RF deck rework

[KA2QFX]

Hi Patrick,
It's been a while since I saw this thread. Looks like progress is being made.

In reviewing I noted a few things, but I'm not sure why you're seeing low efficiency. Frank's concerns about the strapping and plate lead inductance are well founded but you should likely be safe for freq. below 40m...  for now.

1. On the U-joints: the coupling shaft end yokes should always be 90 degrees apart (out of phase so to speak), yours are in-phase.

2. Regarding input power dissipation with tube cold: That's a tough one.
I noticed a series resistor in the G1 lead before you changed the input tank. Is that still there? If so what's it's value? If it's there for parasitic suppression I'd consider (DC) shorting with a turn or two of wire making a more conventional para. suppressor rather than a resistor. Keep in mind that the grid is a pretty low impedance during the conducting portion of the RF waveform.

 

3. Have you considered that the MB-40 is resonating on a harmonic because it doesn't like the impedances it's seeing?  We saw this when initially playing with input circuits for W1IA's Ampzilla. Once we got the ratios figured out things worked better. It looked like what you describe. Very low efficiency, but there was significant (albeit low) output at the desired frequency but the input was predominately 2 or 3 times F input.  

I read the spec. sheet on the MB-40 and wonder if you've changed the input winding for 50 ohms as National suggests?  Is N^2 anywhere near your calculated input/grid impedance ratio?  Etc, etc.



Just thinking out loud... I'll be interested to hear others thoughts and follow your progress. Good luck.

Mark

[Opcom]

OK here are the fruits of today's labor after the Texoma Traders Net.

The normal/test momentary switch was replaced because it had bad contacts. This was used for keying the transmitter momentarily from the front panel. I used a center-off on-off-momentary DPDT switch so that when the key is down, it can be keyed continuously (normal when the key is locked down), on standby (middle pos.), or keyed momentarily regardless of the key position. This extra position is important because of having only two hands.

The dials were keyed on the shafts incorrectly such that the 4-1000 was tuned to 14MHz but appeared to be on the 7MHz band and was being driven on 7MHz. So that explains the bad efficiency, but OK for a doubler.

I have a good set of operating conditions on 7220 but had to stop because that 56 Ohm 2W carbon resistor in series with the grid lead was the culprit and has run out of smoke. I do not know why a 50 Ohm resistor can burn with 30mA grid current going through it but I guess the RF current was much higher.

The number of turns on the MB-40 was set for 50 Ohms before, and is 4 turns on the coupling loop.

Driving with an IC-735, these figures were obtained.
F = 7220KC
MB-40 dial = 15 (my dial is reversed so that is "85" by the book.
Plate voltage = 3000V
Plate current = 340mA
Grid Bias = -160V
Grid current  = 30mA
Screen voltage - 380V
Screen current = 85mA
RF output = 800W
Drive = 28W (forward power)
Drive SWR - 1.7:1

These results are much better. I may be able to replace the resistor without pulling the RF deck and find out what happens next on the other bands. I have not calculated N^2 or the grid impedance. N^2 is going to be tricky because it may be one, or the other, set of coils in play on any given frequency on an MB-40. When it is out I will count up the turns.

Thanks!

[KA2QFX]

The dials were keyed on the shafts incorrectly such that the 4-1000 was tuned to 14MHz but appeared to be on the 7MHz band and was being driven on 7MHz. So that explains the bad efficiency, but OK for a doubler.

Murphy and his minions.
That's essentially what I figured was happening. But it's better that the cause was a mistuned output rather than a mis-resonant tuned input. That's easier to correct.

I have a good set of operating conditions on 7220 but had to stop because that 56 Ohm 2W carbon resistor in series with the grid lead was the culprit and has run out of smoke. I do not know why a 50 Ohm resistor can burn with 30mA grid current going through it but I guess the RF current was much higher.

27 watts of drive, 2 watt resistor...
I would put a turn of wire around the resistor and just make a conventional low-Z grid suppressor.  Let the MB-40 handle all the transformation/matching to the grid. It should take a LOT less drive too but input SWR will be a tad touchier. 

The number of turns on the MB-40 was set for 50 Ohms before, and is 4 turns on the coupling loop.

Sounds like you're almost there. Kewl!

