Water cooling high VOC classic 150

[offgridQLD]

I have two classic 150's One is being fed around 70 - 80v - 4000w and runs nice and cool. The other is being fed 110- 115v - 4200w and runs a bit hot for my liking. It's not in a runaway thermal condition but I do see the turbo fan come on a a bit when it's getting a workout. Even lightly loaded I notice if I run my hand around the back of the case and feel about where the two inductors are imbedded into the case (left hand side about 2/3 the way up the case so top inductor) there is a real hot spot there.

So I have a plan....project.....ok I'm bored I wanted to wick that heat away from the inductor area and dump it away from the controller. So I thought to myself what about I small water cooled chill block like they use on laser cutter heads. Wicking away the heat at the location where its being made (inductor area back of case) and dumping it through a small radiator with a fan on it. Small pump to circulate the water.

The aim was to never see the (turbo fan) come on My tests showed the two small internal fans come on at 48C and stay on until the temps reach 45C and repeat that cycle . If they cant keep tabs on the heat and it rises above 48C and gets to 58C then the turbo fan kicks in.

So I set a target to aim for and that's 50C or less.

Parts arrived yesterday for the water cooling of the charge controller project. I have about $45 invested into it so if it doesn't work It's not the end of the world but fun trying. (I will use it for something different)



I made a small video on my utube channel going over the parts and the plan.
https://www.youtube.com/watch?v=5qjLpfpASCs&feature=youtu.be

Kurt

[toothy]

Hello

Looks like fun.

I would, being the cheap bugger I am, try to use some finned copper hot water baseboard heater mounted diagonally on the wall above the CC and see if a thermo syphon wouldn't dissipate enough heat, and save a few watts.

It would be curious if you had some logged temp data before and after your mod.

Good luck
Wade

[offgridQLD]

I do have temp data logs from both controllers so I will be able to do a direct comparison. I'm not to fussed about the power consumption of the setup under 10W when active during the hot part of the day (I have 8200w of PV feeding the controllers)

Copper is expensive. large heat sinks are not cheap. So you would have to get lucky at a salvage yard. Part of the advantage in the water cooling is you can place the radiator away from the controller rather than just recycling the hot air .Particularly if the controller is in a tight space.

Kurt

[Halfcrazy]

This is interesting. I wonder what would happen if the fins where machined off under the fets and the heat sink was attached there.

Ryan

[offgridQLD]

I wonder what would happen if the fins where machined off under the fets and the heat sink was attached there.

Well I didn't really want to go hacking up or  highly modifying in a non reversible way my lovely looking aluminum case on my new classic. As It's one of two classics powering my house (not a R&D lab disposable sample) So I think the fins will stay on the fet area for now. (though its it would be interesting)


That said a very gentle ....delicate light cut with the mill to create a nice shiny flat surface over the area where the inductors are. The same shape as the liquid cooled chill block might be necessary on the back of the controller case. Just to get a perfectly flat area to achieve good thermal contact.

Given  water cooling offers HUGE thermal conductivity advantage over  air (0.6 vs 0.025 W/(m·K) It just might work.

My thinking is no matter how efficient the charge controllers are when your pumping a lot of power through them 4000w+ and high Voltage PV even 1 or 2% conversion efficiency loss can give you 80W or so extra heat to try and extract. Put that in a compact case and do it for hrs on end. Particularly in Australia where the ambient temp might be pushing 100F and things get toasty.

The power loss from the 1 or 2% isn't the issue it's more the effect that heat has on component life.

A lot of Electric car power components are water cooled (even the little 2200w on board charger on my OEM electric car is water cooled. Not a lot of power its dealing with but  it's often on for 5hrs + and they want them to last)

Kurt

[mike90045]

I'm also of the mind to cool the FET area, instead of the body of the case.  Maybe weave & press soft copper tubing into there to carry away heat.   And use thermal-siphon to flow and cool it.  powered pump, blower, on a great solar day when you need it, the power is not the issue, but it's 2 more things to fail, at the worst time.

Maybe mounting the case to a slab of aluminum plate via thermal pad ?
http://www.chomerics.com/products/thermal/gap-fillers/high-thermal/index.html  I gotta look into that as an option, along with a 1x3 .25" aluminum plate to go on the wall, maybe 18" wide, to span 2 studs.  I'm thinking of doing that to my Suresine, and up the high load time.

