SDU high current consumptions?

[tom91]

Isn't max temp = 85C?

Yes so if you park when the inverter is at 50C things wont get hotter. The heat will start to dissipate into the housing.

I still recommend measuring the temperature increase across the inverter if you are looking to resolve this, that should show you how much heat is going into the coolant.

[johu]

Isn't max temp = 85C?

This default value applies to inverters that measure temperature somewhere on the heatsink. This temperature is low pass filtered by the heatsink mass and offset from the igbt core temperature.
Now water cooled inverters don't have much of a heatsink so instead they deliver the temperature right next to the igbt core. This fluctuates much quicker and is higher than the hs temp for the above reason. Igbt chip temperature can is allowed to be around 150°C according to the data sheets but most lookup tables in the OI firmware only reach to 100°C. This should be extended.
Someone care to open an issue or even fix it? I don't have access to my github during vacation

[alexbeatle]

I still recommend measuring the temperature increase across the inverter if you are looking to resolve this, that should show you how much heat is going into the coolant.

Do you mean to log/graph tmpm? Sorry not quite sure what do you mean by this, the inverter is sealed.

[tom91]

No as in add temp measurement sensors, before and after the inverter. You can already see the "heat" in the inverter. You need to know coolant.

[alexbeatle]

No as in add temp measurement sensors, before and after the inverter. You can already see the "heat" in the inverter. You need to know coolant.

Any advise on the sensors?

I'm considering investing into the vacuum filler. Is -70..-60cmHg ok for the SDU inverter to not cause damage or should be smaller?
https://service.tesla.com/docs/Model3/S ... %20closed.

[alexbeatle]

So I changed the config of my cooling system - no more 90 elbows, complicated valves for series\parallel loops. In essence 1 common bottle y-splits the coolant to 2 pumps (electronics and batteries) creating 2 independently controlled sub-loops, that join together before going through a radiator, which then returns the fluid to the common bottle.
Also added added a valve as a bypass for the charger (can control how much from to the charger, how much bypassed) added inline temp sensors on the input and output of the inverter, pulled vacuum on the system and refilled. Result - same overheating issue.
Attaching the video of my coolant flow bottle->charger+bypass->inverter->gearbox->radiator->bottle

Temp sensor readers. Unfortunately in Farenheight, but as I drove 3 laps up and down the hill at aprox. 25mph with stops on the stop signs, the coolant temp only went up to 32C/91F. The inverter temp however went as high as mid 50sC.

Attached also my latest parameters of SDU (pretty close to my liking).
Running out of ideas how to solve this...
Any help is really appreciated..

[jrbe]

If you park and monitor temps, do the 2 coolant temp sensors match the displayed inverter temp? It's worth checking a few times as it cools too. To see how the range is.

Mid 50c isn't really hot.

[alexbeatle]

The 2 sensors start to match quite fast after park, but the inverter temp takes long time to come down.
Even now after hours of driving, the ambient temp is ~18C, the two sensors report 65F/18.3C, but the inverter is reporting 21.56C.

[johu]

The temperature input isn't that accurate

[alexbeatle]

The temperature input isn't that accurate

Hmm.
1. What happens when the temp reported by the IO board is >85C (my limit set by the parameters)? Does the drive start to de-rate the output?
2. Is that when the inverter becomes too hot and burns?
After driving, when I touch the inverter side casing, it's warm to touch (warmer than ambient), while the temp sensors return to approx. ambient temp after a minute or two.
3. It's a bit confusing, if somebody could help understand, is 85C the limit temp of the inverter's igbt-s or the limit temp of the coolant, i.e. the cooling loop material inside of the inverter section is not rated for too high temps.

[johu]

85°C is a heatsink temp limit, i.e. when the temp sensor is mounted some distance away from the IGBT. On the SDU though we see the chip temperature and this can actually be much higher, like 130°C or so.

Trouble is the sensor range ends at 100°C for the SDU. So setting the limit above 100 will result in no temp derating at all.

[P.S.Mangelsdorf]

Hmm.
1. What happens when the temp reported by the IO board is >85C (my limit set by the parameters)? Does the drive start to de-rate the output?

Yes, it will derate.

[alexbeatle]

This thread should probably be called SDU issues, issues, issues, or similar. I appreciate your patience with me folks.

I drove around today more and at up to 60mph. There's only slight improvement in cooling at higher speeds, but have reached tmpm=80C.
What I noticed is that the radiator gets warm to touch at those temps, while the ambient was ~16C. Radiator doesn't get enough airflow!?
Running the fan helps, I can feel the cool air gets pulled through the radiator, but my fan failed today, so not sure how it will improve. I set the motor pump to 70% permanently, but without luck. The hose from the pump to the inverter becomes hard as it's trying to squeeze that much pressure when converting from 3/4" hose on the pump to 5/8" on the inlet of the inverter.
Also I notice the case of the inverter becomes very warm to touch in this area (within red rectangle, lower is cool too), but the area where the igbts are is cool to touch. Is that how it suppose to be?

