Relay problems

[Opcom]

I have this old transmitter, and am looking for advice, since I am going to have to do something about it.

It's a '57 homebrew called the "Col. Tucker Transmitter", and it is not really a complete transmitter, but a 4-1000A power amp and a 3-500Z class B plate modulator. It has to be driven by RF and audio.

The Colonel Tucker has 5 relays to key it.. No doubt the cause of my present woes.
-One to turn 120VAC to the PA screen supply on and off.
-One to turn the 120VAC to the PA plate supply on and off.
-One to bypass the "conical heater element" in series with the 120VAC to the PA plate supply.
-One to turn the 120VAC to the modulator plate supply on and off.
-One to bypass the "conical heater element" in series with the 120VAC to the modulator plate supply.

The heater elements are cut in when the "power reduction" switch is activated. I am not using this feature yet. At first I thought these were for inrush limiting, but it's not wired that way.

Ok, so when you supply 115VAC (via your KEY if using CW!!) to a terminal strip, these things all engage at the same time, amidst a bunch of blue-green flashes. This is apparently the way Col. Tucker himself used the rig, as well as locking the key down for AM, unless I am missing something or there was a sequencer or controller missing when I got it.

All the relays are toast. In fact, one has a pair of contacts welded together, which has caused all kinds of problems with the modulator plate supply's pair of 673 rectifiers flashing over. The PA screen relay -something happened to it and the armature spring of all things has been vaporized.

So, today all but the screen relay got replaced and I'm doing some tests. Each time the unit is keyed, there is a 8Uf-4H-32uF filter to charge to 3KV. I can see the green flashes in my new 250VAC/30A (1HP@125V) relays. I tried several things: a capacitor across the contacts, and MOV across the contacts, etc. No difference.

For one, these new relays may not last very long. The other thing, this has got to be the totally wrong way to key the transmitter. Should not the high voltage be on all the time to avoid these huge surges?

Assuming I can run it with the HV on and add a couple of tubes and electronically put this thing into the "transmit mode" via the PA's screen voltage and the modulator bias, would that be a good idea? anyone done it that way?

to be specific about my question:
0. transmitter warmed up, 3KV sitting on the plates.
1. put in xmit mode (and TR relay switch to TX)
2. "key" the exciter and speech amp
3. blab for a while.
4. "un-key" exciter and speech amp disabled
5. take out of xmit mode (and TR relay to RX).
6. listen to various old buzzards blab for a while.
7. GOTO step 0.

Any opinions of this? I want to get the beast back on the air. I'm willing to do some more mods, but I'd like to pick a good method the first time.

[WQ9E]

Hi Opcom,

First, to keep you amongst the living I would add a real switch to key the relays!  Next, you have a perfect application for the inrush current limiters (available from Mouser, Digikey, and others).  These have a fairly high cold resistance dropping quickly to a low resistance and are great for reducing inrush current and the resulting effects you are seeing.  Use properly rated ones to carry the current of each of the individual supplies.  Last but not least, the way things are set up if your plate supply relay fails you are going to cook the screen in the final unless it already has overcurrent protection.  I would suggest taking the power for the screen relay coil from the circuit feeding the plate transformer (you may have to switch screen relays or use a stepdown transformer depending upon the transformer primary voltage).  This way the screen supply will not be activated unless the plate supply is also on.

How did the Colonel meet his end?  Perhaps from the wiring of this rig???

73, Rodger WQ9E

[Steve - WB3HUZ]

I didn't know 3-500s were around in 1957. Cool rig though.

Don, K4KYV wrote an article for ER on a TX/RX sequencer he built. You may want to contact him.

[W2XR]

Hi Opcom,

 I would suggest taking the power for the screen relay coil from the circuit feeding the plate transformer (you may have to switch screen relays or use a stepdown transformer depending upon the transformer primary voltage).  This way the screen supply will not be activated unless the plate supply is also on.


73, Rodger WQ9E

Good idea.

I interlock the screen supply enable relay with the HV/plate supply contactor in my rig. I use a spare set of contacts on the contactor for this purpose. This way, if the plate contactor drops out due to a plate supply overload, etc., the screen voltage is simultaneously and immediately removed from the 4-400As as well.

In a rig of this power level, I would definately soft-start the separate HV supplies for the RF PA and the modulator. I don't think I would use one of those inrush limiting devices available from Mouser, etc., for this purpose, but rather a current limiting resistor and time delay relay. In the plate supply for my rig, where the RF PA and modulator share a single common supply, I use a 20 ohm 200W resistor and a 300 ms. time delay. This is in series with one leg of the 230 VAC feed to the plate transformer primary. With the soft-start, there is never any nuisance tripping of the plate overload breaker, or dimming of the house lights when the rig goes into the transmit mode, plus it is gentle on the silicon rectifier stacks.