I've also be quite intrigued about the 2 yellow transformer/choke cores.  With their flat tape winding, why are they so hot, was there a reason a toroid could not be used, or larger core to reduce the thermal concentration?  And then I found the video on how to replace the transformers, yikes, they go bad often enough to have a R&R video ?

[zoneblue]

If youre bored... given that you have two fairly similar arrays , i think youd be really doing the world a favor by trying to model the perfomance difference with different input Vmps. Anecdotally, weve heard this before, that higher Vmps leads to heat issues. But it would be fantastic to actually graph these efficiency curves. With this information in hand the string calculator could be modified to factor the controller losses at higher input voltages. And that would be pretty cool.

[offgridQLD]

I don't think there is much loss worth worrying about  in terms total  kwh production - output performance over the day worth chasing so that's not my concern.

Like I mentioned before it might only be 2 % or so. But 2% of 4000w is 80w and it's the heat from that small amount of power loss than can soak into a small aluminum case. The benefits of lowering it and not allowing the unit as a hole to get hot is the electronics inside should last longer.

Running my hands over the case front,sides and back. I feel there there is more heat coming from where area where the inductors are located (back side where it usually mounts to the wall)  The right side where the fets are located isn't to bad. My units are mounted on stand offs so I can run my hands behind the unit and feel the back of the case. I got a real shock the first time I did .

Utube video of system running under load and the classic temps
https://www.youtube.com/watch?v=5tB70IhgV40



Kurt

[pechan]

you could fill in the fins with epoxy to get the flat surface.

[offgridQLD]

I've also be quite intrigued about the 2 yellow transformer/choke cores.  With their flat tape winding, why are they so hot, was there a reason a toroid could not be used, or larger core to reduce the thermal concentration?  And then I found the video on how to replace the transformers, yikes, they go bad often enough to have a R&R video ?

Yes that was part of my concern in trying to cool the area where they sit in imbedded in the case. Given its also heating all the other electronics in the case infront of it. The fets are then trying to cool of into a heat sink (basically the entire case of the classic) that is all ready heat soaked from the inductor wast heat.


Regarding filling in the fins with epoxy to mount my chill block. As I said I am trying to get the heat away from the inductor area not the fet area where the fins are for now. Though a adhesive thermal mat  or epoxy on the back of the case is a option over milling a flat section. I am a machinist by trade to other than potential warranty issues from taking a light surface cut on the back of the case Its not a big deal for me. (the area where I want to install it is almost flat just that part of the logo embossing is in that location + removing the powder coating and having a perfectly flat surface to contact the flat chill block will work better.

Inductor location


location of water chill block on rear behind inductors

[CDN-VT]

Neat !!

VT

[toothy]

Here is what I was talking about in my post above. http://www.aimradiantheating.com/store/FinTubeBaseboard_3_4_Element_Only.html
These things are all over up here used as people go to full floor heating.

In the summer 10 watts wouldn't worry me but in the winter when the sun gets up over the mountains at 11+ and disappears before 3, ten watts makes me crazier than normal.

Also I like the, keep it simple stupid, theory there is almost nothing to go wrong, famous last words.

Wade

[Halfcrazy]

The reason for the R&R on the Inductors was not that they go bad but we had a few that where shorted internally in a small window of serial #'s do to the manufacturing process. So we made the video for some of our dealers.

[CDN-VT]

Reading this thread & my post here :
http://midniteforum.com/index.php?topic=2162.msg20334#new


If it's clean cool air , This should also work , it's the water vapour , wet Pacific North West wetness that Im worried about air flow 9-20 hrs 7 days 365 . (Run from different solar sensors )

VT

[offgridQLD]

Well some very rough preliminary test of the system .More just to see if the pump,fan and radiator function  and purge the air from the system.

Had a play just making firm contact to the case with the chill block to cool the inductors.

I need to take that unit offline and machine a small smooth patch to get good thermal transfer.

Anyhow having some fun video link bellow

Initial trial of the parts.
https://www.youtube.com/watch?v=JXK1U8AChYshttps://www.youtube.com/watch?v=JXK1U8AChYs

More to come.

Kurt