The AM Forum

Yesterday, my friend Dennis W5FRS helped me with the big old TX I have been working on repairing for years. I've already mentioned it many times as different work was done on it.

It is finally working and on the air  but there are some small issues. The PA is a grid driven 4-1000A, the modulators a pair of 3-500Zs.

A screw was found to be too close to a connection lug on the MB-40 grid tank. The distance was less than 0.005". The connection was to the PA grid. A shorter screw was put in because of the high voltages present there. If it was a dynamic short, that explains much about it starting and stopping oscillation for no reason. Because of continuing issues with the MB-40, Dennis advised replacing the original choke with a 250 ohm 25W WW resistor and not bypassing the center tap of the MB-40, but doing it after the resistor. This all seems to have made it very stable. Some testing was then done. The amp to drive it is a Stromberg-Carlson hi-fi amp, 30W, PP class A 7027s. The 4 Ohm output goes to the 11V 15A secondary of a power transformer, and the 120V primary of that transformer feeds the 500 Ohm input to the TX. Inside the TX a 40 watt UTC driver transformer is there, feeding the 3-500Z grids.

1.) regardless of which way the polarity is done, the modulation on voice hits zero volts before hitting carrier x2. On a sine wave, it is not an issue. This went away by increasing the plate voltages to 3000-3500. At lower voltages, it works right with the modulator voltage kept slightly higher than the PA voltage.

2.) The mod iron, presumed to be from a KW RCA BC TX, seems fine but with the PA at 3KV and the modulators at 3KV, the thing was barely able to get 100%. -this was fixed by increasing the grid drive and screen voltage.

3.) the 3-500Z modulators have no bias and run much current (Pd=800W) at zero signal/3KV. This is supposed to be a class B setup so I have to make some bias for them. I assume this should be heavily bled and well regulated. The old modulator bias supply was for 304THs, and therefore is some 300-400V. This is not usable for the 3-500Zs. Perhaps the modulator will make more power once this is done.

4.) Possibly the modulation driver transformer is not quite suitable for 3-500Z grid-driving. A concern is that it is a higher impedance as may have been for the original 304THs. I will not switch back to 304s. I must use commonly available tubes.

The attachment shows the grid circuit as it is. The resistance is right next to the MB-40 tank. If nothing else it is a lossy RFC. This seems to work so far.

It drives easily.

Today had the first QSOs with it.

There were some mysterious emanations from it this morning that delayed me getting on the air. The reflected power would jump up and down and the RF on the scope was looking very nasty. At first I thought it was the antenna or a loose connection.

The cause was made clear. The "3KW" MFJ-939 Antenna tuner blew up, but I had only been running 400-600W carrier into a dummy. what junk.

Going on the theory that the smoke comes from the fault, the TX "cleared the fault" in the MFJ. At this moment the RFI issues, including audio feedback and someone complaining about their TV, disappeared. The problem was the MFJ-939's SWR/Watt meter circuit. Now that it's dead because it could not handle something within its supposed specs, no more issues.

The TX can work on 7160 with the antenna alone, the reflected power was only 40W or so. I have parts for a real tuner, a future balanced-balanced tuner project to have RF ammeters. I have to check the Tucker on all bands, and if satisfied, it will be time to unload the hoard of parts and equipment supporting it's rebuilding and no longer needed, clear out some space, and work on the tuner.

I also blew up a low pass filter. In its place I put a very clunky and old B&W one. Fixed that.
The original filter may have been weak but it was OK with a 800W PEP amp I have been using. I pick filters up at fests for $10 when they are offered, and if I-fix-it, I win. Can't loose $10 on a box with two SO239's.

today's conditions:
Tucker first transmission 2012 12 28 1223 CST
Po 400
PA:
Ef 7.5V
Eb 2250V
Ib 240mA
Eg -150V
Ig(keyed) 10mA
Ig(drive) 20mA
Eg2 450V
Ig2 40mA

MOD:
Eb 2500V
Ib (resting) 240mA
Eg 0V


The audio cabling and keying cable have been run temporarily and I actually put the back doors on the unit and closed them, removing the vise grips that had been holding the interlock down. I consider that a milestone.

remaining oddities:
the screen supply has awful regulation. 700V drops to 400V with a 50-100mA load. Have to look at that again. no problem, 400V is enough.

