QR...OH MY

[Mike/W8BAC]

Here is an interesting AMP for the "price is no object" contester. I think an am'er will find plenty of headroom at the legal limit. Better reinforce the desk and find all of the weak points in your antenna system before "it" dose.   Enjoy!

http://www.dc9dz.de/en/index.html

[kb3ouk]

i think that timtron's "blockbuster" used a 4CX10000 in it, i remember him saying something like it had 20db gain, only needed 15 watts in to give legal limit out.

[Ralph W3GL]

   WOW, now that's an AMPLIFIER...

[W7TFO]

Smart design, compact yet looks easy to service.  Good thing it is 3ph primary, otherwise the neighbors might notice all the available electrons going down a single path...

73DG

[KE5YTV]

Looks like the amp for the Death Star. 

Mike

[flintstone mop]

That is the meanest mofo I have ever seen!!!!
THREE PHASE power and 10KW out. UNbelievable!!

[KM1H]

I wouldnt want to be in the same room with that beast...its too crowded for long term reliability and if something lets loose say goodbye to all.

[K1JJ]

Beautiful construction. Talented builder.


That said, hmmmm... now the head scratching:

How does that relatively "small" roller inductor handle 15-20KW output without losing its mind?...

The amp needs to dissipate upwards of 15KW of heat in linear service to put out 15-20KW. The bottom of the tube looks kinda air restricted with all the parts in the way and the fan / area is not very big. Don't know how enuff air could flow to keep the tube from melting down.  Especially running quietly as a desktop. That thing must run as hot as a firecracker.

There's a few other things, but that's a start.

Interesting design with only 250V of regulated screen voltage and 7V of bias.  6KV on the plate is good.  

Getting -35db 3rd IMD at 15KW sounds very optimistic.  Most of those grid driven 4CX-XX,XXXX tubes are lucky to get -25 3rd IMD in linear service unless it's an expensive and rare 'J' designed for linear service OR uses a liberal dose of RF negative feedback OR possibly running near class A1.  The standard 4CX-xx,xxxx tube is designed for class C and running one at 10db more than the amateur limit will get someone lynched with splatter reports... . (To our side-channel neighbors, it's the same effect as running 1500W at -15db 3rd IMD )


But more "power" to him.  Always like to see guys push homebrewing to the ragged edge of wetness.

T  

[Mike/W8BAC]

Interesting points Tom. At first blush my reaction to the tank circuit (L and C) was the same and the pictures of the tube base neatly cluttered with circuitry and SO tightly spaced had me looking for the fan that was going to save this rig. Like almost everything else (except for maybe newer switching power supplies and class E rigs) size is everything. The Tsunami might survive limited use on side band but it is not a "brick on the key" survivor. That being said, I still want to be a beta tester.

Mike

[K1JJ]

Mike,

He certainly has an eye for detail.

Thinking more, an MRI pull, the YC-156 high-mu triode (3CX-5,000) would have been a better tube choice.  It's a cleaner tube designed for linear service and will easily put out 5-7 KW with about 200w drive. [200w edited]  Also a better match for the size cabinet and apparent cooling infrastructure. They can be had for ~$300 as pulls.

T

[W1AEX]

You know Tom, with that amp you could build a 50C5 final modulated with a 12AX7A and still strap! I love the warning about running the drier at the same time you try to run this amplifier off your 200 amp household drop.

[Mike/W8BAC]

I am thinking purely fantasy Tom but I like the gain of the 4CX1000 and Karl has put everything together so well. Either he is an absolute genius or he has a team of engineers and machinists, not to mention circuit board designers and so on. Why not bring the German electronics design culture into the twentieth century like the VW and Porsche and develop a water cooled 10K?

Think of it. We could have nice (GREEN) warm bird baths during the 160-20 meter wintertime peaks. I'll put a web cam on mine for the world to see. Question, how do you isolate plate voltage from the cooling water? KAZAP

Mike

[Todd, KA1KAQ]

Water cool the tube and use the bird bath as the cooling pond. 

[K1JJ]

Question, how do you isolate plate voltage from the cooling water? KAZAP
Mike

Yes, water cooling would get rid of the darn blower noise.  But where do you find a clean water-jacketed tube DESIGNED for linear service? Must be VERY expensive and rare.  There are 3CW-20,000's around but think they are class C designed tubes.

For HV isolation, use rubber hoses to interface and I understand that distilled water works well.  Otherwise, a closed system if OK.   Antifreeze might be good for winter. As I understand it, pure water is not a good conductor.


Rob, just think of what an amp like that wud do for your smugness factor.  You might become whirlwide famous in your neighborhood too.

T

[K5IIA]

http://www.ampsystems.com/asiprod-old.htm

they have some neat water cooled rf devices.

[Todd, KA1KAQ]

You could earn the well-known WEE (Worked Everything Electrical) Award or the less known but highly-coveted WAE (Worked All Everything) Award.

ETO Alpha made a vapor-cooled amp prior to the PA-77 single 8877. The PA-70V, which used the 3CV1500 tube. Hard to find today no doubt, but a high coolness factor.

Seems a skilled metal worker could make a water jacket for your favorite tube, though even if it didn't leak, it might not cool enough. Those sorts of things are generally part of the original design.

Still, it would be pretty neat to have a cooling pond for your amp or transmitter, ala WLW.

[kb3ouk]

i would like to know how that worked at WLW, those ponds were outside and i'm guessing not covered either. how did they keep the water pure? anything could get in there and contaminate that, i'd think.

