Active Controller cooling hack

[John in CR]

I decided enough of these hot controllers. I took a small centrifugal fan that came in my Bosch chargers and were left over after that hack. It's a 4cfm fan, so my .5 liter controller will have it's air replaced 3 times per second, actually a bit faster since I'm overvolting the 12V fan a bit to 14.8V nominal off of my light pack. If I wasn't such a caveman hacker I'd put a thermal switch to turn it on. Instead I'll just use a forward/reverse grip switch to turn it on manually for hills.

I didn't have a nice neat way to have it blow into the controller, so I cut a big hole in the controller end cap and epoxied the fan to the cap to work as an extractor fan. Being a centrifugal fan it works fine that way, but hopefully the air intake temp never gets high enough to damage it. The controller tilts down at the back and the fan exhaust is at the bottom, so even if something got in while the fan is off sand. water, etc. won't get into the controller that route.

Intake is at the front of the controller from the bottom near the 1st FET. I used a pair of 3/8" nylon tubing that extend down so debris and water can't easily find their way in, and I plan to tape a little balloon of screen as a filter with restricting air flow. Good thing I put 2 tubes. One restricts airflow too much and slows the fan and it seems to run properly with 2. If I even have to get into the controller again, I'll add another tube just to be sure the fan draws air freely.

 

[grindz145]

Very nice John! I thought about doing this myself. I see no need since I can pretty much smoke any hub motor with my controller as is.... However for a high performance motorcycle motor like a perm132, this would be a very nice solution. You can pull an immense amount of current through FETs, so long as they are cool enough

 

[tostino]

OMG that is just about the exact same mod I was going to do. Beat me to it!

I'd love to see how it works for you though!! If you can post any data, that would be amazing.

 

[John in CR]

My thought is that doing this almost doubles the useful surface area for the case as the heat sink. In addition, I get some direct cooling airflow across the surfaces of the interior components. If this works out, then my cargo/schoolbus bike will knock DocBass out of the top spot, because I'll move my 18fet 150V controller into one of the larger cases I have after adding significant fins in the FET boltdown area. Then I'll go overboard with fans (at only 2W each why not) including at least 2 fans inside blowing directly on the FETs, and 3 to move air through the case. On the outside I'll direct some airflow directly at the fins in the area of the FETs.

Assuming voltage down to 135V, that puts my no-load speed at about 125mph, so I won't get out of current limiting until over 60mph. With all that cooling I should be able to safely push close to 100A through 18 4115fets. Since the bike won't reach max power until above 60mph and plenty of power on tap and plenty of motor rpms to go, it's just a matter of aerodynamics, and my low bike has an aero advantage over typical uprights. Sure the wind gets hard to cut through, but I did hit 60mph on sagging to 75V with a 50-60amp controller, and was with a backrest sticking up in the wind like an air brake and no real effort to explore the most air efficient body position. With more areo bars than the BMX handlebar I use, no backrest, and 135V, I think I can get past the 82mph pedal only record without going to dual motors like I originally planned. 8)

John

 

[docnjoj]

That is some clever engineering, John. Do be careful out there.
otherDoc

 

[John in CR]

That is some clever engineering, John. Do be careful out there.
otherDoc

Thanks Doc. I'll be doing my run on a nice new 4 lane highway at a time of little traffic, so other than a front tire blowout, it's going to take something freakish like an ill timed strong earthquake to knock be off the bike. This is just a one time shot while a new number is still within easy reach. 50-55mph works fine for me in regular use, and only comes in play on the highway. Bat out of hell acceleration and great hill climbing is all I'm after in the long term.

John

 

[grindz145]

It probably wouldn't hurt to add a thermal sensor on one of the heatsinks too. It would be nice to have a guage with both the motor temp and the controller temp.

 

[John in CR]

Thank goodness I added the fan, because I went too short with the shunt wire I added. I had some errands and didn't wire up the fan yet. I had noticeably better acceleration, and with just flat road riding, the controller was scorching hot (70-80°C as best guess hand estimate, too hot for more than about a second of touch), so this will be a real test of how well the fan works, though I will have to ride with it always on. Rain is coming, so probably no results till tomorrow.

Assuming the fan gets my controller case down to moderately warm, even at the peak of hills, does that mean I'm out of the woods, because this level of performance is great for a cargo bike, or do I still risk the FET's getting too hot despite the faster heat dissipation?...Not where I want to get ultimately, but well on the way.

