Watercooled drives/speed controllers.

I was wondering how many of you guys are running watercooling on your drives? I have an ASI BAC2000 I will be using with a crystalite Crown 4080 with heatsinks and ferro fluid. I was looking at how much air flow there will be in the case, and decided to pick up some PC water cooling parts and give that a try on the drive in hope of being able to push some decent power without smoking it.

I got a radiator that is designed to fit two 120mm fans, a pump rated for almost 10 feet of head, and a pair of 40x80mm water jackets to mount the the plate of the ASI drive. I don't think I am going to add fans to the radiator, but it is more of an experiment at this point. I have a em3ev EEB frame and an planing on mounting the radiator on the frame behind the wheel, and making a composite fender to keep rocks and dirt from being slung in the radiator fins.

Have you seen this thread?
https://endless-sphere.com/forums/viewtopic.php?f=2&t=48753


Or any of these (not all are relevant, but most are obvious by title):
https://endless-sphere.com/forums/search.php?keywords=water*+cool*&terms=all&author=&sc=1&sf=titleonly&sr=topics&sk=t&sd=d&st=0&ch=300&t=0&submit=Search

https://endless-sphere.com/forums/search.php?keywords=liquid*+cool*&terms=all&author=&sc=1&sf=titleonly&sr=topics&sk=t&sd=d&st=0&ch=300&t=0&submit=Search

I said nothing about watercooling a hubmotor or any motor in general. I know how to use the search function. I found only a couple of old threads about watercooling drives, so I thought I would bring up the question to see if many users are doing such as it is easy to do and can be of benefit. Maybe you do not understand that an electric drive is what many call a motor controller or an ESC and not a motor. Yes, I did look at the few posts that I found on doing such.

I was originally going to build a much larger bike based off of a Curtis controller, and I found that their failure rate is exceedingly high with just heatsinks; however, in the EV retrofitting world they sell bolt on water cooling plates to replace the aluminum backplates for increased cooling capacity. At the IAH airport they literally throw away hundreds of those drives on a regular basis due to overheating and subsequent component failure. Maybe you didn't mean it that way, but I would not have started a new thread if there was not a plethora of discourse on the subject already. I will say that after examining those drives it is rather obvious why they have cooling issues.

footbaldd said:

I said nothing about watercooling a hubmotor or any motor in general.

Click to expand...

FWIW, "drives" as I've seen used here on ES at least could mean motors or controllers, didn't know that you were excluding info about motors and only wanted controller info. (for instance, terms such as "mid-drive", "hub-drive", etc refer to the motor location and not the controllers specifically).

However, the systems used to cool motors are often just as useful to cool controllers, so the info is still applicable. Plus some people that are cooling their motors are also cooling their controllers, and may have also posted that info in their controller posts/threads.

I know how to use the search function. I found only a couple of old threads about watercooling drives, so I thought I would bring up the question to see if many users are doing such as it is easy to do and can be of benefit. Maybe you do not understand that an electric drive is what many call a motor controller or an ESC and not a motor. Yes, I did look at the few posts that I found on doing such.

Click to expand...

Didn't mean to offend with my post; just posted the info I knew existed in case you hadn't seen it. Didn't want you to miss out on useful info that others might not post in your thread for you to see.


(Most people do not know that they can search, or how, and have no idea there is already good info out there, especially when they've only been here a few months. Since I've been here at least several tims a week (usually almost every day) for almost a decade I've seen a lot of threads and discussions, and try to point out where such info exists for those looking for it. Unfortunately many people don't like that, for whatever reasons).

footbaldd said:

I will say that after examining those drives it is rather obvious why they have cooling issues.

Click to expand...

Most controllers are built in ways that create thermal bottlenecks, and can usually be improved with a few mechanical casing changes, depending on exactly how they're built.

One of the biggest problems is that an optimal electrical design may often preclude an optimal cooling design, given a specific casing-size or -shape limitation that might be required.


THe old Curtis brushed controller I used on CrazyBike2's last powerchair mid-drive, have uses two L-shaped FET mounting bars taht then bolt to the external casing. I think there may actually be another intermediate plate between those, making the heat path even worse, but it's been years since I've had it apart.

Most of the common "ebike" and "scooter" controllers use a similar method, moutning the FETs to an aluminum bar that then bolts to the outer casing, which either is the heatsink or bolts to yet another one, adding layers of heat-path interference. (made worse by the poor fit most of them have, and either excessive or insufficient (or zero) thermal paste, rough non-lapped mating surfaces, etc.).

Some members have modified the heatpath, reducing the number of layers, by removing casing parts and replacing the mounting bars with complete single-piece heatsinks.

Some have used water or other coolant pipes inside those. Some of these used various pumped-coolant radiator systems, and some just used passive-movement systems.

Some have even used multiple segments of heatsink on different parts of the bridge, that are not electrically connected to each other, so that no isolation pads have to be used on the FETs, improving the heat path even more.

Some have used oil- or other liquid- filled controllers, or thermal potting compounds, but these may cause heating of other components that would not otherwise reach as high a temperature (or not stay there very long), and decrease their lifespans (electrolytic capacitors in particular).

Some simply removed the endcaps and added forced-air circulation in one way or another.

Some have completely designed their own controllers and/or power stages from scratch. One in progress puts them in a triangular "tube" layout that could optimize both electrical and cooling requirements.


Sorry I don't have direct links to any of those threads or posts, but they should be possible to find if you're interested in them.

Thanks for the response. That is much more interesting. I do not want to rip apart the BAC2000 just yet, but the Curtis controllers I have sitting here disassembled have three layers of PCB boards, the bottom being flush mount fets with a cooling plane in the PCB, the intermediate being a capacitor bank, and the upper being a 64Mbit microprocessor board. I am planning on sandwiching a pair of aluminum cooling jackets to the back of the ASI drive, and an aluminum plate to secure the sinks and add thermal mass.

When I get to mounting the radiator and cooling jackets, I will share some pictures if anyone is interested. Right now I am busy getting the battery ready, and I just realized that I need to wire a proper outlet at the box outside this house to get rid of the crazy voltage sag on the branch under welder load, as well as for charging at 20 amp. I am going to throw a 30 amp with GFCI at the box. I tried running my welder today and it was a lost cause on the current outlets. I have also been thinking about water cooling of the BMS and maybe forced air cooling of the battery.

Any updates on this?

So I have been fairly busy trying to get the bike up and running. It turned out there was no good way to fit both the BMS and drive into the case cleanly even after I widened the body by 1.5" to make more battery and padding space. I have my drive on the underside of the frame for now behind the front wheel. That is where I had planned to mount the radiator, so I am not going to worry about the watercooling just yet. I have all the parts, but I decided to get the kinks worked out on the bike first, and then worry about cooling the BMS and drive for higher power.