Supplemental cooling for the charger.

[Don]

There is an OEM fan on the radiator at the front of the car - Any thoughts on just making a mod to use that fan?

Many ICE cars have 12 volt cooling fans behind the radiator which are energized using a thermal switch on the radiator tank - When it gets hot enough, the fan comes on for long enough to cool things down until the coolant switch de-energizes. Something like this shouldn't be too hard to fabricate. Hook power to the thermal switch 24/7 and just let it do it's thing . . . . the fan would automatically run anytime it's needed

If the charger is water cooled, I like the idea of making that work, rather than blowing air on it

Don

 

[Aerowhatt]

There is an OEM fan on the radiator at the front of the car - Any thoughts on just making a mod to use that fan?

Many ICE cars have 12 volt cooling fans behind the radiator which are energized using a thermal switch on the radiator tank - When it gets hot enough, the fan comes on for long enough to cool things down until the coolant switch de-energizes. Something like this shouldn't be too hard to fabricate. Hook power to the thermal switch 24/7 and just let it do it's thing . . . . the fan would automatically run anytime it's needed

If the charger is water cooled, I like the idea of making that work, rather than blowing air on it

Don

So far my data clearly shows that the fluid cooling of the charger is not designed that well. Sure it helps to blow air through the radiator and/or run the circulating pump full time too while charging. Even with these steps the charger still runs hot. Cooling the charger directly with air actually brings the coolant temperature down more than blowing air through the radiator does. Are you drawing some other conclusion from those numbers? When the car is in ready mode the radiator fan works just as it does on an ICE car as you described. However if the design of the chargers fluid cooling is less than adequate (which the data shows) you would need to refrigerate that fluid to do a good job of cooling the charger down.

Aerowhatt

 

[Don]

There must be much more heat generated when driving the car, cooling the motor and the inverter - I can't believe the cooling system that handles that cannot handle the charger alone. In your tests, do you think you were moving as much air through the radiator as the stock cooling fan would? If the water cooling loop can't adequately cool the charger alone, then Mitsu really missed the boat . . . . which is a tad hard to believe

Don

 

[jray3]

While driving, the waste heat loads should be higher, yes, but there is more than just the radiator fan at work. The turbulent air in our motor bay has to provide some cooling as well, compared to the hot stagnant air during a recharge. I'm just glad I never got around to making that rear belly pan for aerodynamic improvement!

 

[Don]

For now - Until we find a way to energize the OE radiator fan - I'm using this blower on medium speed placed in front of the car. It forces at least 90% of it's air into the radiator with almost none hitting the bumper cover

https://www.samsclub.com/sams/pivoting-utlity-fan/prod8670795.ip?xid=plp_product_1_5

That's 435 cfm compared to the 50 cfm that Aerowhatt found decreased the fluid temperature by 9 degrees. Lots more air should provide lots more cooling. I still think some sort of switch to energize the OE fan (as siai47 suggested) is the ultimate answer. I think it's position mounted on the radiator might be even more efficient than blowing air in from the front

Edit: After charging L2 @ 12 amps for an hour, with the car sitting outside in the 95 degree afternoon sun (pretty much a worst case scenario) the inverter isn't hot - I don't have a non-contact thermometer, but I can put my hand on it anywhere comfortably and keep it on there for as long as I care to. It's for sure warm (to be expected) but it's not hot. The DC to DC converter on the other hand is quite hot . . . . too hot to leave my hand on for very long. Since they made no effort to cool it, and it sits atop the inverter, I hope it was designed to run hot. The coolant pump is cycling normally and only stays running for a minute or so when it does kick on, so I think the car's cooling system is probably operating as it should - I think it's pretty apparent that it needs a radiator fan though

Edit 2: Checked it again after another hour. The inverter does feel a little warmer to my hand. The coolant pump is kicking on after about a minute and running for about 30 seconds. By far, the coolest thing back there is the coolant hoses - A little warmer leaving the inverter than on the incoming side, but considering it's 95 degrees outside, the hoses don't feel much warmer than that to me - Much cooler than the inverter itself. If we don't come up with a better solution, I'll be using this fan for all my summertime L2 charging