You may also want to consider a time-delay for the T/R relay. High power transmitters can store a lot of energy after the plate supply is shut down, when going to the receive mode. I use a 700 ms delay before the T/R relay contacts transfer, after the plate contactor contacts transfer open. Otherwise, you run the risk of burning the T/R relay contacts in fairly short order.

Good luck with the new rig. Sounds like a nice transmitter.

73,

Bruce

[Opcom]

Ok, about the key, I have attached a 250V 3a SPST switch.

As far as the Colonel, I don't think the 120V or HV got him. He died in old age, however, right in front of him at the operating position, was the quite large diameter military surplus antenna beam direction indicator, chock full of thick radium paint, and showing about 5mRad/Hr through the glass (gamma) to this day. If he sat in front of that hot dial admiring it's glow along with the glow from his radios, 3 hours a night a couple 3 times a week.. who knows what it may have done. (image.. It is the large clock-like item above the receiver)

As far as the wiring, 90% of it is done in #12 and #14 single strand house wiring, with black insulation. The other color of wire, for the little stuff like the light bulbs and such, is blue, about #18 or 20. Before I got it, even the 3KVDC wire was the black house wire, albeit it was kept well away from everything else, like bus bar would be.

I don't know when 3-500Z's were first made. The rig originally had 304TH or 304TL tubes for modulation. After being cheated by liars on eBay and at crafty old bubbas at hamfests 5 times, I decided it would be less costly to put in a pair of $130 Chinese 3-500Z's, even with the cost of the blower, etc, than to keep buying desk ornaments from  crooks.

The rig has a screen overload relay of the latching type, as well as a plate overload relay of similar ilk. If the plate overload goes, the primary of the screen supply's HV transformer is disconnected. It's not possioble to fault the unit in such a way as to cause the screen voltage to be present without plate voltage, unless one has it operating and then cranks down to zero the plate control variac, but the screen grid overload will trip then. Finally, the screen relay's coil is in parallel with the plate relay's coil. I could move it to the switched plate primary volts, but the 120VAC to the plate there at the PA plate relay is actually from a variac, so it's possibly 60V at times, or 140v if I want to crank it.

I'll look into these inrush limiters, and see what they have that would be suitable. I like the idea of a resistor and a time delay, either independent timing or controlled by the voltage after the resistor as shown in ARRL handbooks, but I prefer DC because an AC relay always buzzes in these applications. I picked up four 10-Ohm 50W resistors today, might do parallel or series, depends on the drop and the load. Power consumption with the filaments, bias/screen filaments, bias HV, and 3KV HV on with bleeders, but no final or modulator current, is about 5A per 120v leg.

But, just either 3KV HV supply itself consumes only 300 watts idling at bleeder load.  -so this is all I have to supply through any inrush resistors, and then switch them out. I'll also go look at the Mouser products.

I just want to avoid burning the new relays. Also picked up relay #5, the screen keying relay. might as well do it also.

both the HV decay to zero in about 4 seconds via the bleeders. it's bled to just past the point of Lcrit on the chokes, about 100mA. That's one of my mods, it had no bleeder before except the HV metering string.

Bruce, Please explain, about burning the TR switch contacts. Do you mean with RF? If that is the case, I have to also make sure the RF would be gone (due to external drive keying) before that connection is broken.

[W2XR]

Bruce, Please explain, about burning the TR switch contacts. Do you mean with RF? If that is the case, I have to also make sure the RF would be gone (due to external drive keying) before that connection is broken.


[/quote]


Hello Opcom,

(Sorry, I don't know your name!)

Great photo and shack! Very 1950s, including the cowboy-garbed JNs. Come to think of it, my parents used to dress me like that too, when I was a young JN in the late '50s.

With regard to the delay sequencing of the T/R relay; yes, RF. That should be all you are switching through the T/R relay, unless your rig does not employ a DC plate blocking capacitor, which would be representative of very poor design practice. You never want to hot-switch any T/R relay in the Dow-Key class, etc., at the power level your transmitter seems capable of.

You want to give a lot of thought as to the correct turn-on/transmit and turn-off/going-to-receive sequencing of any high-power radio transmitter. In other parlance, this is known as "controlling the fire". In the case of going from transmit to receive, you want to sequentially:

1) Turn-off the plate supply and screen supply simultaneously

2) Remove the RF drive from the RF power/final amplifier

3) Transfer the T/R relay contacts

4) Un-mute the receiver once the T/R relay contacts have fully transferred

I use interlocked digital time delay relays in my rig to control the sequencing of these functions. These relays allow very precise setting (to 1 millisecond) and repeatability of the delay or turn-on time, and new-in-box examples are posted on e-Bay frequently, and at very reasonable prices. Mine are made by a company named NCC (National Controls Corp.). They also exhibit excellent RFI immunity, although I do bypass the input to each relay with a pair of .001 uf caps to ground. You can view all of the information on the various NCC product offerings by going to their website.

