Charlie was great at designing triode amps, they were works of art and they worked well. I first met hm at Henry Radio, he was able to get out from working in that sweatshop and started u his own biz.
I was often down in SoCal at his shop where he did 98% commercial / industrial. The last 2% a mix of CB and Ham customers.
Sadly the 4CX20,000A linear he built was the last thing that came between us, when he learned I was a friend of Bruce Gary who had bee asking me for help from the Middle East where he had to do a major redeign of the chassia and PS to make it work, it was too much for him to accept and the last call from Northridge occurred when he hung up abruptly. Loss of face they say, a common Asian malady.
One problem disovered was that despite my advising "never let the screen grid float" thats what was occurring wen the vac relay in series with the sceen to the screen supply was energized and it let the plate current go towars infinity.
I also do not recommend using a FW bridge followed by umpteen hundred Microfarads in the filter. A recipe for a tube failure if/when a glitch occurs, Bruce added a glowbar resistor to replace a wirewound resistor that had blown into shrapnel. All the stuff I used to recommend to has was sadly ignored on the AmpSystems calendar.
Reid W6MTF
I echo what KH6AQ advised. In addition an increasing load SWR could be a factor, I saw that happen once on a smaller triode. I would assume its being monitored closely though, so focus on the plate tank, especially a blocking cap that might be heating up. Do replace the grid sensing though and continue to report. All hail the mighty YC156! take good care of my baby! 73 de Reid Brandon W6MTF
Is this THE "Reid", friend & mentor to Bruce Gary (SK)?
N0UN
Bruce Gary was a very smart guy. He flew over to the middle east, and fixed all those grid driven Asamoto amplifiers, that all oscillated when fed into a resonant ant..( and blew up the xcvr to boot)....but worked just fine into a DL.
I just went through a similar scenario with yet another Asamoto grid driven tetrode. Worked fine on 75m into an ant, or DL on 75m....( worked on 75m but real high gain, like 22.5 db)... but would osc on all other bands, whether into a resonant ant, or DL...and that was with amp keyed, and NO drive applied, go figure. Would osc with a DL on both the input and the output.
The fix was to completely rebuild the PI tuned input. The mile long path was eliminated. Terminating resistance reduced from 1000 ohms, down to 500 ohms. On the advice of Robert, W0VMC, a small parasitic suppressor was inserted right at the grid input of the socket. The new coupling cap was wired directly between the 500 ohm terminating resistor ( 2 x 1k flat packs in parallel, and heatsinked to rear chassis) and the new grid parasitic suppressor. Virtually zero lead length anywhere. Bypass caps installed on both sides of the cathode. (oem config only had bypass caps on left side of cathode).
Only one end of both C1 + C2 caps used to terminate the stator.
Ok, now it's 100% rock stable on all bands 80-10m. Gain reduced to 20 db. Neutralization not required. (Oem config didn't have neutralization either). Input SWR is 1:1 from 80-17m. On 15m, it shows a 1.5:1 swr, which is caused by the huge 100 pf of input C of the 4x10 tube...( and 160 pf on a 4x15). That C can't be absorbed by the C2 air variable on the PI tuned input. The fix is a relay switched small coil, to cancel out the XC of the tube. Only used on 15-12-10m.
...
The original oem config had current traveling through a mile of coax,from input side of TR relay, then up to one end of the hotside of the C1 air variable cap, then through the entire length of the stator of the C1 cap, then through the roller coil. Roller wired wrong, with current flowing the entire length of the central shorting bar, when on upper HF bands, then through the entire length of the hotside of the C2 air variable, then a mile of strap to the hotside of the 1000 ohm globar. Cold end of same (really long) globar bonded to chassis via a tall, solid aluminum standoff. Return current path is then through the aluminum standoff to chassis, then through a mile of chassis..and back to bypass caps at only one side of the cathode..at the socket. No grid suppressor used. It's no wonder it oscillated. It was a dog's breakfast from day 1.
On another grid driven amp, this time a hb 4x5 grid driven affair, a 450 ohm globar is used in conjunction with a 1:9 un-un. No pi tuned input. "no tuned input required". All sounds wonderful, until you realize the tube's 100 pf of input C is directly in parallel with the 450 ohm globar. That just screwed up the input swr, on all bands. An equal amount of XL has to be inserted in parallel with the 450 ohm globar....and that's a lot of coil required for the lower bands, since the Xc of 100 pf increases... as freq decreases. EG: 100 pf = 418 ohms of XC @ 3.8 mhz. 418 ohms of XL = 17.5 uh. A small 20 uh roller coil had to be used, and tweaked one each band. On 160m, it's worse, with 35 uh required.
