... and finally, success!
Everything went very well, considering. Pictures are worth a thousand words, right?
That's a quarter of a pound of your everyday silicone/zinc oxide heatsink compound. A layer on the battery, and a layer on the heatsink. The stucco appearance is from separating them after repeated mating checks to make sure there were no gaps. I figured 2 oz would cover both batteries... But that one pound tub sits right at about a half tank with all said and done, and there was very little oozing; maybe a half ounce total if I had to guess.
Anybody that's attempted to remove a heatsink from a CPU with a spreader plate on the die knows that the stuff gets a grip. It holds true in batteries as well apparently.
Eventually, it would give way to fatigue and come off, or slide off to one side and become detached. Couldn't have this happening, so back on with the zip ties.
On the back side, I used aluminum angle to spread out the pressure that would have been on the case of the battery itself. They are cinched down much tighter than they were for the dry test, and I didn't want to risk damaging the casing. Everything is much better off if the acid stays inside the cells.
There has been a bit of debate as to the ideal location of temp sensors. Some argue that the sensors are best placed on one of the posts, the idea being that the lead conducts the heat of the core back to the post and provides a more accurate reading than if the sensor were placed on the side of the battery. For a battery/cell that is open on all sides for ambient temperature air flow, I'd agree. However, due to the way these two batteries are housed and insulated, the side of the case actually provides a more accurate reading than the posts would. Being forced air cooling, the posts would see a lower temp by the air flowing across the top of the batteries. In the pic above, there are 3 sensors, all within about 2 inches of each other in a V. A gasket 2" wide and 3/4" thick, sealed both on the inner and outer perimeters by tape provides an airtight seal and a chamber that completely blocks out errors caused by the cooling air. The original layout before heatsinks were ever a consideration involved only a tape seal around the outer edges of the batteries, and this did cause the measured temp to have some notable "play" in it when the fan was turned on or off. With this, it is now rock solid.
Here is a shot of the heatsink <-> seal interface. This was very difficult to pull off, as it took forever to get the grease completely removed from the faces that needed to be taped. As you can see, it involved getting creative and using the most accessible faces of the heatsinks in combination with the structure of the tape itself, and the unbroken seal on the inner perimeter. The result is a good solid seal. At the very bottom in the center is the intake temp sensor for the differential thermostat.
The much improved passage for the temp sensor wiring.
Everything worked exactly as planned once it was all put together. I stayed up with it to tweak the thermostat so it would cut off at the appropriate time, and checked on it from time to time throughout the day to find out the point at which it would bring it back to life.
Along about the late afternoon hours, I noticed the LED that indicates the power for the fan was on. At first, I only thought of it as a little early for it to engage, but a quick peek at the thermometers revealed that it was in fact on it's mark. The day began with heavy rain, and remained cool for almost the entire day. I had some extra insulation on the batteries and had actually lit off the SMPS charger to fill them back up as well as artificially raise the temperature when I realized the sun wasn't going to do it naturally.
But then I noticed that the LED was cycling, and had an odd flicker in it right before it would cut off. This definitely wasn't normal. A quick check of the fan revealed that not only was it not turning, but it's motor had become very hot for some reason. I have my suspicions, but I don't know for sure. Either way, it was dead at that point. One high performance 4" brushless fan, junk.
To get things back going again, I had no choice, I had to use the fan that had been venting the cab to pull air through the battery box. It's a very similar fan, slightly larger and has a thermister in it that I bypassed to make the motor run at full throttle all the time. It's knee isn't in the right spot, so it would slow down way before I needed it to.
To protect the fan, I ran it through a current limiting buck converter (it needed to be slowed down anyway, so this worked out). The batteries are now once again cooling and with 5 hours or so before even twilight, there's little left for it to do. It's drawing a fraction of the power that the other scheme was as well.
Both last night and tonight so far have demonstrated that the core temp can be brought within about 2F of the intake temperature, which (up until the fan failure) could generally be brought to within a couple degrees of ambient outside.
All in all, I'm pretty satisfied with how this has gone. Much thanks once again to DaveW for the heatsinks.
Now on to reinventing the cab vent