I have yet to mute the Sunair GSB-900DX receiver. Using a Johnson TR switch, the 4KW one. So there is one button to key up, and a knob to turn down the RX volume. Almost a one-push operation.

This set wants to run 600-800 watts carrier. It's most efficient with 3500V on the plates. The best way so far to reduce power to keep the peaks legal has been to turn the 4-1000 plate voltage down, leaving all else alone.

The gentlemen that were nice enough to talk with me were:
Darrel   WA5VGO   35mi. N of houston   KW-1, 1/2 wave @ 60FT
Dennis   W5FRS   Irving, TX   250TH, 90W carrier
LEE   W0VT
Joe    W4AAB
John   K5SEE   N of Houston   Henry Clasic, 350W carrier, R75 RX

I got good signal reports, a good hum-free report, very clear, pleasure to listen to, and report of 'no buckshot'. One had a fancy receiver and tuned around looking for defects but none were to be found.

I hope COL Tucker is happy. I did not want him to spin in his grave due to the TX not getting used, but given up on and parted out.

Finally, the Dallas ham club refused this thing many years ago when it was still in one piece and usable. Hell has no fury like a transmitter scorned.

Great news, Pat.

I am proud to have made a slight contribution to making your maul come to life. :D

73DG

Glad to make the QSO today. It is just karma that the Dallas club passed on it, and you resurrected it. I have benefitted from the appliance operators being afraid of high voltage and tubes :-).My antenna doesn't favor Dallas like it does Houston. I am thinking about running multiple slopers on 40. Sorry you found out the hard way that mfj stands for miserable freaking junk. Look for a BC-939 tuner. It should handle the power.Wishing continued success with the Tucker rig.
                                        Joe W4AAB

FANTASTIC!

More than a slight one Dennis! The many times you have sent me parts or materiel, and our couple of trades bear some responsibility for the rig's operation!



There are a couple of gutted BC-939's here, the big rollers mainly, also some other parts that are heavier duty.

My question is what values are really necessary to cover 160-10M and a wide range of impedance.

I had in mind the rollers might be good for a balanced-balanced arrangement. Do you believe they will really take a 1KW carrier and 4K peaks? What about the output of a 3CX3000 amplifier? I have been repairing that as well. slowly. I do not plan to operate at those levels but the stuff has to be able to stand up to it if it happens by accident. I don't want to reproduce the MFJ and have another crapout.

Funny the MFJ-939 and BC-939 numbers. presumptuous?  or coincidence.

There's been an idea for a while to use a shallow BUD rack and build up a big tuner on rack panels so the components are inside the cabinet -better for shielding- and face the panel side to the wall near where the feed-line comes in. The 'back door' of the rack would face into the room. I have motors and the rest to turn coils and caps, so the action can be remote, with pushbuttons at the operating position. All the main parts seem to be here.

I want to overbuild the tuner because my antenna is short for 160M  - it is really only a half decent match on the 40M band. It is very hard to match on one band, i forget which, but the feed line voltage gets very high. I think each leg is about 65 FT, but there are two legs in 'parallel' on each side of the 'dipole'- like a fan dipole with all 4 legs the same length.

The basic idea of the tuner is attached. nothing new.. image from the www and edited.

T1 would be a 1:1 current balun, a commercial 5KW unit for 160-10M.

L1 and L2 could each be:
a BC-939 roller - lots of L.
or
two 15-20uH broadcast-level rollers in series -indestructible.

C1 could be
an existing 1400pF per section dual air variable with 1/2" spacing. It can be configured for two 750pF sections at 2x higher voltage because it's really made of four 750pF sections.
or
a single Jennings 30KV vacuum cap, 100-1500pF
or
there are a couple of 1100pF 3500v air variable caps, but they seem small for this whole matcher thing now.

