TMC GPT-10K band switching question

[Opcom]

No I don't have one, wish I did.. but I have a few important parts including the tapped PA tank coil which I want to use for a 3CX3000. The PA plate tuning coil unit incorporates the band switch somewhat like the way the B&W 851 or 852 plate tank units work, with a shaft down the center and a rotary switch at the hot end, the mechanical operation being carried out from a shaft at the cold end.

Because it is physically large, the band selector shaft has a worm and pinion gear. The worm gear drives the pinion to rotate the bandswitch shaft.

The issue with re-using this tank coil is that it is not a matter of a 360 degree rotation of a front panel knob. It takes many turns of the worm shaft to go through the bands. Additionally, the bandswitch has high-pressure fingers as contacts, and unless a serious crank is to be used, it is not something that anyone would want to turn by hand. Turning the bandswitch shaft directly and getting rid of the worm would mean using a medium size handwheel maybe 8 to 10 inches diameter, also be hard to turn, as well as eat up a lot of panel space. This is really not a good option and it seems better to try to do it as the TMC equipment was intended.

The manual is not too clear about how the indexing is done. There appears to be no direct knob for this, but instead it looks like a motor in he pictures that drives the worm. That is not too much of an issue, but there is no indexing switch on the bandswitch to indicate what band it is on. Putting one on the worm drive does not seem to make sense because it turns so many times.

How did TMC set this up? There must have been a way for the TX to tell which band it was on and then use the motor to go to the band for use. The manuals go into detail about the electronics but not this little tidbit of information.

[Opcom]

The motor might be this B900 strangely shown in one version of the manual but not another, or maybe somewhere else. but the shafts can be seen going to the front of the unit.

Just trying to understand how the worm-driven bandswitch position is known or aligned.

edit - S901 senses a detent on the band switch in each position so that it is not in-between, although I don't see this sw anywhere but on a schematic. I guess I'll keep searching to figure out the way the band switch is moved.  B900/901 is an air circulating blower.

[Opcom]

I stand corrected. There are a lot of manuals on this, so far about 12 of them. The PA band is switched by a small knob on the front panel. OK well it must take a gorilla to turn it. Would be nice to have pictures of this part of the unit as I swore there was a motor to turn that shaft and there is an odd bracket-looking deal in there on some of the manuals' RF deck pictures - see pa_01.png

The numbers near the tuning elements seem to correspond to the 100 series numbers on the front panel picture. that is, 117 on the pic relates to 17 in the schematic.

[W3RSW]

1. Old way:
I would simply drill about five small holes (not to big or you'll ruin the integrity) into the bottom facing edge of the sector gear at the same radial spacing as the switch contacts. Thread these holes and insert a small screw sticking out to contact a switch to ground, mounted on the chassis underside. The switch contacts go to a light on the panel signaling 'make.  Then you or a chip make the count.  The switch could simply be a light homemade slightly springy leaf contact.

Better logic but more switches, just one hole and screw but five switches with five panel lamps indicating which band your on.  Find five switch locations that don't interfere with the four coil mounting spokes.

2) modern way:
Or go optically with five small mirrors and photocells.  Or opaque optical interrupters, whatever.  Use TTL logic of choice and LED drivers if you want to get fancy.

Turn the worm with any crank/knob worthy of much smaller torque but with many turns.

[K6IC]

Patrick,

There probably was a motor driving this switch.   Since this worm drive is very solid,  the determination of which contact  had been selected could have been done with a small gear setup at the motor end of the shaft.

These manuals are very detailed,  and this can make it very hard to find the correct parts of the manual to find things ...  sometimes many pages need to be referenced,  all at the same time ...   you know this ...   can be a bit frustrating and time-consuming!

The panel selector may just be a switch that activates the motor/control system.

A very studly inductor and switch.

There is one member here who has a 10K,  that is complete (+/--),  and probably can speak to way band changes occur.

I will try to PM him ...  he works in a different state from the main residence,  and may not check in here often in the Summer.

Neat project.   I,  too,  have one of these inductor/switch assemblies,  but it lacks the chassis and gear,  so am following your progress at a distance.

73  GL,   Vic

[Opcom]

Rick, the optical method seems like something I could try, and would only need one hole in the pinion or sector gear. It could as well be mechanical switches but space may be an issue even as large as that box is.

I should look at it again to discover where this 'detent' switch should be -look at the shaft itself. I don't see any extra holes in the pictures. There is also always the real low tech way of putting a plastic ring on the shaft at the top, just above the switch, with numbers on it for bands, and viewing it through a window.

because of the existing holes in the gear, it might not be possible to mount pins on the gear in all the right places and I sure hate to risk that drive. The gear may be some rather soft metal. I have to take this apart at least partially and try to clean and lubricate it.