I hope that this helps.

73,

Bruce

[Todd, KA1KAQ]

I recall when PJ got the Tucker rig some years back, and yes - it's a very stout, well-made rig by the looks. A piece of work (and history) like that deserves to be set up and handled properly. There is little doubt in my mind that the lack of a good sequencing set up was a factor in my zorched mod iron, combined with running fast break in.

Excellent explanation, Bruce. Thank you for sharing it.

[w3jn]

Often you can find these at hamfests for little/nothing.  Last hamfest I went to, someone had a box of a dozen or so brand new ones for $5.  Easy to build up a sequencer as described by Bruce. 
http://en.wikipedia.org/wiki/Solid_state_relay

I use a NTC thermister as an inrush current limiter in my GPT-750.  The disadvantage of these is that if you're doing break-in, there's often not enough time for them to cool down so as to effectively absorb the inrush surge.  The advantages are they are cheap and effective if you're, say, talking to Don KYV or The Tron.

[Opcom]

I added my name and so forth and updated some other old info.. Sorry about that, I lurked for long time.

So, that picture is Mr. Tucker some 40 years ago, with his sons. In my day (decade after), we also played at the cowboy games, but for me, more often to be seen in a military getup with a wooden M1 carbine or a plastic .45 squirt pistol, since my father was a KC-135 tanker commander in the air force and we lived on the SAC base in Omaha. (Military brat, base housing, Catholic school -- we got away with NOTHING! -and I'm thankful for it now, yes sir.)

I met Durward Tucker's son, George, a few years ago when he came to Dallas to see to their old house. The desk, square clock, and ceiling lamp were still there, as well as the radioactive tower direction indicator. He now has the station components as well as COL Tucker's call sign "W5VU". He brought the "binder" to me and let me scan it all. He is now living in Woodstock, NY. He was surprised to see the big rig still existed after all this time. Being a gentleman, I offered it to him, but he has no place to keep it, so it's mine till someone else inherits it. I hope to have it on the air before then!!

Back to the technical discussion, I like the idea of a timed and controlled sequence of events. With today's quite stable plastic capacitors in the tens of uF range and the possibility of operating common 120VAC relays on DC, it may be possible to use RC constants and avoid having solid state devices inside the rig. Have to think about that. Time to think about how much coil DC current a small and common 120VAC relay would need -same amount as on AC, but lower voltage. Do it with RC constants alone? I have alot of 6AU5's and 6AV5's on hand and a 6AU4 or 6AX4 could be a good rectifier for 120V. Do it with tubes? Solid state inside rig=no good. solid state outside=good.

Disregarding the view of the entire station for now, about keying just the Tucker itself, suppose the rig was in the state where the TR relay was set to XMIT, and the HV was ON.
-if I were to create a state where the PA screen voltage was connected to GND, this state I would call the "hot standby" condition.
-next, to bring it from hot standby to operate, I would propose to apply RF drive and switch the screen voltage from 0 to 500V. The rig would then be in the carrier condition. I believe this would satisfy the requirements of reasonably fast keying (for AM), without having to thump the high voltage on and off each time. It should not harm anything to apply drive with the screen voltage at zero for a fraction of a second. Driving power is 12 watts and the grid dissipation is 25 watts, by the book.

Biasing off the modulator to maybe -200V during hot standby would also protect the mod iron from unloaded operation.

The way the rig is built, it would take two relays, one for screen control, and one for modulator bias. They would connect to the wiring from the front panel VARIACs and max out the mod. bias and kill the PA screen voltage during hot standby, then let 'er rip for "operate". Maybe I can draw a schematic of this including the 'sequencer' with tubes. Is this nuts or just a period way to do it?


So, I am going to digress now from the rig's actual keying possibilities, and tell about the rest of the rig's circuits.

I am proposing this approach because I have a laboratory PLL type signal generator and am putting together a low distortion solid state broadband power amp for the purpose of driving the Tucker. I could also use the Kenwood TS-430 for a variable RF source. Either can be made to provide a keying contact closure to signal the Tucker that it's time to operate.

