I've already looked at the stator mount, several times. In fact, most recently I looked at it (while replacing the clutch pushrod) soon after seeing your idea to use heatsink goo in that joint. It is perfectly obvious that heat flows from the stator windings to ambient mainly through that cover.
Hmmm. In my service manual you don't have to remove the side case to replace the clutch pushrod or seal.
Nobody is contesting the idea that reducing stator winding temperature would increase its longevity.
Agreed. But I would add, reducing the temp of the stator core implies lower winding temps.
The fundamental problem is that Suzuki (or the magneto designer they used) has elected to run that #18 AWG wire at a current such that there is only about 135 circular mils per Amp. This is less than half the area at the bottom end of the range used in most applications of magnet wire.
I am sure they had their reasons. But you keep wanting to cast this as an electrical engineering design exercise and it is not. We don't get to change anything regarding the design of the stator wiring, B & E flux, magnet strengths, coating materials, ad infinitum. Its a done deal for us riders so we can only modify the system as is. Perhaps if you designed this generator there would be no failures but you didn't work for Suzuki at the time.
The cover's temperature is, indeed, a proxy for the temperature of what is thermally connected to it.
The problem I tried (unsuccessfully) to bring to light is revealed when one attempts to decide how good a proxy it is for stator temperature. Obviously, those temperatures are related, and if that relation stayed constant the cover temperature could be used to assess improvement (reduction) in stator heating.
Not true. The relation doesn't have to be a constant they just have to be correlated which they are certainly are.
But the whole point of your heatsink goo application is to change the relation between the two temperatures, to bring them closer together by improving thermal conductivity.
Sort of true but not my point, that would be a side effect, and the way you state it you are reversing cause and effect. The stator is primarily cooled by conduction to oil and to the side case (we can ignore radiative cooling). For any given operating point, if I improve the conduction of heat into the side case (all else held the same) the temp of the stator must go down.
You brought out this cover temperature measurement in the context of applying heatsink goo to the stator-cover joint, implying some relevance. But all that measurement can show in your case is what is already perfectly obvious, that heat flows from the stator to the cover.(Now, if a series R/R was installed, that temperature would become interesting.
You can't have it both ways, claiming it has relevance to prove the series R/R is lowering stator temperature but has no significance for proving that stator must run cooler if the side case extracts more heat (i.e. gets hotter). Proving the series reg lowers temp is a simple measurement (its been done, the case is closed), proving increased heat flow in my plan is complicated (see below) but it is just as real and just as valid of an approach.
More interesting would be the temperature difference from the center to outside edge of the cover, as that would be a decent measure for the amount of heat flow.
The IR photo shows exactly that. There is a color-code-to-temp conversion chart to the right of the graph. The photo also shows the crankcase bolt cover is significantly hotter than the rest of the side case. You haven't explained how that is possible.
The reason I mentioned thermal resistance, of the stator-to-cover connection in particular, is that it is relevant when considering the meaning of the temperature measurements at the cover. That, too, is basic physics. What do you see as the likely effect on cover temperature of adding your heatsink goo? And, with your "basic physics" firmly in mind, why?
I've stated my argument on this multiple times in my original thread and in my replies. It is the primary context of this discussion and a context you keep dropping to discuss electrical engineering design questions. If you still don't get it reread it if you must, I'm not going spoon feed you.
What you said was, "it is unlikely I will be able to prove it definitively with data." That does not concede the point at all, and the fact that you've taken to attempting to educate me away from using my physics books and pretending I've overlooked the stator-cover thermal connection suggests you do not even get the point.
Yes, it does. Your point was I won't be able to validate my plan measuring side case temps. Your original argument was that they are uncorrelated and can't demonstrate what I claim because I am not directly measuring wire temp but, as I mentioned, our reasons are different.
Whether that goo helps or not cannot be discovered by measuring cover temperature.
I absolutely disagree with you here because I have first-hand knowledge of how its done. I have been lead engineer on this exact type of engineered monitoring system. I used to design and install nuclear reactor feedpump diagnostic equipment that would (primarily) monitor temp and vibration of various components many of which we could not measure directly but had to be monitored "by proxy". Some of the proxy chains I have monitored were far longer and more complicated that this simple example so I have absolute confidence that meaningful data about the stator temp can be derived from the access cap temp. I know it can be done. The reason I say that just measuring the access cap temp can't help is it takes tons of data over long periods of time with a high-quality baseline. And the data has to be analyzed statistically for trends and correlations to infer the status of critical components.
To apply these principles to this case I would need to take a completely stock bike and monitor it for months across a variety of operating conditions. The bare minimum would be a database of road speed, ambient temp and access cap temp. This would be the base line from which I could statistically estimate the average heat flow out of that access cap or side case. Repeat the experiment with a modified bike and the case would be proved - average heat flow would have to go up (and by implication stator temps down) because the thermal compound and a heat sink have improved the conduction of heat out of the stator. (How much of a difference and whether it would spare the stator or not are completely separate questions).
In any case, this all getting a bit ridiculous since neither I nor anyone else is going to do such a test. Moreover, as you suggested, it would be easier and more informative to just stick the thermocouple directly to the stator wire and get direct measurements (they aren't radioactive, lol). But I am not going do that test either because I am not interested in proving this one way or the other. From basic principles I know that thermal compound and a heat sink will, at minimum, move the system in the right thermal direction and that is enough for me. Moreover, I have spent less than $20 and had fun hooking up the temp sensor and thinking about this and all that it entails. I am not posting to convince anyone to do this if they think it is a dumb idea or a waste of time. I think its sparked some interesting discussions and people have even commented as much. If you are happy with the series R/R and think my idea is dumb then commit to your decision and move on.