- Joined
- Jan 10, 2012
any updates on this project?
-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
New to the forum. I have a unique need for liquid cooling and need experts!
- Thread starter Jayls5
- Start date
- Joined
- May 15, 2006
There is nothing wrong with asking questions like he has. While we are "a computer forum", we can still provide our expertise when it applies.
- Joined
- Jan 10, 2012
I like the odd projects like this.
- Joined
- Mar 18, 2015
If the size of the radiator becomes an issue I'd like to make a suggestion. An air conditioning condenser coil for a car may work. Much smaller tubing (1/2 in to 3/8in) instead of the full size car radiator (1 1/2 in to 2in) I've considered this for a wanky water cooler I was going to build in the future. 
if it's a dumb idea just say so, no need to yell at me
if it's a dumb idea just say so, no need to yell at me - Thread Starter
- #25
Ok, cool. I initially stopped posting because I don't want to bother a forum that isn't interested in something like this.
Most of the parts arrived, including a cheap reservoir/pump combo. I did some initial testing on the cooling block with the little radiator, pictures shown below. I put a 150w load by mounting 3 TEC's on it @ 12v. The PC fan size is perfect for this radiator, and the screw mount holes appear to line up perfectly with it. I ran the system with no fan on the radiator vs fan on. The former was mainly to get a subjective idea of how much heat that water could soak up before the system even needed the radiator. The fan brought the temps back down significantly within about 1 minute. Sorry no concrete numbers, my IR thermometer is back at the build site with the car.
That's with a single pump with tiny built in reservoir, fan, and radiator. I have the larger reservoir, an extra reservoir/pump, 3 more cooling blocks, endless fans, and an extra radiator. Keep in mind that ANY improvement in cooling on this controller means more fun acceleration with the car before it starts dialing things back on me at 85-90c (which is good, considering I don't want to boil water in my cooling system).
- Joined
- Jun 6, 2002
see what you can find for a "water chiller" this will be your best bet IMO, instead of trying to make one. even with the pelts they will need heatsinks on the other side to remove the heat from them.
where i think you need to really start is:
what is the going to be the nominal temp of the controller, ie 50c/60c, to not cut current back. as from there you can figure out how much of what your going to need to keep it at a operational temp with out current being cut back.
i know i dont know all but what do you mean by potted? if im looking at the right picture with that thin copper folded over how to you plan to remove the heat effectively? i would think they would actually need to be tied directly to those blocks you have. to many attachments to remove heat will make things worse and not as effective at removing heat.
what made you want to build a controller vs buying one? there are a few out there that can handle the wattage and be over kill plus they can be liquid cooled as well.
where i think you need to really start is:
what is the going to be the nominal temp of the controller, ie 50c/60c, to not cut current back. as from there you can figure out how much of what your going to need to keep it at a operational temp with out current being cut back.
i know i dont know all but what do you mean by potted? if im looking at the right picture with that thin copper folded over how to you plan to remove the heat effectively? i would think they would actually need to be tied directly to those blocks you have. to many attachments to remove heat will make things worse and not as effective at removing heat.
what made you want to build a controller vs buying one? there are a few out there that can handle the wattage and be over kill plus they can be liquid cooled as well.
- Thread Starter
- #27
The box is an aluminum military surplus medical supply container with latching seals. I wanted really good thermal insulation and electrical insulation for safety. I lined it with a ceramic matting for battery fire protection, then a wooden box to house all of it: batteries, management circuitry, charger, controller, contactor, etc. I bought this controller originally because it stated that its active air cooling required no external heatsinking to perform as advertised. Other controllers are usually meant to be mounted directly to a metal surface, so this feature was attractive for my build. Perhaps their ratings or testing conditions were generous... who knows. I'm sure 50-60c would be fine, but the problem seems to be about heat saturation buildup after 5-10 minutes of spirited driving. I figured my hairbrained idea would address that.