Multiband, grid driven tetrodes, using large tetrodes is not as simple as they appear. If done correctly, they don't have to be neutralized. If they do require neutralization, that opens up another can of worms. The typ, small value vac cap used, can, and will develop 'whiskers' with B+ applied to one side of the vac cap. A special, inert gas type is typ used, or the conventional vac cap has to be over rated. Depending on input config used, the neutralization cap will interact with the tuned input, throwing both off. Then you also have the screen and bias supplies to deal with, plus screen and plate over current protection. Ok, rant over.
Agreed. On yet another asamoto 4x10, this time a 6M cathode driven affair, with screen bonded to chassis, it would not resonate on 50-54 mhz. Blown up vac relays in the screen and bias supplies, ( blown to bits) and a massive 2 kw CCS screen supply used. 1 amp fuse in the screen supply blown to bits..and ditto with the 5 amp fuse in the B+ supply. Massive grid xfmr used as well.
Most bizarre tank circuit I have seen yet. 50 kv (50,000 volt rated) rated bypass caps at the cold end of the fil choke. 10 x huge bypass caps at the cold end of the too small plate choke. Vac load cap in series with the output.
5 of us worked on it for months....remotely. The amp had sat in storage for 10 yrs and was never used, or tested / used by the original owner......then was re-sold...10 years later.
Cut to the chase. With a resistor from anode to chassis, to simulate the plate load Z, it would only resonate on 70 mhz wtf? Part of the bizarre output network used a pair of 17 pf, HT-57's in SERIES..from anode to chassis...( 8.5 pf total)
1 of the 17 pf caps was removed ( now 17 pf in total)..and presto, amp resonated on 50-54 mhz. 8.5 kw out with 250 watts of drive. (up to that point, everything in there got vaporized, since it did put out power on 6m, just not a lot).
Turns out charlie had ..'borrowed' the brand new eimac 4x10 from the 6m amp..to use in a FM broadcast PA. 6M amp sat with no tube for months. When original owner came to pick up.... charlie stuffed a new, just returned, rebuilt eimac 4x10..that had just come back from ECONCO. Well, the ECONCO tube had internal capacitances, that were just a few pf different from the oem EIMAC tube. ( anode to screen, screen to control grid, control grid to cathode). The ECONCO tube runs superb, but the internal C's was the last thing I would have thought of. Last year.... No help from Charlie....he initially denied building it..... until the PDF schematic with ASI, Northridge was pointed out. Still no help. "Worked when I tested it into a DL, 10 yrs ago". "you are on your own".... then hung up.
Checking real old notes, I remember reading warning notes from one of the commercial FM broadcast PA manufacture's about serious issues, when end users swapped from EIMAC tubes to SVETLANA tubes. Similar deal. Output circuit had to be re-tweaked to handle the russian tube, due to differences in internal tube C's. In some cases, the output circuit flat out could not be re-tweaked, and component values had to be changed out. This was on the 4x20 series of tetrodes.
On that 4x10, grid driven HF amp, on my previous note, charlie used a a single 5 kv rated, microwave oven fuse, in series with the B+...and only a few inches from the anode. Those chinese fuses are not much longer than a 3agc fuse, and the 2 x piece, plastic, inline fuse holder, leaves a lot to be desired. If it opens off, and with screen voltage still present, the owner is gonna have a really bad day.
At least on the Aerocom brand series of grid driven tetrode HF amplifiers (4x10, 4x15), they used extensive (adjustable) anode, screen, bias protection. Anode, screen, bias current protection. Same contactor feeds the primary of both the plate and screen xfmrs. Anything goes sideways, anywhere, and it all kicks offline.....and fast...and anode + screen V removed, in the correct sequence. Drive removed etc.
John Lyles, K5PRO, who used to work for BE, etc, and now works at Los Alamos, recommended using the Type AS tubular resistors from Globar (now owned by Ohmite). These types are made specifically for high energy discharge applications. I managed to buy 4 x 200 ohm versions of these Type AS resistors from SSON years ago, cheap. 18" long x 1.5" diam. 4 wired in parallel = 50 ohms. Each is rated for 120 kv..and each is rated for 119 KJ...( < 10 msec pulse). Preceded with a BUSS HV fuse. I use this setup on the hb 5500/7700 B+ supply. Buddy used 2 x 100 ohm, wire wounds in parallel, (225w each)..and one day had it overfused..... and it cut loose. The wire wounds had gone open, all flaked on the sides. Looked like u ran a propane torch on em. Even the globar type SP resistors, used in big dummy loads will work in glitch service. Ok, enough of all this. Take care Reid. All the best for the new year.