The relay or switch may be a broadcast antenna switch on hand (thanks to DG), rated some 20KV/15A. It's motor driven. If the tuner were to be very simple, it could instead be one of those big knife switches.

Ammeters -where to put them? after the L?
I have two 15A (I think) thermocouple units been saving. That's too big.
Because of the wide difference in current at a given power and frequency, the reading could call for almost anything from 0.5A to 10A under common conditions. No one meter range, it seems, would be best.

I regret I won't be on today. I was planning on this Saturday, however yesterday I got hoarse and was coughing up stuff, so I think I have a cold and need to stay inside in the heat otherwise it could get bad from previous experience.

FB project. Thanks for the writeup.

Zener bias the 3-500s. No worries about building another supply.

Make sure there are parasitic suppressors in the modulator plate leads if not already there. Usually just a large power resistor in each lead. Check out broadcast rigs.

A balanced-balanced tuner can handle anything you care to throw at it. I've built many. Also built plenty of link coupled stuff too. For QRO or inexpensive design simply skip the roller inductors. Wind your own and make adjustable taps. This circuit can be difficult to adjust initially but is well worth the effort. Matches stuff a Jaw-and-son Matchbox simply can't.

IIRC my 4-1000 plate modulated rig used to run at 38 mills grid current. Not enough grid current and the modulation linearity goes out the window. The grid tank in mine had an arcing problem from lack of spacing. Great way to make friends!

Patrick,

Very nice job and extremely well document as others have noted. 

Tell me more about the name Tucker.  It sounds familiar maybe since I remember a Tucker car.  But is does sound familiar in radio terms. 

Help!!

Joe, GMS 

Hope to catch you on 40m with it soon, Patrick. 
Dallas to Houston isn't a stretch.  Bring the thing to 3.880 sometime and call CQ. 

Using the 1980 Eimac specs the 3-500Z is rated zero bias only to 2500V in AB2 GG SSB linear service and yet they claim the same as AB2 modulators at 3000V and 300ma idle. This is really pushing it even with air system sockets and chimneys. Thats too much IMO and Id suggest biasing for 150-200ma for the pair which is within the ratings of a SB-220 type of fan or a good all metal EBM-Papst muffin fan as used in the LK-500ZC which also directs air at the filament pins.

You can use a string of 1N5408 diodes (reverse connected compared to a zener)and a switch to select different idle currents if you want to run distortion tests. Use a fat electrolytic, 10K uF or more, across the string for stiff regulation. The SB-220 runs 2500V key down and about 120-130ma idle and distortion products are down around -33dB which should translate to a clean AM signal from the tubes anyway using the GIGO rule.

Carl

That was the last name of the guy who originally built it.

And here I thought it had something to do with Tucker Electronics down there in TX

Durward Tucker W5VU( SK) at one time was the RTTY Editor of CQ( in the 1960's after CQ got rid of their original RTTY editor and editor-in-cheif(guy named Green). Durward Tucker was a Colonel in the Texas State Guard.

Durward Tucker - a short bio is here along with some pictures and references of his writings and an article about him.
http://www.bunkerofdoom.com/tuckerkw/Tucker_Transmitter.html

The "Who Is Durward Tucker?" article may not be in QST, not sure, but there is one in a ham rag.

Latest news, very poshumous - In 2007 Durward J. Tucker, W5VU (SK), who helped promote and popularize RTTY in the 1950s was inducted into CQ Amateur Radio Hall of Fame.


Obviously the TXs web page needs major updating. The most recent info on it picking up where the bunkerofdoom leaves off is here and on amforever.
The members of both AM boards have been very helpful in this work, offering everything from advice to criticism to praise.

The pic. might be about 1957-1960.

Maybe fabbing up a pair of adapters to use 83's rather than those 5R4's would limit the droop in the smaller power supplies.

Rather than using diodes, that is. :P

73DG

Thanks for the information.  Very interesting indeed.