Don

 

[JoeS]

So, yesterday was a busy day with a non-stop 65+mile round trip to SFO and by mid-afternoon when I returned home I had run the car down to two bars so I just left it sitting there to cool off, with CaniOn showing the battery pack at around 31°C (88°F). Plugged into L1 at around 10pm. Stuck a 6" fan in front of the radiator and placed an 8" fan underneath the left rear of the car, facing upward and left everything on all night long. I just wanted to move air and not let it stagnate, rather than force-cool. By 7:30 this morning (15 bars) here were my readings with an uncalibrated infrared non-contact thermometer -

Car chassis (we cool off nicely at night): 66°F (18.9°C)
Charger chassis top front right side: 95°F (35°C)
Charger chassis top front left side: 87°F (30.5°C)
Charger chassis top rear right side: 90°F (32.2°C)
Charger chassis top rear left side: 87°F (30.5°C)
EMI Filter case: 95°F (35°C) [Edit: had incorrectly called this the BMU and doghouse ]
Fluid hose (left side of Charger: 77°F (25°C)
Motor drive inverter chassis: 74°F (23.3°C)

Even with the air circulating (but not blowing cooling air directly onto the charger), this has resulted in a charger chassis temperature that is 29°F (16°C) above ambient, pretty-well corroborating Aerowhatt's readings. (Sigh)

 

[Aerowhatt]

There must be much more heat generated when driving the car, cooling the motor and the inverter - I can't believe the cooling system that handles that cannot handle the charger alone. In your tests, do you think you were moving as much air through the radiator as the stock cooling fan would? If the water cooling loop can't adequately cool the charger alone, then Mitsu really missed the boat . . . . which is a tad hard to believe

Don

It's not the cooling system of the car. It is well engineered and delivers coolant back to the parts needing cooled at just a couple of degrees above ambient temperature and only 11F or 12F above ambient with the AC cranking full bore while being driven. So the cars cooling system is awesome. The problem arises inside the charger (in the design of the liquid cooling of the charger). The coolant coming out of the charger(while charging) is only ~ 1F hotter than the coolant going into it. There is not good transfer of heat from the hot components of the charger to the coolant loop. If you run coolant through a device that is 56F above ambient on it's case it should come out a lot more than 1F hotter than it went in. Remember Jray's original observation that the coolant outlet hose of the charger was warm but not hot, while the case of the charger would burn you? I measured mine's inlet hose and outlet hose in real time with an infrared thermometer. The heat transfer to the coolant in that charger is criminally pathetic.

Since the charger design does not allow for good transmission of the heat to the coolant it doesn't matter very much how much air is put through the radiator. Most of the heat from the charger never makes it to the radiator to be removed. Most of it is forced to migrate to the aluminum case. Ergo the best cooling (without redesigning the charger) is by cooling the case . . . as the data shows.

Aerowhatt

 

[Aerowhatt]

BMU case (sits on top of Charger): 95°F (35°C)

Do you have a picture of the charger setup in the 2012? It sounds different than the 2014.

Aerowhatt

 

[JoeS]

BMU case (sits on top of Charger): 95°F (35°C)

Do you have a picture of the charger setup in the 2012? It sounds different than the 2014.t

Dunno what possessed me to call it the BMU. What I was referring to we've been calling the doghouse portion of the OBC/dcdc. Edit: Bumbling along, as kiev properly pointed out, it's the top of the EMI filter. Corrected my post.

