Fixing a Corsair’s Digital Readout

 

The Corsair II developed two problems simultaneously. First, the radio went “dead” in terms of RX and TX, except for a little noise. Secondly, the digital readout blanked (to 9.0000 Mhz), and stayed there. I’ve worked on Tritons and Omnis, but up to this point I’ve had limited experience with the Corsair II. Deciphering the fix that was needed in order to get the Corsair II back on the air put me on more intimate terms with this fine little radio from the year 1983.

If the display had been stuck, but the radio had been otherwise OK, I may have taken longer to get to the answer. As it was, I immediately targeted the PTO/Osc/Mixer lineup for a first stab. The aim was good, as it turned out to be the Osc/Mixer board (80975) that was at fault. For those who’ve owned Ten Tecs, the workings of them are pretty well trodden information. But for the passerby who might think that an old-but-not-yet-antique piece of gear might be an attractive addition to the radio desk, I’ll explain the backstory. While some digital radios are intricately wound into the display and control mechanisms, and need them in order to work at all, the Ten Tecs (at least the early ones) were still all-analog radios, with the digital part added as a more convenient feature for frequency readout. This is the case with the Corsair series, which work perfectly fine without the digital display, as the radio also sports an analog tuning knob with ticks for frequency on the front panel, just like the earlier radios like the Omni A,B, and C, as well as the Triton and Argosy, etc.  “Perfectly fine” may not be perfect for everyone, because one would need to count the number of complete knob rotations to know which part of the band he was in (hi hi).

The display can stop operating due to the display drivers and such, but after my investigation and eventual fix of my own Corsair II,  I’m convinced that some of the display problems seen in Corsairs are really in the Osc/Mixer lineup, and not in the display itself.  There is a RG174 coaxial run into the counter box, which is center/front/top of the Corsair enclosure.  It enters the left hand side of the counter box.  The other end of the coax connects directly to the J-36 connector on the Osc/Mixer board 80975, traversing a thru-hole in the divider that separates the top and bottom circuits in the enclosure, since the Osc/Mixer board is in the bottom half of the Corsair. 

What I’ve always loved about Ten Tec is the way that every major block-diagram part has its own board.  So, working on a Ten Tec is just working on a block diagram in silicon.  Right off the bat, the signal coming into the display/counter box was seen to be too low to switch the input gates (about 93 mV).  The J36 connector on the 80975 Osc/Mixer board also showed me about 93 or 100 mV peak-to-peak on the scope.  The frequency was correct (5.xx MHz when using twenty meters), but the amplitude was too low.

Switching the band-switch allowed the correct frequency to appear on J36 and also on J61 (J61 feeds the actual radio mixer stage, instead of the display box) – for each band, but in all cases there was not nearly enough amplitude.  To me, this screamed “Mixer/HFO board output transistor failure.”  Indeed, a quick look at the bias voltages indicated that the last two transistors on the board were in the tank, and internally damaged.  So, I ordered a new MPS3693 and a 2N3866, and waited a few days for the parts to arrive.

The output of the Osc/mixer board really is two separate outputs. The MPS-3693 feeds the counter box thru J36, while the 2N3866 feeds the whole rest of the radio thru J61. Both outputs were low. When I went to unsolder the transistors for replacement, both of them pulled off easily without heat applied.  They had gotten very hot indeed!  They had MELTED their own solder leads! By my reckoning, they broke down, and subsequent to that generated even more heat, until the pads separated from the leads!

There was another problem.  The Osc/Mixer board cannot be removed without removing the entire band-switch assembly.  Oh nightmare of nightmares!  I have worked on Ten Tecs for a while, and usually don’t run into this kind of design-time miscalculation by them.  There was no way to get to the back of the board in order to resolder the new components, except by total dissassembly.  So, I decided to top-tack the replacements onto the board, point-to-point style.  Ugly, but a whole lot less work. Below is shown the empty space after removal, an then the ugly point-to-point result:

Figure: Empty space on board, where defective transistors were removed

Figure: The transistors were “top-tacked” onto the leads of connected associated parts (i.e., resistors etc)

So, the rig is back in operation, and I have a little more appreciation for what holds it together. I have a few comments about all of this. I love Ten Tec gear, and it is the darling of DIY fixers, no doubt. It’s cleanly built stuff, block-diagram board-built, and usually seems to have great design decisions. But, the replaced 2N3866 transistor still runs quite warm. I haven’t done the calculation to show it, but I think it may have been picked too close to its dissipation limit. The device is spec’d at 350 mW, and I’m guessing it’s grabbing every bit of that. The dissipation is coming from somewhere, albeit I can’t see under the board, where there might be lurking a cause of which I’m currently unaware.  – Just a note for the next folks to build a Ten Tec Corsair (wouldn’t that be nice?!)

It’s likely the 2N3866 failure cascaded to the MPS-3693 failure, as that happens.  So, both the radio and the display got zeroed. The fix works, but is ugly. So, I’m thinking I could cut a little daughter board to hold the fixed transistors, and make it a proper plug-in fix rather than a hack.  It would be easier to mount heat-sinks for the devices if they were properly affixed to a daughter board.  Waddya think? Here’s the MPS-3693 part, and the 2N3866 is only a little more, but I haven’t drawn that in my Gimp sketch, and am too lazy to switch to the electronic-CAD:

Figure: The daughterboard for a cleaner fix would be simple

The MPS-3693 that drives the counter box also seems to run very warm, but it’s not as seriously hot as the 2N3866. I could see the possibility of a failure of either or both devices over a long period of time.  A failure of the MPS3693 would cause the display to continuously readout the IF “9.0000”, or “91.000” – yet allow the radio to continue to work in analog-only mode. This may point to easy fixes for display-stuck Corsairs in certain cases, since many Corsairs on the used market are sold that way. If I were an adventurous DIYer, I might give some of them a shot. There’d be no guarantees, of course, since the display components themselves CAN go bad.  A bad display section might show nothing on the readout display.  So, this rambling adventure is not meant to be advice for anyone else, as I could be entirely full of it, hi hi. 73’s.

Figure: The Corsair before the fix, showing a “stuck” display

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.