EEVblog Electronics Community Forum

General => General Chat => Topic started by: NiHaoMike on October 13, 2012, 05:48:06 pm

Title: Interesting ECU teardown
Post by: NiHaoMike on October 13, 2012, 05:48:06 pm
http://barsomauto.com/transmission_diferential_files/Nissan/xterra/Xterra_TCM.htm (http://barsomauto.com/transmission_diferential_files/Nissan/xterra/Xterra_TCM.htm)
That's definitely an unusual way to design an ECU. Just pieces of bare silicon assembled on a ceramic substrate and hermetically sealed. Then the enclosure it was put into wasn't waterproofed properly so it failed anyways...
Title: Re: Interesting ECU teardown
Post by: rollatorwieltje on October 13, 2012, 06:14:37 pm
Bosch ABS computers use the same technique. Those bond wires look very fragile.
Why don't they use a conventional PCB and pot the whole thing? Designed to fail?
Title: Re: Interesting ECU teardown
Post by: tom66 on October 13, 2012, 06:22:17 pm
ECU must survive 150-200°C engine temps and lots of vibration from the engine.

That being said I worked at a company that made down-hole oil well equipment which used conventional PCBs but high reliability solder, the equipment was rated to 220°C for one year I believe.
Title: Re: Interesting ECU teardown
Post by: FenderBender on October 13, 2012, 07:13:36 pm
Bosch ABS computers use the same technique. Those bond wires look very fragile.
Why don't they use a conventional PCB and pot the whole thing? Designed to fail?

That's what I thought. I took a peak inside my car's ECU which is a 1993 Dodge Spirit. Now granted it's not located in an extremely hot part of the engine, but it's just an FR-type board suspended in about 2cm of potting compound on either side. It's last almost 20 years like this.

Those bond wires scare me also...Why didn't they just have a PCB that extended all the way out and then used through hole pins instead. Wouldn't that have been easier anyway?
Title: Re: Interesting ECU teardown
Post by: mikeselectricstuff on October 13, 2012, 09:56:26 pm
Hybrid modules like this are quite common in high reliability power applications - one reason being very efficient thermal coupling. Lack of packaging also reduces mass, making it more resistant to vibration.

When they fail it's not always Game Over...
www.youtube.com/watch?v=DGqH8LGjZdY (https://www.youtube.com/watch?v=DGqH8LGjZdY#ws)
Title: Re: Interesting ECU teardown
Post by: SeanB on October 14, 2012, 06:02:44 am
Hybrids have a big advantage in that they can be made as bare ceramic plates and have all the testing and trimming done by laser cutting resistors during test, and then they are guaranteed to be stable cal wise for at least 30 years. You do all the testing to week out defective units at the board level, before the expensive encapsulating stage. Even if the hybrid ceramic costs $50 it will be the cheapest part, as it is assembled into a $1000 engine controller. Those get tested quite well during assembly ( aside from using bad capacitors which fail in 5 years with monotonous regularity) so that they are at least as reliable as the mechanical units they replace but provide enhanced functionality.
Title: Re: Interesting ECU teardown
Post by: krish2487 on October 14, 2012, 12:50:23 pm
It is not scary at all!!

When you consider that a 300A 1200V IGBT module has just those kind of bond wires.
and they perform quite well.

potted with a transparent sticky gel and bare bond wires.

and these power devices ARE quite rugged, and they get HOT.

Now take this TCM, it is subjected to thermal stress, but not generated internally (I assume so , because the die looks more like a custom microcontroller and several medium power drivers but nothing high power to cause significant increase in the internal die temperature)

I think that failure is more likely to be a passive failure (tantalums) rather than a silicon failure due to overtemperature
Title: Re: Interesting ECU teardown
Post by: SeanB on October 14, 2012, 02:02:57 pm
Hybrid most likely did not have damage, more likely is that water got into the assembly, and picked up ionic debris from the oil ( it does nothing in the oil until it gets moisture in it) and caused a short between pins outside the hybrid assembly with permanent supply. If you look the hybrid stayed sealed in the housing, inside which it was very well protected secondarily against contaminants by the Parylene coating over it for vibration and mechanical shock absorption.
Title: Re: Interesting ECU teardown
Post by: NiHaoMike on October 14, 2012, 03:41:02 pm
If they took the extra step of conformal coating the connections outside the metal can, they might have been able to prevent a failure after the primary seal failed.
Title: Re: Interesting ECU teardown
Post by: SeanB on October 14, 2012, 03:50:20 pm
Oil insulated, until the radiator intercooler failed.