By "potted" controller, I mean they used a black waterproofing epoxy to completely cover the circuit board on the inside. I guess it's an EE industry term. The only exposed parts are the mosfets secured to the copper bands for heatsinking. The copper is in the cooling path of the controller's internal fan. I was using the peltiers in testing just to simulate a heat load, and i'm probably not going to bother with them in the end since it adds complexity. The plan was to direct mount the cooling blocks to the copper with a thermal adhesive like "arctic alumina". It would be practically on top of the row of mosfets. I just checked the cooling block clearance for that and discovered the copper fold narrows too much to slide them in. I would have to bend/widen the top fold of the copper to fit... not too happy about that. The same company makes 2 more powerful models, but I wanted to exhaust cheaper options before having to resort to that. (Their top of the line model costs nearly a grand!)
- Joined
- Jun 6, 2002
what electric motor are you using? its hard to tell from the pic, i know the company name use to be Etek, thats what it looks like to me.
As well controllers will have two different ratings, a nominal output and then a peak output something like a 30sec or 1 minute output rating. as in the case it sounds like that controller might be able to do 20kw for 1-2mins, then it drops down due to thermal limits. this is the case with all of them so that is something that should have been taken into account. as lets say you need 20kw nominal output but they are listing them by peak instead, you might have needed a 30/40kw one. since i dont know who made that controller or where its from, its not one i have seen before though i dont search that much for them. most of the ones i have seen used in high power applications like the "white zombie" use zilla controllers or others built for that type of application.
As well controllers will have two different ratings, a nominal output and then a peak output something like a 30sec or 1 minute output rating. as in the case it sounds like that controller might be able to do 20kw for 1-2mins, then it drops down due to thermal limits. this is the case with all of them so that is something that should have been taken into account. as lets say you need 20kw nominal output but they are listing them by peak instead, you might have needed a 30/40kw one. since i dont know who made that controller or where its from, its not one i have seen before though i dont search that much for them. most of the ones i have seen used in high power applications like the "white zombie" use zilla controllers or others built for that type of application.
- Thread Starter
- #29
I'm impressed. You know your stuff. Zilla has an internal contactor in their controller. I have an external 400A contactor. SPM72400 model "Alltrax" is the brand. American company, but I'm not sure what that means anymore.
Etek basically applies to an old briggs and straton design. I'm not going to bother with radial vs axial air gap design explanation, but basically... there are neodymium motors available up to the 14 kW continuous range in small packages.
My motor has been custom adapted with an automotive fan of the same diameter:
Per the manufacturer, this could increase continuous cooling as high as 275-300A continuous. That will make my 1000 lb ride good for blowing away average traffic.
Overheating motor controller still performs well:
- Joined
- Jun 6, 2002
so exactly the air cooling is for the controller or the motor? im a bit lost on that part, the cooling would help with short bursts for higher torque though not really more power per say. what motor though you never did say or link to it, so is it from mars llc? i haven checked their site latly to see what updates they done to thier motors. from what i recall the guy doing mars llc is one of the guys that help come up with the etek motor B&S used.
is that a video of your "buggy" running? probably not what you want to hear you might have needed to go with a higher rated controller, like the 72500 or 72650. i guess that would have been dependent on your minimum contentious torque output. another thought is too, you didnt say what type of batteries your using[lead,gel cell,lipo4]? you sure it might be the controller and no the batteries sagging after some spirited runs?
my thought if i was going to truely liquid cool that controller for higher rated (in time, ie 1min/2min) current output. i would get those blocks as close as possible to those output transistors. as in if possible removing the copper that is attached and directly epoxy the block to them. if that wasnt possible remove as much of the copper fins in side that i could as in excess that would not be in contact with the water block. if its 4in wide then anything past that 4in gets cut off, this will now mean you will need a rad with active fans and then one that would get cooling while in motion or just make both with active fans. from what i read about that controller you want to keep the temp max 85c to keep it running. though it designed as is to operate in this range, im not sure what really is going to be added by water cooling it. lets say if your lucky with the water cooling keeping things in check to get another 5mins of peak current ratings added across the board to what they list. would that be good enough? what is good enough, 50ft lbs/100ft lbs?
