the limits where you run out of drive capability  and still have big peaks.   I drive my pair of 4X1s linear with a class C 813 -  series modulated by a class A 813 pair.  Plenty of CLEAN drive available.

T

Tnx Tom...   That is good infomation for me.  I'll let you guys know what I find. FUN Stuff.

You're going to need 25 volts or more of bias at 5kv, and this is dependent on the condition of the tubes.

You want next to nothing for idle current key down no drive.   On my Sb220, that meant 15 mils key down.

My sb220 had zero issues making legal limit pep on am that way.

--Shane
WP2ASS / ex KD6VXI

=========

Excellent use for CRT bezels! That plastic may not last wothout an air blast. On the othe rhand, if you have a skilled glass shop around they can make you a square piece with holes in it.

Class BC linear service is a nice way to run tubes for an AM only amp. When i used to run a FT-901 on AM i would even use it that way, there was enough range in the bias adjustments that you could pull the tubes below cutoff, and enough drive to be able to still drive it. Less heat out the back, slightly more power out, and I noticed you get slightly extended positive peaks vs running in AB.