John

 

[markobetti]

i like it very much... Very good mod and idea

 

[John in CR]

i like it very much... Very good mod and idea

Thanks Mark. I'm trying to figure out a way to provide power to the Colossus on the cheap.

John

 

[John in CR]

The rain didn't make it, so I had a chance to test with the fan, and it definitely works. I've got a fan problem though and the fan has slowed to just a fraction of it's normal rpm. I think maybe it got gummed up by fumes from the contact cement I used to attach the rubber gasket to seal the end cap to the controller case. Note to self- Mount the fan last, and bolt it on don't glue it to make replacement easier. I took the bike for a harder ride than made the controller so hot earlier with the fan running, albeit slowly, and after the ride the case was barely warm. Time to get some proper air flow through there and go tackle some big hills. Bottom line, it works even better than I hoped, because the stagnant air in a sealed case can get very hot.

Once I get a fan working properly I'll dig up my temperature probe do some measurements with and without the fan.

John

 

[AussieJester]

Very neat John...reminds me of a similar mod i did to a modem router that used to get a lil too hot in summer...

KiM

 

[amberwolf]

Good mod; basically the same thing I talked about doing to the Lyen 6FET after I blew up caps in it from the heat here (although I have not yet done that; I also have not ridden as far with that kind of a load on it, and I have been running it with both endcaps off, inside my cargo pod).

As for damaging the fan by the heat from inside the controller, it's unlikely. In PC power supplies, UPSs, and car stereos, as well as a few carputer projects, I've seen even really cheapo fans survive for years under heat sufficient to curl up their little plastic/foil labels and actively brown the paper ones. As long as the plastic doesn't distort from the heat (also unlikely), I think it'd be fine.

Regarding the amount of airflow, well, nearly *any* airflow is better than none, as far as I can tell. Just leaving the ends open on mine has helped dramatically, since I could get the case hot enough that I would not want to keep holding it before doing that, when pulling the kennel trailer full of dog or dogfood. Now it's just hot. A fan would help even more.

 

[John in CR]

There's definitely something wrong with that first fan. It runs fine for 10 seconds or so and then it slows pretty drastically. I have its replacement ready to go with a quick endcap change in the morning. The replacement will mount in a slightly different manner utilizing some of the materials from the charger.

The cool thing is, not only does this stretch the performance of my bike with it's everyday controller, but I'm now comfortable in running one of my high voltage controllers with my big hubbie with only RWD, instead of having to wait until I finish my 2WD to give my HF controllers a first test run at something north of 130V. Hey Doc, you hearing footsteps yet, or in this case tire hum?

John

 

[tostino]

What kinda fets and how many are in that controller by the way? I didn't see you say that before .

 

[John in CR]

It's a 15 fet controller that came with my motors using the 846MCU board# EB715. I don't know the FETs because they sanded off the face. It's a 60V30A controller stock, and the only info I have is that my factory rep said the engineers said don't exceed 83 or 84V fresh off the charger. It does have 100V caps. Based on experience with other controllers on this bike, it seems that my shunt mod increased the current limit to about 70A.

What controller it is doesn't matter that much, because my 2 turn 16Kv motors are hard on controllers. Partial throttle on a hill killed my v1 Methods 100V100A controller running at 74V nominal with a 23.5" tire. I need the active cooling, and will even do so on my 24fet controller.

This am after my school bus duties, I continued on for a 10mi run including hard running and some moderate hills and the controller was still barely warm, despite the low fan speed. I'm still going forward with the fan swap, because at slow it's just too quiet and I'll forget it on all the time. That's no big deal, since my light batt pack would run the fan for 3 or 4 days. I want a little noise to know it's working and I want max air flow.

Here's fan2 mounted to the end cap. The white part is the intake and is the part on the toolpack charger that mates with the battery pack to draw air through while it's charging to cool the pack and the charger at the same time. I have plenty of length available under the motorcycle seat on my bike, so the fan side sticking out of the controller isn't an issue for me.

 

[tostino]

*ordering 3 blowers right now* =D.

I have two 12 fet 4115 controllers built up (other than wires) already, I think i'll test it on them and my current controller on my bike.

I'm thinking about ordering 72 4115 fets with method's group buy, and seeing if I can get a couple bare 36 fet boards from Keywin to build up two of them after I end up killing those 12 fets XD. Oh what controller fun there is to be had!

 

[John in CR]

*ordering 3 blowers right now* =D.