[W1ATR]

How about a water to water heat exchanger to keep the toob clean? Distilled water on the tube jacket, and bird bath water on the dirty side. Distilled water is less conductive than tap water and a probe can be tee'd into one of the hoses leaving the tube. I don't know what a useful rating would be, but I would say a 10 or 20mA panel meter with the probe on one term and the other term to ground would be a good start. If you start to see a few mills current on the meter, it could be time for a water change. 

[Mike/W8BAC]

ETO Alpha made a vapor-cooled amp prior to the PA-77 single 8877. The PA-70V, which used the 3CV1500 tube. Hard to find today no doubt, but a high coolness factor.

The guy that sold me the 20V-3 was working on an Alpha 77 looking amp when I visited the first time. It was water cooled and he was "fixing" a leak. He had that blue silicone glue goobered on everything. Definitely a hambone approach. It was an interesting amp and the water cooled tube seemed small compared to it's air cooled counterpart since it didn't need a heat sink.

[KM1H]

Thinking more, an MRI pull, the YC-156 high-mu triode (3CX-5,000) would have been a better tube choice.  It's a cleaner tube designed for linear service and will easily put out 5-7 KW with about 100w drive. Also a better match for the size cabinet and apparent cooling infrastructure. They can be had for ~$300 as pulls.

Not even close. About 3KW with 100W at 5KV; 300W makes it talk. Bonus is no socket or neutralization. Con is 8 minutes warm up time to be safe.

Yes, water cooling would get rid of the darn blower noise.  But where do you find a clean water-jacketed tube DESIGNED for linear service? Must be VERY expensive and rare.  There are 3CW-20,000's around but think they are class C designed tubes.

It still needs a small blower for the 7V @ 100A filament seals. The 3CW20000A7 is designed for linear service and is a 3CX10000A7 with a water jacket.
The photo shows tube, socket, blower, and filament transformer. Its a pretty cheap tube considering what it can do, a bit less than a pair of new 8877's from Eimac. Its also available from China and likely with Eimac supplied tooling
http://www.g8wrb.org/data/Eimac/brief/3CW20000A7.pdf
http://tubedata.tigahost.com/tubedata/sheets/164/3/3CW20000A7.pdf

Carl

[WA1GFZ]

Remember the big amp Dean had. I think the tank roller was 3/8 tubing.
I bet this thing is a bomb waiting to go off. It is beautiful though but with a smaller tube running 5KW max. I agree with Tom a YC 156 better choice and no screen to deal with. less heater power.

[flintstone mop]

I thought Tom K1JJ would appreciate the Amplifier "under glass" effect. All of his known named efforts are in a plexi environment.

Looking closer it does remind me of the cramped ill-designed TenTec amp I had. Titan 425..........evil design. Cannot be repaired easily and prone to failure. Nice idea of separated power supply, but I'll take my Ameritron Amp any day over the TenTec.

[K1JJ]

Carl,

Yep, that's right about the YC-156 needing 200W for about 6kw out. I was thinking of the FT-1000D driver and forgot it was 200w, not 100w...


Good on the 3CW-20000A7.  It seems most of the high mu (7 designation) make good tubes for RF linear service, since that is what they were designed for. Easy to drive too, for triodes. But the tolerances between the fil and grid are so tiny they can be a bear in QC.  What does a good 3CW-20K-A7 cost these days? Do you have just one tube or a spare too?  Is yours an A7?  You wud need a stout IPA for it, of course.

The filament running at 700 watts is really not THAT bad compared to other tubes in that power class. But still can't beat an indirectly heated tube.

The blower noise is really a BIG issue for me. I refuse to have less than 40db S/N on my signal, so that's one reason I located my blowers on the outside wall, even for the pair of 3-500Z's.  Water cooling wud be a great method once we get a handle on the techniques.

Frank, you've probably had the same experience I've had with the roller inductors. Usually the wire holds up but the roller wheel contact itself pits and melts itself. The circulating current in the tank being Q* IP puts it in another league. Every attempt to use a roller inductor in a big linear here resulted in later cobbing in a big copper coil and bandswitch - or plug in coils.

Fred, yes I did notice the Plexiglas box...  My first thought was, "hmmm... I wonder where he got that idea"....

T

[W2VW]

Maybe Econco can stick a water cooled anode on a YC-156. Use dielectric oil through silicone hose to glycol to oil heat exchanger. Run the tube like a 3CX15000.

[KM1H]

ETO Alpha made a vapor-cooled amp prior to the PA-77 single 8877. The PA-70V, which used the 3CV1500 tube. Hard to find today no doubt, but a high coolness factor

They also had the 70A which used a 3CX1000A7, a much better tube than the 8877 for reliability; the socket cost was the killer. Thats the tube I use to loaf at 1500W on 2M, a early 70's made pull when I got it about 27 years ago and power hasnt dropped a bit since.

Good on the 3CW-20000A7.  It seems most of the high mu (7 designation) make good tubes for RF linear service, since that is what they were desined for. Easy to drive too, for triodes. But the tolerances between the fil and grid are so tiny they can be a bear in QC.  What does a good 3CW-20K-A7 cost these days? Do you have just one tube or a spare too?  Is yours an A7?  You wud need a stout IPA for it, of course

Last I heard about $1895. I only have that one lab display, never saw HV; just filament and water. My LK-500ZC could almost get it to the max and the 76PA would have some reserve for SSB/CW. For AM linear the LK would be sufficient for a 6-7KW carrier

There is no reason for another grid at that power, no need to attract attention with poor IMD. Thats why the YC-156 is so popular, be loud but clean when the juice is needed.

Maybe Econco can stick a water cooled anode on a YC-156.

Thats a good idea that has been suggested by several on other forums. However since the tube is not rebuildable I dont know if they would be interested. A good machinist should be able to shave the cooler for starters.