I think I've fixed all the problems with why the fan stayed on too long, mostly down to the switch being in the top not bottome hose and the system being scaled up. So I moved the sensor and when that didn't fix it fully, (though I didn't want to do it this way) I descalled with a washing-up tab - actually 2. I put one in and ran for an hour and flushed, which made things a bit better but not enough. So I used another and drove round for a couple - three hours, cycling the temperature a couple of times. That made a big difference. I also fitted a 75 degree thermostat instead of the 82 degree one; though I think I'll switch back for the winter.
I'd checked the voltage stabilizer, which was wrong, at about 11 volts, but seemed stable. And as that does not affect when the fan switched on or off, I had left it alone, so I could see the effects of all the other changes all in the same context. But at some point in the process, the temperature and fuel readings started to get erratic: after reseating the top fuse once, I gained nearly half a tank and the temperature went to full scale before the fan kicked in - luckily I still have the rad instrumented, so I knew I wasn't really overheating. But clearly the sabalizer had gone. I used the kino-restorative technique and got it back to working poorly, but consistantly - the reading was around 2/3rds to 3/4 full hot with the fan on - but thought now was the time to fix it.
I wondered, it being mechanical, if I might be able to fix it. But not actually knowing how it worked, I got a new one first. That appears to be a solid state one - I can see the base is actually a small circuit board. I suspect its an L7810 in a TO-220 plastic can.
I managed to get the cover off the old one without buggering it too much and what I found inside is shown in the attached photo. The interesting thing is that there's clearly an adjustment. So, if you have one that just mis-reads, but is otherwise okay, it may be possible to re-calibrate it to 10v or so, rather than buying a replacement. Well it might work.
I think, looking at how it works, part of the problems I had was a poor earth, which connects at the sliver tab (front right) to the can, then through the mounting screws to the bulkhead. As it's the current through the wire from the output to earth that heats the bi-metal strip, and reduces contact pressure on the points, if that earth is poor, the strip won't heat up enough at the right voltage, and the output will be high. If it loses all earth, it obviously has to output battery voltage - unless the points are corroded and there's nothing to earth anyway.
I did try putting it back, after cleaning all the earth contacts, but I had played with it a bit - to see how the contact resistance changed as I bent the strip (not very well) - and it didn't work properly any more – while the gauges were stable, the measured voltage was all over the place. I suspect it was oscillating at a frequency that aliased with the sample rate of the DVM - it jumped about as if possessed; I think that it was p155 (to paraphrase Mike Absalom's Saga of the Ancient Britton).
So I put the new one in, and that's brought the reading down to just below to just over half driving round town in the 30 zone. I want to take it down the M55 for a bit and check its all okay now.
Still, I though the picture of the internals of the old Smith's stabilizer, and that there is a possible adjustment might be worth a note.
Graham
Attachments: |
DSCF3949.JPG [ 135.32 KiB | Viewed 1260 times ]
|
_________________ The 16v Slant 4 engine is more fun than the 3.5 V8, because you mostly drive it on the upslope of the torque curve.
Factory 1977 TR7 Sprint FHC VVC 697S (Now all of, but still needs putting together)
B&Y 73 Dolomite Sprint UVB 274M (kids!)
1970 Maroon 13/60 Herald Convertable (wife's fun car).
|