I have two 12 fet 4115 controllers built up (other than wires) already, I think i'll test it on them and my current controller on my bike.

I'm thinking about ordering 72 4115 fets with method's group buy, and seeing if I can get a couple bare 36 fet boards from Keywin to build up two of them after I end up killing those 12 fets XD. Oh what controller fun there is to be had!

2wd with actively blown controllers may get you where you need with the 12fet controllers, even with the higher resistance 4115's. How high were you wanting to go with the current? Now I feel much more comfortable now giving my 12 & 18 FET controllers with 4115's a shot at my motors, and I probably have the hardest to drive motors on the forum. I was going to try one with 37s for a few speed runs really babying the controller with flat road only. No more babying it now.

I just have to figure out how to true and balance the wheel with my motor in a 14" motorcycle rim. It gets a pretty strong vibration above 60mph, and that's probably why this bike won't punch past 60 with my 20s pack. First, I guess I should straighten the spokes I've bent on the handful of occasions that I forgot to take the chain off before moving it.

John

 

[tostino]

Lol, honestly I am kinda wanting to be able to smoke the tires from a stand still on dry road . Not sure how possible that even is though with a 9c, I'd probably need a x5 to get the torque required to do that.

I was wanting to try and push 100a phase and battery (1:1 ratio) pr controller, possibly more if they will take it. Right now i'm pushing 160+ phase amps on my 12 fet 4110 controller with a delta/wye 9c, and that isn't near enough to get the tire to slip, though it is enough to throw my recumbent in a wheelie XD.

 

[John in CR]

Not much use going past 90A with a 9C, since the stator saturates at 90. Justin tested it.

You gotta get video of a recumbent wheelie. Unless it's one of those high short wheelbase bikes that's something special, though not something I'd even dream of attempting.

If you get tire smoking torque, what are you going to do if it rains while you are out? We really need a dual program settings switch. The throttle limiter isn't enough.

 

[tostino]

It helps that my battery pack is right over the back wheel, so it doesn't add anything to the front of the bike weight. It does take just a small (very small) pedal effort, and the front wheel comes up a like a foot off the ground. It is scary as hell, and incredibly unstable. I'll see if I can get someone to video me being retarded.

I know it starts to saturate after 90a, and you get diminishing returns, but it's not completely useless to go a little higher. I do notice a huge difference between 100a phase currents and 160a phase currents on my bike now.

I am not too worried about smoking the wheels in the rain though, I am pretty good with throttle control. I will probably be setting my CA to do torque based throttle up on this bike anyways.

 

[John in CR]

It sounds like leaning back to far in a chair to me. I ride like an old grandma compared to many of you guys, yet I'm the one that gets a hard time from Doc and AJ just because I only wear a helmet when I perceive greater risk (unfamiliar roads, wet roads, or when my ride will take me on the highway). I'll leave the wheelies to you guys.

 

[tostino]

If it were an upright bike, i'm pretty sure i'd have no trouble pulling wheelies without any pedaling, and the batteries in the triangle. it's not extremely easy to lift the front tire off the ground by hand without me on the bike, so it's not solely a weight too far over the back problem .

 

[markobetti]

John in CR this was my idea, similiar to yours. Sony vaio - big one- old edition uses one very small cooler , and efficiently cools the motherboard and other components . Look at its design :



Uploaded with ImageShack.us

Maybe it would be good idea to put copper on behind the fets (sepparated with thermal textile) like this cooler uses extended copper part to cool of the processor and put ventilator in housing ..

Second idea is this : http://www.madshrimps.be/articles/Larkooler-complete-liquid-cooling-solution-leeghoofd-28743.jpg
dimensions are (L)48mm x (w) 14mm x (H) 14mm, weighing in at 135 grams ( you would need few ones for your fets)

Putting directly at the back of the fets waterblock . Ofcourse in the middle of waterblock and mosfet we would need to have again thermal textile ..

Or take the whole Lacooler kit and put controller on
http://www.madshrimps.be/articles/Larkooler-complete-liquid-cooling-solution-leeghoofd-28769.jpg
http://www.madshrimps.be/articles/Larkooler-complete-liquid-cooling-solution-leeghoofd-28742.jpg

I havent been able to test this things yet since our controllers are not here

 

[tostino]

I love the idea of the mini water block and radiator! Would be amazing if you could get it working properly. That, in conjunction with air cooling would be simply amazing (air blowing over the front of the fets still can't hurt).