Joe, W3GMS

Congrats, PJ!! Glad to hear it's finally back on the air. I know you've been tinkering on it forever. Damn, I remember when you got that thing, must've been at least 10 years ago? Isn't it blue or blue-gray?

There's something to be said for bringing a piece of gear back from a long sleep to become a working machine again. Maybe not the same as seeing your own homebrew creation come to life, but right up there. I was able to finally get an old rig running this summer with the help of Johnny, W3JN, after it had been off the air since long before I got it in 1985. Felt good to see it making RF again and to hear someone answering Johnny's call. I'm sure you were all smiles.  ;D

Hope to work you on it sometime soon.

That's really cool !  Definitely looking forward to working you with that rig.

Hope you are feeling better, also.

At night, we should be able to easily work each other on 75 meters assuming we don't have all kinds of static.  Will be around tonight (Sunday) and other evenings as well.

Again, nice bit of work.

Regards,

Steve

There's nothing like getting an old xmitter back making RF after a long sleep!

Pat, I'm wondering with all the blown up stuff if you have a parasitic somewhere, or excessive harmonics.  You have access to a spectrum analyzer, correct?  It'd be a good idear to check the output spectrum (be VERY wary of overloading the spec an's front end - start out with plenty of attenuation).

I like your suggestions John.  The fact that the LP filter blew up may have been caused by too much energy in the filters cutoff region.  Then again, it could just have been a bad LP filter.  You may also want to check to see if the final is neutralized.  Big projects like yours I always tend to re-establish the baseline and first start with the transmitter into a dummy load and start checking for stability of the final and other parameters. 

You may want to consider getting rid of one of the audio driver transformers.  A properly selected PP plate to 8 ohm output turned backwards would provide a solution to couple your Hi Fi 8 ohm output right to the grids of the 3-500Z's.

Jeff, W3JW is using a pair of 3-500's to modulate a single 833 in his HB rig.  I am sure an email to him would get you some particulars on his modulator.  I know it sounds very good on the air.

I like zener shunt type bias supplies when biasing modulators as mention by Dave, W2VW as well. 

I really enjoyed hearing the Tucker story! 

Congrats once again for bringing it back to life!

Joe, W3GMS       

And here I was expecting a Tucker with a steerable headlight!  :)

With my HB modulator for the GPT-750, I use 3-500zs for modulator tubes, with a zener diode string between the centretap of the modulator filament transformer and ground and a good amount of capacitance across the zeners. I chose the zener diode value so the modulator tubes would idle at approx. 90ma. I have a 100 ohm 2 watt carbon comp. resistor in the grid of each 3-500z.

Works great, with no instability noted.

Patrick,

    I have some 6v and 12v TO-3 case zeners. I can send them your way, just PM me.

It might be more practical, and Col. Tucker era compatible to just use a simple unregulated DC supply, heavily bled (> 100ma), and bypassed. Not being regulated can be a good thing since the bias varies with line voltage. This helps keeps the idle current more constant. Throw in some selenium rectifiers as well since they were in common use around 1960.  :D

Still waiting to here you on....

Jim
WD5JKO

While using a zener string or other voltage between the C.T. of the heater transformer will provide bias, it also reduces, by the bias voltage, the voltage available for modulation.  Albeit, it may be small, but if it's, say, 100V, that's 100V that's not available for modulation.

Most mod transformers in use today have poor ratios with respect to getting a lot of positive modulation.

I've actually run tube modulators with *negative* voltage on the heaters to give an additional voltage to the modulator for more [positive] peaks.  Bias supplied by the driver, of course.

Anyway, just thoughts  ;)

BC939 rollers work FB in the balanced^2 tuner. I used them at QRO for a few years then decided to use my 80 meter 1/2 wavelength doublet on 160 meters. There the coils were not happy. Sold the pair on ebay to some guitar player from Studio City.

Looking at the schematic. Lose that capacitor center tap back to ground. Not necessary. Split stator not necessary in this app either.