im not sure how much height you could add but if you went the pc rad route you could go "A" frame style on the top and get rads mounted up top that way. then if you got enough length you could do a 4x120 rads or if you dont go to a 3x120 rads.
is that a video of your "buggy" running? probably not what you want to hear you might have needed to go with a higher rated controller, like the 72500 or 72650. i guess that would have been dependent on your minimum contentious torque output. another thought is too, you didnt say what type of batteries your using[lead,gel cell,lipo4]? you sure it might be the controller and no the batteries sagging after some spirited runs?
my thought if i was going to truely liquid cool that controller for higher rated (in time, ie 1min/2min) current output. i would get those blocks as close as possible to those output transistors. as in if possible removing the copper that is attached and directly epoxy the block to them. if that wasnt possible remove as much of the copper fins in side that i could as in excess that would not be in contact with the water block. if its 4in wide then anything past that 4in gets cut off, this will now mean you will need a rad with active fans and then one that would get cooling while in motion or just make both with active fans. from what i read about that controller you want to keep the temp max 85c to keep it running. though it designed as is to operate in this range, im not sure what really is going to be added by water cooling it. lets say if your lucky with the water cooling keeping things in check to get another 5mins of peak current ratings added across the board to what they list. would that be good enough? what is good enough, 50ft lbs/100ft lbs?
im not sure how much height you could add but if you went the pc rad route you could go "A" frame style on the top and get rads mounted up top that way. then if you got enough length you could do a 4x120 rads or if you dont go to a 3x120 rads.
- Thread Starter
- #31
Ahh, sorry. Somehow the part where I stated the model got deleted. My motor is a ME1003. Motenergy is the actual company. It's their most powerful brushed neodymium magnet motor. The owner specifically told me that forced air cooling could hypothetically increase power handling between 275-300A continuous. It has a peak output in the neighborhood of 60 ft-lbs. The vehicle only weighs about 1000 pounds.
Yes, that is a video of my buggy running. You and I seem to think alike. I really wanted the cooling blocks on the fins of those MOSFET heatsinks that are attached to the copper to pull that heat away as efficiently as possible. Basically, I want to get to the point where sustained highway speeds are capable and a spirited run up to that speed isn't a problem. I was putting off a purchase of the SPM72650, but went ahead with it after discovering the copper wasn't wide enough to slide in the blocks without modification. I didn't want to kill the resell value of a good controller that I'm clearly using for purposes beyond its capabilities.
Batteries are hardcase LIPO. It's LiCoO2 cathode, about 3.7v nominal and 4.2v peak. Total pack size is 90AH @ 74v nominal. Voltage sag is negligible. Assuming I wanted absolute peak 400A output to the motor, the individual cells would only see 4.44C peak discharge and they're supposedly rated for 20C continuous. Even if you factor China's tendency to overrate, highway cruising speed should hypothetically be under 2C, which these batteries would laugh at. After all, this 1.1 pound battery can jump start a mac truck, and I have 90 of them.
- Joined
- Jun 6, 2002
yea that is the company i was thinking of, forgot they changed names to montenergy, they use to be mars.
well you got two things going on, looks like the peak power is going to drop off the motor side first looking at its 1min rating. the 1003 is rated for 72v@400amps@1min, thats 28.8kw. you may have the right sized controller for the motor and your just asking more continuous then what either can do. yea that motor can do 60ft lbs but that falls into that 1min rating. doing the math from some of the other motors listed, he doesnt list nm/amp or ft lbs/amp on the 1003. just assuming it is the same as some of the other of the same weight and near the power output KW, it should be around .15nm/amp. that means with 400amp*.15nm=60nm=44.25ft lbs, looking at the table it has 315lb/in or 26.25ft lbs. that is assuming the 315lb/in is going off of the non-fan cooled motor at 200amp continuous that works out to be .178nm/amp.
you would then be looking at 200amp+275amp=475*.178nm=62.3ftlbs to the high side of 200+300=500amp*.178nm=65.64ft lbs. now im starting to see where that controller is actually going to be your limiting factor and that a higher rated one is going to be the way to go vs modding the current one. i think its a bit to bad that a higher rated ME1003 is out there with liquid cooling support something like the ME1302.