I will do more plots tomorrow to be exact, but I see my voltage sags quite a bit - ex. BMS reports 46% when stationary, ~36% when accelerating, ~43% at speed, back to 46% when stopped. Perhaps cooling and parameters are related!?

[tom91]

Radiator doesn't get enough airflow!?
Running the fan helps, I can feel the cool air gets pulled through the radiator, but my fan failed today, so not sure how it will improve.

Your radiator is mounted where? You realise any radiator needs air moving to remove heat. This is basic thermodynamics, mass to mass movement of heat energy based on delta temps.

What is the temperature delta across the radiator?

[alexbeatle]

So originally vw t2 had no radiator at the front of the vehicle. It has these side ears that guide airflow into the engine compartment. I used that as the guide for mounting the radiator. The fan is meant to pull that air to assists air entering through those side ears. My thinking was that there would be enough airflow when driving at speeds >20mps, so do not need to run the fan.

I will add temp probes on the radiator to measure the temp.

Side ears and the air channel where the air is meant to travel down the engine bay..

Radiator and the bottom of the air chambers coming from the ears at the top...


This is how the radiator is positioned looking from under the car...

[tom91]

Really buddy, you every look at an "aircooled" engine. They are forced cooling. http://www.volksbolts.com/faq/cooling.htm

[jrbe]

Your coolant bottle has a y on the returns. Are you sure coolant is making it through the inverter well? The pressure you feel on the inlet is a good sign but the temp difference on the red square area is odd. Aluminum conducts well so there's something going on there.
I still think there's an air pocket, some big bypass, or a restriction in that area.
Does running the pump slowly help at all?

I wonder if the pump is cavitating and pushing air into the inverter. Do you have anything to catch / purge air at the inverter or on the way there?

If your radiator isn't getting hot, lack of airflow isn't the cause of your current temp. issue. It may pop up later though as you fix issues.

It's easy to forget or not realize this but there is a low pressure area behind the vehicle, as long as that is able to suck air through the radiator it's getting at least some airflow pulling in from the sides.

[alexbeatle]

Really buddy, you every look at an "aircooled" engine. They are forced cooling. http://www.volksbolts.com/faq/cooling.htm

Well this is embracing.. I must run the fan.

Does running the pump slowly help at all?

I will double check.

I wonder if the pump is cavitating and pushing air into the inverter. Do you have anything to catch / purge air at the inverter or on the way there?

Practically no vertical changes from the pump outlet to the inverter inlet. Not sure why air would trap there, but good point to observe...

Ok plan of action
1. fix the fan controller - fails when I take off, but runs fine when stationary, may be sag in 12vdc supply to the controller...
2. observe deltT on the radiator
3. introduce clear tubing per Gregski's suggestion. I think I will add it on the output of the inverter to see the flow leaving the inverter. May also do that on the input.
Thanks everybody. At least there's a plan to go forward

[alexbeatle]

Fixed the fan controller.
Inverter temps still go high. Didn't drive much and <45mph, but was in 65C-70C.
The difference with the fan - faster temp drop when stationary. Waiting on the thermal camera to see delta, but radiator is cool to touch, while it was warm before.
Observing the temp on my gauge, I find the recovery is faster when the fan is pulling air (the way I setup initially, to pull through the car side ears). When I switch to pushing - recovery is slower.

Does running the pump slowly help at all?

Double checked. It doesn't.

[jrbe]

If the radiator is cool while the inverter is hot and the pump is pumping you have a bad heat path from the inverter somehow. Where does the heat "stop" after the inverter? A hose, connector, etc?
Is the coolant out of the inverter warm / hot / near the reported temp of the inverter?

Possible you have a paper towel / rag or something you missed during assembly in the feed hose to the inverter that's restricting it? It may have pushed into the inverter heatsink too.

[alexbeatle]

image.png

I believe this is the area that I refer to here that gets warm. Looking from the inside is that some kind of isolation monitor? Suppose to be warm?

I can see now why the inverter needs to be vacuum filled, there’s a section for the coolant that’s above the outlet port.
I can also see why the inverter would limit the coolant flow. Those are some hard turns inside past the inlet/outlet ports.
I’d be curious to know if anybody actually had luck with SDU coolant flow inverter -> gearbox exchanger, as opposed to to the way Tesla did it by splitting the flow.

The outlet holds the coolant sensor? I thought tmpm was reading the igbt temp, but actually this temp probe.

[johu]

The outlet holds the coolant sensor? I thought tmpm was reading the igbt temp, but actually this temp probe.

There are 3 IGBT sensors which we display the maximum of (tmphs). Then there is a fluid and a stator temperature on the SDU and we only display stator temperature (tmpm). So the fluid sensor is ignored.