Quoted from mbwalker:
This topic really caught my attention because I deal with this daily and thought it was very interesting. Can it be true? Let's take a closer look. Apologies for a slightly long winded thread.
First off: I'm not a thermal engineer, but rather a EE. But I design with it in mind. Thermal engineers don't tell me what to do, but rather did I do it right? I need to get the heat out of a part to the thermal guys so they can deal with it. No means an expert, far from it. But I design with an understanding of it and maybe I can clarify some of the comments here at the '70,000 ft level'. Not entirely from a flipper/solenoid standpoint, but a thermal issue in general - but it does apply to solenoids (and even transistors and heatsinks) so I will try to focus on solenoids and I hope you find it interesting. The 'things' I design can dissipate upwards of of ~700W in the space of a thumbnail. Ouch! Getting the heat out HAS to happen...no "if, and, or buts" about it or it's a 'bad design day' with big ramifications.
So here it goes...
1st important topic: Kurt and others mentioned thermal expansion, which is a real issue. But we will assume that is not an issue here (or could be easily determined) - but I surely won't discount his (and others) comment because it is valid concern, just maybe not here. Let's assume everything remains friction free.
2nd important topic: Thermal resistance. This is a key aspect and directly relates to the coil temperature (but it gets more complicated in a bit!). This is measured (in my case) C/W (celsius/ per watt) into an ambient thermal sink. In other words, if the thermal resistance of a device (i.e a solenoid) is 10C/watt...for every watt dissipated in a device, the device temperature rises 10 deg celsius. NO clue what a coil dissipates, NOR the C/W, but C/W is the standard by what we go by. Let's say for argument's sake a coil really is 10C/W (into a heatsink). Also let's say it dissipates 5W, that would equate to a 50 C rise in coil temperature...but where does it go? Another heatsink (great)? Or just air (bad)? That is an important question. Everything has a thermal resistance. Let's say there was a heatsink to cool a flipper coil - that connection between the heatsink and flipper also has a C/W number assigned to it and it's not perfect - it can't dump all 100% of the heat. Of course, we all know there's no heatsink to the outside world in a flipper coil - it's all sort of stuck there all by itself w/heat going nowhere, which means the C/W is actually higher since most manufacturers quote C/W into a heatsink. At least the parts I deal with do, don't know about solenoids. Sure, there's metal around it with the brackets, plunger in the middle, etc...but no direct thermal path (i.e. low C/W path) between it and something else...so honestly - I'd write all that off as thermally not connected and it falls out of our of the equation entirely. So whatever the C/W is in air..that's what it is. A coil is a nice little heat source with no where to go, and therefore, it gets hot! Some manufactures only quote the C/W into an ideal (translate: doesn't exist) heatsink.
3rd important topic: The C/W I mentioned above...guess what? It's not uniform. It varies in the depth of coil windings (yeah, now it is getting a little more complicated, but hold on it gets worse). At least the windings near the outside have air it can dissipate into. But air thermally sucks compared to metal, plus there's no air FLOW across. So almost nearly terrible as nothing, but at least there is stagnant air, for what it's barely worth (ugh!). What about the windings inside the coil? There is zero air movement there and basically zero thermal path. The insulation on the winding has a C/W that basically acts like thermal insulator. The nylon sleeve..great thermal insulator too. The plastic form? Forget it thermally - useless. The bottom line..it can get hot and the only way to tell what's going on in there is to insert a thermal probe in the coil itself and measure (which would make a great white paper). If people mention the outside of the coil is toasty, the inside will be REALLY toasty. Thermal resistance of the wire itself (i.e. heat in one end, heat out the other end) is so high it doesn't count unless maybe the wire is ~0.5 inch long - doesn't apply here. Bottom line- the coil is really sitting there all by itself thermally with no place for the heat to go. Wait...we're not done...it's get's slightly more complicated.
4th important topic: So what about the wire? Nerd alert. Copper wire has a thermal coefficient of resistance of +0.00393 per C based on 25C (room temperature). Simply put, as the coil increases temperature (remember it can really heat up on the inside of the coil?), the resistance will increase, current therefore will decrease (more on that later). Spoiler alert: Generically speaking...YES, the coil will become less 'powerful' as it heats up since it is drawing less current. Therefore, it will actually will cool down slightly due to self-regulating w/respect to temperature, which means it will then get hot again since the resistance lowers, which means it will cool down again, which means... Wait, what did I just say? Now even I'm getting getting confused myself! OK, nitpicking there with minor details.
5th important topic: "So my coil draws less current when hot and becomes wimpy, HA! I knew it!" Maybe...the other piece of the puzzle is what about the transistor or FET that drives the coil? It is essential acts as a very small resistor when on. When you draw less current, it drops less voltage and the coil gets a little more voltage which which inherently helps the coil gain just a bit back in current. Not much, but it does happen with a cheap transistor (that's why some transistors have a heatsink). But it can't be ignored. BTW, in this case, the 'on' resistance can be referred to as the 'source impedance'. The rail voltage also has a source impedance, but should be fairly low. There's also some thermal aspects of the driver transistor/FET which I won't cover which can impact the the 'on' resistance on the transistor.
6th important topic: Let's not forget, exactly "How hot is the wire getting and how does it impact resistance?" 1%? 10%? 20%? Real temperatures are needed, not "it's hot" because it doesn't hold its weight (not picking on anyone ). But it is very important to this discussion.
So let's wrap this up: Theoretically, YES a fan can help. The thermal resistance of everything in the coil/bracket/etc. still royally sucks - but on paper, a fan blowing air on the coil will make a difference, the solenoid will be cooler. I hear the grumbles already - "you're full of BS". I'd be the first one to say 'won't help much', but at the same time it DOES help. If it's enough, then 'good for you'. It sure doesn't make it worse. In this case, cooler is better. Pulsing coils and a low power switch mode on a coil can also help immensely. Again, I'm not going to cover it in depth.
In closing, I've just barely scratched the surface. But perhaps I've shed a little light on what's going on. Please don't nitpick my comments - I get it, but I wanted to keep it at a simple level for everyone (and I'm very tired of typing and it's getting late) in at least a somewhat brief explanation (failed at that). This can easily turn into a Ph.D topic if warranted!
There will be a test later!