search

rusefi / hellen-NB2-issues

0 stars 1 forks source link

Pin Error While driving around TC4427 maybe alternator damage in USA #23

Closed BeerMoneyMotorsports closed 2 years ago

BeerMoneyMotorsports commented 2 years ago

https://rusefi.com/forum/viewtopic.php?f=19&t=2190

Related to https://github.com/rusefi/hellen-NB1-issues/issues/25

shipping Matthew Kennedy the failed board via usps for inspection.

BeerMoneyMotorsports commented 2 years ago

B7EE37D6-8448-411A-A13E-249B95FE0BEA

rusefillc commented 2 years ago

Some relevant notes on https://github.com/rusefi/rusefi/issues/3558 There is an opinion that we need https://github.com/andreika-git/hellen-one/issues/118

mck1117 commented 2 years ago

I have the board in-hand now.

Notes:

I'm not sure how that happened, but a 1k series resistor (optionally a pulldown too for safety) would likely prevent it by avoiding whatever overcurrent condition caused the steering diodes to fuse.

BeerMoneyMotorsports commented 2 years ago

@mck1117 thank you. Can you Write a ticket for us to add the resistor to future boards? I need to order nb1 immediately. Is it too soon to make this call while you are still diagnosing everything?

Do we need to find a way to add this to current boards in customer hands?

mck1117 commented 2 years ago

is this not that ticket? https://github.com/andreika-git/hellen-one/issues/118

mck1117 commented 2 years ago

It will certainly not hurt. Many proteus/mre/frankenso boards have been driving around with those resistors in place.

BeerMoneyMotorsports commented 2 years ago

@mck1117 thankyou. Sorry I overlooked the ticket.

mck1117 commented 2 years ago

Actually I take that back - Proteus/MRE have pulldown resistors on the input side, but no series resistor. They do, however, have a resistor on the output, also absent from Hellen.

BeerMoneyMotorsports commented 2 years ago

Per @mck1117 reverse engineering OEM ecu. He found a 1k resistor in series and a 100nf to gnd. Picture below to show circuit. This is being added to NB1 production. NB2 is being tested with updated circuit.

image

related ticket- https://github.com/rusefi/hellen-NB1-issues/issues/25

BeerMoneyMotorsports commented 2 years ago

Modified board to test.

@mck1117 or @andreika-git , i would need some help with how to torture testing NB2.

Customer experiences failure in about a week of normal driving. Customer is willing to test long term again, however we have to test prior to his testing.

Current test plan. @rusefillc will be shipping h72c to torture test(if he has stock).
1- Using a 2001 miata, run unmodified board on customers tune. Log several drives with alternator pid debug. Potentially 3-4 days of driving? 2- Modify some settings to purposely fail a board. This might be dangerous on a customers vehicle in the case it damages the chassis circuits. Bench testing a viable option?

nb2 alt mod

BeerMoneyMotorsports commented 2 years ago

This ecu with updates has been on for about 200 miles of driving. Alternator seems to respond appropriately. Noticing about 1.3v total swing with headlights and AC running. Most of the time you see 14v +-.2v

BeerMoneyMotorsports commented 2 years ago

Board was tested for 250 miles then returned to the customer for further testing. No failures during my testing which mainly consisted of lots of idling and driving. This is with AC, lights, wipers, etc all running.

BeerMoneyMotorsports commented 2 years ago

This board has 750 miles on it with zero problems. Fix has been applied to rev nb2 rev E and then NB1 rev B. Miata NA does not have alternator control circuit. so 90-95 and 96/97 will have the 0.3 output module with protection but not a specific alternator control circuit.

BeerMoneyMotorsports commented 2 years ago

user has in excess of 3000 miles on ecu. no issues.

*note that this should not be combined with https://github.com/andreika-git/hellen-one/issues/119 as it will result in poor alternator performance.