I am going to have to pull the RF deck (once the power and keying is sorted out), and re-do the grid tuning circuit. It sometimes behaves in an unexpected manner, I suspect old damage. It is also bandswitched, yet the plate circuit is continuously tuned. I find this incongruity of function slightly annoying. I have never had this apart, due to about 120 screws. I may go with a swamped design using a 50 or 70 ohm non inductive resistor in parallel with the primary of a 1:6 (voltage ratio) broadband transformer, or, if I am lucky, the National MB-150 might fit in there and I can keep some kind of grid tuning. I also have to fix the plate circuit for parasitics (discussed in another forum).

thanks,

Patrick J.

[W2XR]

Hi Patrick,

With regard to your statement;

"Back to the technical discussion, I like the idea of a timed and controlled sequence of events. With today's quite stable plastic capacitors in the tens of uF range and the possibility of operating common 120VAC relays on DC, it may be possible to use RC constants and avoid having solid state devices inside the rig. Have to think about that. Time to think about how much coil DC current a small and common 120VAC relay would need -same amount as on AC, but lower voltage. Do it with RC constants alone? I have alot of 6AU5's and 6AV5's on hand and a 6AU4 or 6AX4 could be a good rectifier for 120V. Do it with tubes? Solid state inside rig=no good. solid state outside=good."

There is no need to have this aversion to solid-state anything within the rig. I am for whatever technology or design approach does the job best, be it vacuum tube or semiconductor.  I suspect that most of my AM brethren here on the board subscribe to the same philosophy. The old engineering adage, "Form Follows Function", is as true today as ever. For example, I use an all solid-state Icom 718 as the RF driver to my homebrew 2x 4-400A plate modulated rig, as I can continuously adjust the RF output, it is as frequency and amplitude stable as a rock in the CW mode, and the frequency display is extremely accurate. I also use all solid state rectification in this rig, with the exception of the 2x 866As for the +1200 VDC supply to the push-pull class A 845 audio driver.

In my opinion, the concept of using vacuum tube rectifiers, as you have described, is a waste in this particular timing/control application. Obviously less reliable, they will need a filament supply, etc. Look for the digital time delay relays John/W3JN, and I described in our earlier posts, and you will have no issues. They operate directly from 120 VAC, and the ones I use from NCC include built-in voltage transient protection, which really is a necessity with all of the large power supply inductances and related back-EMF present in a large radio transmitter environment. Why reinvent the wheel? They are cheap, available, and reliable, and are a superior solution for this application.

Using an RC time constant circuit to provide the required time delay will work, and indeed this approach is used in most of the lower-cost continuously-variable time delay relays available. But the difficulty in obtaining and maintaining very accurate delay times on the order of several hundred milliseconds, as well as the repeatability of the delay time, makes these relays less than desirable for these reasons. They are widely used in the less-critical process control/industrial timing applications. The more critical timing applications, such as transmitter control and sequencing as described above, will benefit from the time delay relays of the digital design.

Just my 2 cents worth. Whatever you use, have fun and experiment! When you're flying the soldering iron and ratchet wrench yourself, you can utilize whatever you consider to be the best engineering solution in your rig.

73,

Bruce

[Opcom]

After the discussion, I'm reminded of how finicky and imprecise the old tube-type industrial timing controls could be. Thyratrons.. Regulated power supplies required, for any accuracy at all. could be a big mess.

Maybe I will have to look at solid state items. I am aware that one 4x5 circuit board, which I could make in a week, could control every relay in the entire transmitter off of a 12DC/10mA input signal for the purpose of keying it. A soft-start could help save the HV control relays. Thanks y'all for the good advice!

[W2XR]

Hi Patrick,

You are very welcome.

Now get working on that really nice Colonel Tucker transmitter and let's hear it on the air!

Good luck with this nice project!

73,

Bruce

[Steve - WB3HUZ]

This may provide the proper timing/control. You'll just need to have it drive some big relays for the portions.

http://www.advancedreceiver.com/page20.html

[Opcom]

Their logo reminds me of Amplifier Resarch corp.  The board is a good value for the functions it does. Considering the value of one's time, it would be cheaper to buy it than make one.

Last weekend I shoved some RF through the rig and was getting about 400W into a dummy load. I lost one of the plate blocking caps after about 10 minutes. This reminds me I need to add some circuit breakers on the primary as I blew one of the screw-in fuses. Upon inspection the 50-year old caps have little cracks. Probably should replace them all. There are five of the 500pF 20KV cylindrical doorknob caps - from old TV sets I suppose.

I have some caps made by "plastic capacitors". One is an "OF200-103", a glass type with oil inside rated .01uF/20KV. Not sure if it is going to handle RF, their page says it's a filter cap. The rig's lowest frequency is designed for the 80M band. I'll have to do some digging around maybe for a more suitable kind of cap.