Discussion in 'Amateur Radio Amplifiers' started by NE7X, Jun 5, 2011.
Here is my workbench and the Henry 2K Classic under repair.
I'll answer after the big lightning storm stops.
I think I figured out what the problem is. Henry has another error, AND you did not do all the tests I suggested. I suggested you remove D106 and R109 totally from the circuit, connecting D107 to the negative rail. I think if you had done that, the meters would have worked.
Here is another error in Henry's design. D107 will have .7 volts drop. D106 will also have .7 volts drop. The sum of voltage drops of D107, R116, and R118 will exceed the threshold of D106. When that threshold is exceeded, plate current will divert to the chassis through D106.
That whole circuit around the lower end of the filter capacitor bank is a very poor circuit design. It does not protect the meters very well, and it has too much stuff to go wrong, and it also will be unreliable as configured. I know what they are trying to do, but it is a "half-fast" way to accomplish what they want.
They are trying to use R109 as a surge limiting resistor. That would require the meters (D107 and R119) be connected to the junction of C114, D106, and R109. The problem they missed is they drew the schematic wrong, and this silly idea makes threshold voltage of D106 critical. It also puts the ALC circuit and zener D105 at risk for failures.
This really explains it all. It explains why your meters are screwed up, and it explains why D105 failed.
Their idea would JUST marginally work if D107 had a lower threshold voltage than D106.
If it was my amplifier, I would add a proper well-insulated HV fault limiting resistor in the HV lead at J102, and get rid of F101, R109, R119, and D107 totally. I would have D106 in the circuit tied directly to the negative rail of C112 and R108. I would tie the junction of R115 and R116 directly to the negative rail like it should be.
If you do NOT want to do that, and do not mind suffering with the poor reliability created by their silly idea of letting the negative rail do the fault voltage absorbing, you could just add two more diodes in series at D106. That would make D106 three diodes in series.
As the circuit is, any connection of D106 to the chassis would make the meters right at the edge of not working correctly. It would be a horse-race between D107 and D106 to determine if the meters worked right or not.
All of these problems started because they wanted to protect the amp for HV faults with a resistor in the negative lead of the supply, which is the dumbest place on earth to do that. Then they probably had meters popping out of the panel when the negative rail lifted to -3500 volts in a fault, so they added D107 and F101 to fuse and protect the meters. D107 and F101 would save the meters, but make the circuit critical for conduction threshold differences between the diodes.
When this is added to the fact letting the negative rail fly up to -3500 volts in a fault will blow the ALC circuit and D105 out, it makes the entire concept they had "half fast".
If they had spent $2 for a high surge rated resistor like a RCD 175P, and put two 10 ohm 10 watt 175P's at J102, they could have eliminated all the mess of parts they added to correct issues. It would make diode D105 more reliable.
It is profoundly clear to me what they did, and the problems they had, that made them wind up using so many useless and critical parts.
This circuit can be fixed two or three different ways, BUT do NOT resize shunts to make the meters stop pinning. That is NOT a good idea at all. It is like using bailing wire to hold a fan belt together. Fix the real problem.
It does not surprise me the schematic, and the design has errors. I was forewarned by others, including the guy who I purchased the amp from that is was the case. He gave up after months of troubleshooting and this is when I cut a deal with him for the pile of hardware. I knew there would be major difficulties in getting this amplifier working, however the amp is too pretty to just scrap it for parts.
As the amp stand now, it’s working. The tubes are running a nice soft rose pink color and putting out a full KW+. I think the main issue I had was D1 zener wired per the schematic and in backwards. Once I rewired it correctly, the amp stated playing real sweet except for D105 100V zener blowing. Right now I have D105 and R113 removed. The only major issue now is resolving why the Ig meter pegs with more than 30 watts drive.
Reading your feedback, I believe this is your modification suggestion:
Remove F101/R109/D107 totally. Also remove R119?
Connect the Cathode of D106 to D101/C112/R108/R112. Leave C114 in?
Connect junction of R115/R116 to D101/C112/R108/R112
Add HV limit resistor (10 ohm 10 watt) in series with J102
Please validate if I have your modification suggestion correct.
Also, do I leave D105 / R113 out of circuit or put them back in?
73s, Thomas NE7X…
This sound slike a very useful article / web posting -- either on NE7X or W8JI web sites .. so that other Henry owners can correct these errors / short comings
Yes, 119 goes away.
I have not run the ESR specs on the capacitors Henry used, so 10 ohms might be an insignificant change in ESR. Henry tried to use 25 ohms there (R109). My guess is they put the surge resistor in the negative buss to save on insulation, because they would have had 4000 volts on R109 if it was in the + line. The thing they missed is if R109 is moved to the - line so a surge is limited, the negative rail would elevate to 4000 volts!!! This voltage is what back-feeds the 100V ALC zener and pops it. It also would destroy the meters. To prevent the meter destruction they started throwing parts at it, and that is why a very simple circuit became so complex and difficult to understand.
Since Henry screwed so many other things up, like D107 causing D106 to have inadequate threshold, I doubt the engineer ran ESR specs. So I'm going to counter his guess with my own guess, and assume the ESR is about 5 ohms. Without a B+ surge resistor fault current could be 800 amperes. 800 amps will blow a grid apart if a tube arcs, or explode the RF choke. A ten ohm resistor would take to fault to 4000/15 = 266 amps. Many parts will survive that, but you would be better at 100 amps or less if it is workable. That would dictate 35-40 ohms and a 50 watt resistor.
I'm not trying to sell you a part, you use what you like, but I have suitable ~40 ohm resistors here as a single unit, and suitable insulators. If you didn't want to use a single resistor, use a 175P ten ohm and connect three or four in series.
This would be a conservative guess.
They can go back in. If you make these mods, you will probably never blow D106 again.
After all this work, Tom should write it and we should at least share it. This wore me out, but at least I learned something I'll never use again.
Thanks for taking the time to help troubleshoot this and identify the mod. I understand the HV limiting resistor should be somewhere between 30-40 ohms, maybe 50 watts and on HV stand-off insulators. Let me check my junk box inventory, I have some high wattage low ohm resistors on ceramic insulators. I need to see what I have. If not and if you have one, I will let you know and happily pay you for one.
Once the mod is in and everything I working, I will write up a short article and add it to my http://ne7x.com web site along with all my Henry QSK modification information. I agree, its good info and should be shared with other Henry 2K classic owners.
OK, I officially consider this thread closed. Now, who buys the coffee?
This has been a good thread to follow.