I have been real sick since then, not sure when I caught it but I think it's just a real bad cold and I might be getting better, maybe, so there's no way I'm going out in the cold building. Got the heat cranked to 78, and a pot of water boiling to keep the humidity up to 50% or so. Been in bed a lot.

suppressors in 3-500Z plates sounds good. There may be things I don't see on the scope. I can do that.

I want to change the bias supply to a low voltage variable unregulated bias and keep the front panel variac functional and original appearance.

Need to make a new supply something like 0-25V, a 2A bleed. The bias and screen supplies are all made on 19" panels that lie flat across 2 rails inside the TX. This should be simple and I can also keep the original around.

This way the layout can be preserved and a couple of high current low-L chokes can be used and make an LCLC filter similar to the existing higher-voltage one. Have not looked yet to see what is on shelves. A 2A bleed may be worthwhile considering the peak grid current allowable. The modulator bias meter is a 500V scale, but changing it is not mandatory, only changing the full scale needle deflection.

The original included driver transformer, a 40 watt CG-512, was made for tubes that use 2x the grid swing and 0.5x the grid current. It should be OK to keep using that because it's not near the power limit at all.

I realize it is shady to step up a 30W tube hi-fi amp's 4 Ohm output to 500 ohms using a power transformer then send it to a transmitter's 500 Ohm to modulator-grid-driver transformer, but there's already a plan for that and a backup.
I have a Norton amplifier with pp-par 838's driven by 6B4's and it has a 500 Ohm output, and there is also an Altec 1570b converted to a 1570x thanks to Bear sending me a 500 Ohm OPT for it. There is also a 500 Ohm output transformer or two around here if I want to build a speech amp from parts. The audio setup right now is temporary and was just for tests. There is an EQ and a couple of compressors, a Kahn phase rotator, and Altec mike mixer. There are a few audio bandpass filters, the typical +/-200 to +/-3500Hz types, as well saved from various equipments over the years. -so that all should help. I guess the audio is the second part of the overall thing.

I've read about "biasing on" a 3-500Z for class A SE audio purposes at lower range plate voltage in order to get the current up, example 1KV @ 400mA for SE class A. I'm not going to bias positive, I think. Keep it simpler.

Ok about the BC-939 rollers not being happy on QRO. They are big but not that big. The contact area is not the best on them either.

Thanks and happy new year! with I could be on the air, need to go get back in bed and get better.

Just to be clear Patrick the 939 coils worked just fine until I used them on a really short doublet. This is uncommon.

Feel better.......

The MFJ 989 tuner burnout finally got investigated. There is no real evidence that the balun got hot but it has not been completely removed and inspected and there is a slight stench. It is a 2:1 voltage ratio step-up balun. So, it makes 50 Ohms coming from the tuner into 200 Ohms for the dipole ladder line.

I bought two baluns from DX engineering. Toroid-type current baluns, not the ferrite bead ones.

measuring the antenna with a1:1 toroid core type 2-wire current balun shows me about 70-90 Ohms in most places of interest

The tuner sees 17-22 Ohms with its own balun. That is a burden of high current for it. 9.3A at the crest of the modulation cycle at legal limit. 4.1A at 500W.

measuring the antenna with a 1:2 toroid core type 2-wire current balun shows me about 15-25 Ohms in most places of interest.

This is similar to the tuner's balun result, but the measurement difference might be the voltage type vs current type + the individual antenna.

The larger stench is from the tuner's wattmeter/swr meter board. A 5-25pF trimmer capacitor caught on fire, and burning, fell loose from the soldering to the board. This might be repairable using a better cap. They used an inexpensive one. Also noticed was the tiny size of the coupler core and its very small center wire. That didn't burn though. I think it was peak RF voltage that blew up the meter circuit's trimmer cap. The tuner may not be able to tolerate high VSWR of 2:1 at its input at normal AM levels.

There was no real evidence of arcing or burning on the function switch, caps, or roller. It may be worth it to repair the tuner for use with lower power AM transmitters or sell it off to a SSB user.

Having looked at the MFJ coil and caps, it is easy to see the 939 military roller coils will be fine on a 1KW AM size tuner.