 

[kiev]

During the course of dismantling jay's charger i discovered that the 'clear' plastic sleeving and black heat shrink tubing on the inductor and transformer wiring had somewhat melted and fused together. i don't know at what temperature that would occur, but it is likely greater than 360 F. i've never seen heat shrink melt while using a heat gun during application and it is rated at 360F max. i posted some pictures of the internal aluminum heat sink for the soldered board in the troubleshooting thread.

i connected to level 2 for charging this morning and measured some temperatures inside the box.

started at 8am, everything at ambient, 76F

checked at 9 am, ambient at 80F; 10am, amb at 83
water pipes
MCU 88; 94
OBC 89; 94
EMI filter top (joe's doghouse), 102; 114
side of obc case, 96; 98
large cap, 127; 144
large inductor rear, 208; 194
transformer rear, 195; 184
trans front, 165; 210
small inductor front, 174; 212
top board surge supp potting, 158; 165

So it's gettin all Hot up in Herre

 

[Aerowhatt]

During the course of dismantling jay's charger i discovered that the 'clear' plastic sleeving and black heat shrink tubing on the inductor and transformer wiring had somewhat melted and fused together. i don't know at what temperature that would occur, but it is likely greater than 360 F. i've never seen heat shrink melt while using a heat gun during application and it is rated at 360F max. i posted some pictures of the internal aluminum heat sink for the soldered board in the troubleshooting thread.

i connected to level 2 for charging this morning and measured some temperatures inside the box.

started at 8am, everything at ambient, 76F

checked at 9 am, ambient at 80F; 10am, amb at 83
water pipes
MCU 88; 94
OBC 89; 94
EMI filter top (joe's doghouse), 102; 114
side of obc case, 96; 98
large cap, 127; 144
large inductor rear, 208; 194
transformer rear, 195; 184
trans front, 165; 210
small inductor front, 174; 212
top board surge supp potting, 158; 165

So it's gettin all Hot in Herre

Wow thanks for taking the time to get these numbers Kiev :!: I was tempted . . . but thought better of opening mine up and risking warranty refusal, god forbid. The rub is that these components will run even hotter without the charger cases cover removed and they bathe all the components around them in the hot sauna that they are creating.

Aerowhatt

 

[Aerowhatt]

While driving, the waste heat loads should be higher, yes, but there is more than just the radiator fan at work. The turbulent air in our motor bay has to provide some cooling as well, compared to the hot stagnant air during a recharge. I'm just glad I never got around to making that rear belly pan for aerodynamic improvement!

Me too! it was on my list, happy to cross it off

Aerowhatt

 

[Aerowhatt]

OK . . . so does this make sense? Looking back at the numbers the lowest coolant temp and case temp numbers were both achieved with aggressive air cooling the charger case only. If the coolant temperature is lower then the components that Kiev has just pictured on the aluminum heatsink (troubleshooting thread) (which are in contact with the coolant loop) would also be running cooler right? Otherwise the coolant temp would stay higher right? I'm thinking perhaps the coolant cooling system was designed to cool only these components on the heat sink. It's the rest of the components on the other boards that are driving the temperature up. From the layout I'm surmising that the liquid cooling was never intended to cool the other components. The case was being relied upon to conduct it out. Which if mounted in a different environment would really do much better. But, Mitsu mounted it in a stagnant (when parked and charging) hot air trap making things much worse.

I'm fine tuning my ventilation strategy before finalizing the design. Once the bay hatch was all put back to the normal configuration I'm seeing a modest rise in the temperatures over the original (best) test numbers. I'm assuming that some hot air is getting recycled through the system once it is all closed up. So right now I'm testing different adjustments to mitigate that back to the best case numbers.

Aerowhatt

 

[JoeS]

...I'm just glad I never got around to making that rear belly pan for aerodynamic improvement!

Me too! it was on my list, happy to cross it off

Same here, although I never could figure out how to cover the transaxle.

kiev, did you leave the cover in place and then remove it to make your measurements, or was the cover off the whole time?

What's troubling with kiev's measurements is the high temperatures achieved by the magnetics. Even if they used hi-temp insulation, the heat generated and dissipated within the confines of that closed case would be especially damaging to not only the semiconductors but the capacitors as well. Aerowhatt, I agree that it looks as though non-heatsinked components and charger location were probably not properly addressed. Pity that's such an environmentally-dirty/wet area which complicates/precludes adding holes and a fan directly onto the charger...

Beginning to look like bonding a finned heatsink to the case with the air blowing over it and perhaps some means of extracting that trapped hot air under the rear deck might be indicated ... jray3's initial list isn't looking that far-fetched.