*edit*
found what i was looking for, did you ever see this?
http://www.evwest.com/catalog/product_info.php?products_id=364
get both motor and controller for $900 is a steal for that kind of power and high RPM.
well you got two things going on, looks like the peak power is going to drop off the motor side first looking at its 1min rating. the 1003 is rated for 72v@400amps@1min, thats 28.8kw. you may have the right sized controller for the motor and your just asking more continuous then what either can do. yea that motor can do 60ft lbs but that falls into that 1min rating. doing the math from some of the other motors listed, he doesnt list nm/amp or ft lbs/amp on the 1003. just assuming it is the same as some of the other of the same weight and near the power output KW, it should be around .15nm/amp. that means with 400amp*.15nm=60nm=44.25ft lbs, looking at the table it has 315lb/in or 26.25ft lbs. that is assuming the 315lb/in is going off of the non-fan cooled motor at 200amp continuous that works out to be .178nm/amp.
you would then be looking at 200amp+275amp=475*.178nm=62.3ftlbs to the high side of 200+300=500amp*.178nm=65.64ft lbs. now im starting to see where that controller is actually going to be your limiting factor and that a higher rated one is going to be the way to go vs modding the current one. i think its a bit to bad that a higher rated ME1003 is out there with liquid cooling support something like the ME1302.
*edit*
found what i was looking for, did you ever see this?
http://www.evwest.com/catalog/product_info.php?products_id=364
get both motor and controller for $900 is a steal for that kind of power and high RPM.
Last edited:
- Thread Starter
- #33
Wow. I had not seen that. Unfortunately, the inverter's supply voltage requires 250v-400v. My easiest rewire of my current pack would only give me 45AH at 148v nominal. I am maxed on space for batteries. Then I'd be looking at a new coupler to mount the motor itself. Then I'd be worried about exploding the 1970's transmission with those RPM levels. Otherwise, I'd jump on that immediately! I wish I wasn't so far into this design. Oh well!
You did the same torque estimations I did. And once again, since I'm forced air cooling with an external radiator fan, I get to deliver more than the 200A continuous.
The reason some of the brushless Motenergy motors are able to be liquid cooled is their axial air gap design. ME1003 is radial air gap and more difficult to supply coolant lines to areas in need.
- Joined
- Jun 6, 2002
thats true but the outer shell will still get hot, you could have water channels in the outer casing to remove the heat. its just a thought i had but i wouldnt try to mod one of those, if i were to do something like you are. if i had the money doing this project i would have been more inclined to go with the Lem 2x2 D95B setup for the 6k rpm. if i could live with the lower rpm the D135 RAGS for the higher torque, being 30nm vs 42nm. in a 2x2 configuration you could parallel connect them so your torque would doulbe to 60/84. would be tuff call though series[higher pack voltage] connect them or parallel[higher pack current]. would be interesting if they had a controller that worked like the zillia on the white zombie where it would switch from parallel connected to series.
- Joined
- Jan 10, 2012
- Joined
- Jun 6, 2002
that looks more like a heater core then a computer rad. though in the early days some of us (me) used heater cores with custom made fan shroudes for 120mm fans. i have thought about tring to do a electric bug. the company went under i think but they offered a adapter to have two 9in kostov motors mounted in the bug, more of a hot rod. i dont know the motors that well but i think that might be a hpevs or was the old name, AC-31 more the likely or the AC50/51 motor.
after reading this thread though, i though about trying to find and older Suby awd/manual. use one of the AC motors with about 180 to 200ft-lbs, in a awd setup that could be fun.
- Thread Starter
- #37
That pic of the bug looks like one of the larger Curtis A/C motors. I stayed away from A/C initially because of the cost of the controllers, which is ironic now that I've ordered the SPM-72650. It has been on backorder, and it will finally be coming in this Wednesday. A quality controller is not cheap, and I don't have the spare time to design my own. I certainly hope this solves my issue. Looks like the computer cooling stuff will actually get used on a computer after all.
Similar threads