Maybe the tiniest hole like a 4-40 or even some epoxy could mount up a single mirror to reflect at each of the nine positions, although that is a 240 degree arc, about. So the chip could count and be reset when there is a delay or missing pulse while the motor is turning.

Some positions are across from each other. One way to prevent the light from one lamp triggering the sensor across from it is to incorporate the lamp and sensor pair into an oscillator and having the sensor see the mirror, and the reflection of its own lamp, through a lens focused on the place where the mirror will be.

This sounds complicated but it is not, only tedious. The lamp and photocell are mounted in little opaque tubes attached side by side like a double barrel shotgun.The lamp runs off a class A amplifier's quiescent current, and the photocell is on the input, so that when the light hits the cell it closes the loop and oscillates. The AC voltage can be detected a low impedance, bypassed circuit. A refinement was shown to me in an article using an op-amp as the oscillator but I can't find it, only the simpler circuit.



Vic, you are right about the manuals. The sections I found are all the electronics but no mechanical stuff. I am guessing there is a frame section that would shoe the exploded views and the motors and switches. I have the interconnections diagrams but that's it, they seem not to be concerned with that, just the radio part.

It also looks like in the pictures, which are not very good, that there is some kind of unusual mount or thing at the front panel that seems more than just a turns counter. It seems too deep behind the panel to just be a turns counter and there is also a closed box area. The idea of a motor with its own gears and a switch arranged so that it opens the motor contacts when the big switch is in the correct position, would be perfect. It makes a lot of sense that the knob could be moved to the right position and the mechanism start running until it gets there.

Thanks for PM the guy with the set, maybe he could take a look or pictures, or has the correct part of the manual when it is convenient. There is no rush on this project, it has been going on a long time already.

On your coil, you do not have the chassis, but I am guessing you have the T-shaped lower bracket so the thing is held together?

[Opcom]

Been through http://www.tmchistory.org/tmc_manuals/tmc_commercial_manual_page.htm again to check and see if I missed something. There is a ton of great stuff there but unfortunately like so many of the accessible manuals for military/industrial stuff some are are incomplete. Maybe they were not always complete and have several volumes that are uncommon, I don't know how TMC supplied the schematics. If they are large sheets it might not be simple to scan and it is a lot of work to scan them in pieces then paste together in paint or something. I'll keep looking, there has got to be some mention of the band switch mechanism somewhere.

[KD6VXI]

If you still have the Harris RF-103 check how they do it.   A single pot is ganged to the roller inductor.   It's a 10 turn pot.   It uses a comparator to reference the voltage from the pot to a predetermined setting of the roller (set by precision board mounted 10 turn pots).

This allows tuning to be done either remotely,  by watching a meter for preset positions or by watching a wattmeter and tuning to max qro.

I hated it troubleshooting it,  but after I fixed it I gained respect for it.

--Shane
KD6VXI

[Opcom]

I don't have it any more, but that's really not too big an issue. A pot is fine for human reading the position, but on this inductor there does not seem to be a place to add anything very conveniently with a 1:1 turn. A plastic toothed, thin cog could possibly be put on the shaft just below the coil, and a rubber belt used. The shaft is very large diameter though, it would not be a standard cog, possibly a modded one. Just another idea  but stuff like this is where the hams with the lathes and mills have an advantage.

I've worked a lot on the electromaniacal/electronic servo stuff. The Collins 618S-1 is an example. It positions the crystal selector and band rotary switches by means of the kind of switches where all but one contacts is normally closed.  They are wired together in series with the motor control relay coil. When the switch on the knob is turned, the relay energizes the motor and it turns the other switch until the open position is found. Those switches are not very common any more but it ought be possible to logically invert the state from a regular switch such as 1P12T rotary type appropriate for this tank coil.

[Opcom]

Here's my explanation of the positioning, from some work on the 618S-1. I imagine the GPT-10K to use something similar.

[K6IC]

Patrick,   seems very plausible to me.   Good sleuthing ...   Thanks.
Big parts are fun!     73   Vic

[KL7OF]

The GPT 10-K has a tapped inductor very similar to the one you have pictured Pat....It is manually operated by a BIG knob on the front panel and a set of straight cut gears and has positions for different band segments...The transmitter has continuous coverage from 1.5 to 29 Mhz...

[K6IC]

...  Interesting that the pa 03 schematic that you posted,  Pat,   shows that the lowest frequency for the plate tank is 4 Mhz.

There must have been a number of variants to this basic rig.    FWIW.
Thanks Steve for the added info.    Vic