Mark

[Opcom]

Pulled the RF deck back out. After spending 2 days cleaning up the place and clearing off the workbench. Just never time to work on everything.

So the 50 Ohm resistor ought to be replaced by same with a turn around it as a suppressor.

[ke7trp]

Patrick.  Fire up the filiment and blower (or rig a fan) on the bench. Then you can tune, adjust and experment with the grid while watching the SWR meter.  I had all kinds of trouble with my rig in this area and after 40 having high SWR, I ripped the deck out and set it up on the bench with the filiment on.  You dont need the plate supply here.

I used the MFJ259 and adjusted the coil for the best match. I got it 1.0 50 ohms on all bands and now only needs HALF the input power it used to being out of match.

C

[Opcom]

That's a good idea. 100+ screws to remove and replace is a bust.

I'd rather get rid of the resistor because it was not originally there. It was added to get rid of a parasitic, but the real condition causing that has been fixed.

The grid uses an MB-40 and besides the tuning function it also has a swinging link of 4 turns to be closer to 50 Ohms. I had it set full into the coil but I'm wondering if it ought to be set up another with another shaft and swing the link for loading just as if it were on a low power plate circuit. There is already an spare hole in the panel and there's have another spare vernier drive knob somewhere.

[ke7trp]

The scews are a pain.  Go to lowes and get the orange black and decker mini screw gun.  Its cheap, Has Lion battery.  It will spin the screws out in minutes and the battery lasts for a very, very long time.  I use it almost every day.  Everyone that has seen or used mine, went down and purchased one. 

C

[Opcom]

I have an electric screwdriver. They are flatheads, which is an inconvenience.

[N0WEK]

I have an electric screwdriver. They are flatheads, which is an inconvenience.

I assume that you're going to use Phillips head screws going forward?

I mostly throw all my flathead, straight bit screws away if I can find the equivilant in Phillips or Square drive.

From the pictures, the guy did love his screws! 

[Opcom]

I will likely go back with non ferrous Phillips machine screws. Each and every hole is also tapped, yessiree.
Durward Tucker (Colonel Tucker) took two years to build it so there was time to drill and tap some 400+ holes. He completed much of the work at the Flagpole Hill studios near White Rock Lake, Dallas, the transmitter site of the city-owned 5KW WRR-1310 AM in during 1956-1957. He operated AM and also wrote several QST articles on Phone.

How did he get away with this? He was the managing director at WRR, chief engineer of the Dallas municipal radio department, director of civil defense in Dallas (that office in WRR's Fair Park studios), and communications officer for the 9th MP Brigade, TXSG. It was not likely anyone would have dared to said a word about the extracurricular activity as long as things ran smoothly. There's a bio here: http://bunkerofdoom.com/tuckerkw/about_Durward_Tucker.html

Time to make a T/R switch for this thing too as it was never in my hands. A simple large relay will do.

[W7TFO]

Classic fasteners require a bit of skill & talent to master.  Old gear.  Old guns.  Old cars.

How many have we seen that have been AFU by someone with the wrong or poorly fitted tool?

If you build to period pre-WW2 style, you wouldn't even think of using a Phillips, square, Torx or the like.

Straight blade is the most popular from the time (run with properly hollow-ground drivers), with full hex and some Allen acceptable. 

Stainless threaded fasteners tend to gall and become irreparably locked at the worst of times.

If you build modern style, well, just glue it together.

73DG

[Opcom]

I mastered slot type screws a long time ago. I am not sure why they were used so widely except maybe manufacturing cost for the screw and drivers.

I did operate the RF deck on the bench to fiddle with the MB-40 as suggested. The link must be in all the way for best results here. When tuning it around the frequencies, the R is about half the X as displayed by the MFJ. Its never 50, and X never goes near zero. The SWR on that MB-40 is poor depending on band but the ricebox will drive it enough that I don't care.

RF deck is back in. removed 56 Ohm burnt resistor and it is now just a piece of wire, added from grid to GND a 3100 Ohm 40W non-inductive "glowbar" resistor in series with a mica transmitting cap for some swamping. 10W drive is making 500W output. Any more and the screen overload trips at 60mA and I do not know why as before it was running 85mA no problem. The pot has probably not been touched in 50 years.

Tried to have it back together for the rally but no dice on a speech amp or T/R switch yet. Listening is OK for tonight.