 

[Don]

A question - What controls the pump cycling??

With 435 cfm going through the radiator, my pump kicks on for about 30 seconds and then is off for about a minute and then repeats when I'm charging L2 @ 12 amps. If it's a thermal switch cycling the pump, then it is carrying away an awful lot of heat generated somewhere. I noticed charging the other car this afternoon using L1 @ 12 amps with no air blowing in the radiator, the pump doesn't seem to be cycling at all

I just put that big 435 cfm blower under the back of the car, pointed up at the electronics, so I'll check later and see how hot things are getting with it running. It moves a LOT of air!

For the time being, since it gets so hot down here in the summer, I've got an original, unmodified 8 amp OE EVSE and I think I'll start using that from now on - We nearly always have 12 to 14 hours between the last time we use either car and the next time we need it the next day . . . . I really don't need to be charging faster than that 98% of the time and it's GOT to stay cooler using 960 watts than it does using 2,880

Don

 

[DkubusEV]

Hi Guys ive been reading this thread with great interest and im thinking it might be well worth getting a bit extreme on this modification for what im building (ImiEv based lightweight race car)....

anyway I think im going to "port" the Charger case and then positively pump cooling airflow into and out of it via dual sealed "dust and moisture" proof pipe work from and into the cabin air space.

This should assist with those burning hot individual component temps?

cheers for the data guys.

Mike.

 

[kiev]

...
kiev, did you leave the cover in place and then remove it to make your measurements, or was the cover off the whole time?

What's troubling with kiev's measurements is the high temperatures achieved by the magnetics. Even if they used hi-temp insulation, the heat generated and dissipated within the confines of that closed case would be especially damaging to not only the semiconductors but the capacitors as well.

Yes Joe that's exactly what i did--left the cover in place except to quickly open and get a reading.

i was trying to figure how to drill a couple of holes in the case to mount a fan--maybe a small squirrel cage blower with a dust filter on the inlet.

Those big caps are Nichicon GX(M) 680 uF, 420 VDC, rated at 105C

 

[coulomb]

Those big caps are Nichicon GX(M) 680 uF, 420 VDC, rated at 105C

420 V? The datasheets seem to say 400 V max for 680μF. The GX series are moderate ripple current, long life (5000 hr @ 105°C and rated ripple current). So: about as good as Nichicon make them*, and I've only seen one longer life model, 10,000 hr (one of the United Chemicon series, quite rare).

*Edit: Oops, I see that Nichicon do make some 10,000 hr models, e.g. the (L)GR series, but those GR series have lower ripple current ratings.

It's possible that they make special order capacitors for these chargers (with a non-standard 420 V rating).

 

[Aerowhatt]

Beginning to look like bonding a finned heatsink to the case with the air blowing over it and perhaps some means of extracting that trapped hot air under the rear deck might be indicated ... jray3's initial list isn't looking that far-fetched.

Yeah, I don't think a heat sink is in order. It might even get in the way with the close quarters when the "hood" is down. The key is lots of fresh air. These blowers I am using are amazing. Thinking computer fan would be way off the mark. These things kick out high velocity, large quantities of air in a small package. The key is to blow so much air over the unit that the entire bay is turbulent and in motion. This allows rapid mixing of the normally stagnant hot air with ambient air being draw up from below by the turbulence. I use a "smoke" test to optimize the turbulence. Put a plexiglass lid on the compartment let the hot air collect, add smoke, Turn on blower and time the interval until the smoke clears. Adjust mounting, rinse and repeat until the results are optimal. There isn't a lot of room to work with, so the solution has to be simple, elegant and effective. No space (or need) for more complicated Rube Goldberg solutions.

Aerowhatt

 

[kiev]

The box has a nichicon part number, and looking thru their website i found they take credit for the miev OBC. In fact they claim that it worked so well for Mitsubishi that Nissan converted over to using their design in the Leaf.

Here is a page from a product brochure:

So i would say that nichicon is to blame if there are any thermal issues. Wonder if we could buy them direct without getting middle-manned by Mits?