Fixes overactive and noisy brake pump during braking on Honda cars #413
Conversation
…ump activity and noise than necessary. Patch supplied by @riccardo from Comma.ai
|
Well, I tested this in my morning drive today. I do think it seems to have quieted down at higher speeds but I could be wrong... I'm also using it with Pedal and that is just all out noisy and terrible as always when you're at a complete stop. I'm not quite sure why it needs to press so hardly on the brake at a stop... is it just ON or OFF? Can we adjust the AMOUNT of pressure it takes to keep the car from rolling and just stop there? |
|
The BRAKE_PUMP_REQUEST signal is binary: ON or OFF. I haven't seen a pressure indicator signal. Without the pedal the difference is massive, especially when the car is almost at a stop and creeping. There is no pump sound at all during slow creeps. In any case matching Honda's behaviour for stock ACC is probably better than just turning the pump on whenever there is a non-zero braking applied. I agree that right at the complete stop OP steps on the brakes hard which causes the pump to go on because it is increasing brake pressure. There are wheel speed signals so I don't see why OP can't just use minimum braking pressure to keep the car at a standstill. The other thing that can be done is that if the car is close to a standstill, just don't turn on the pump.... any pressure in the pump should be sufficient to hold the car at a stop since it's already going ~0mph. |
|
This PR is solid. Restores brake noise back to factory. And it also got rid of a rattle I was hearing while the car was running... like a ticking that was constant while the car was on. Not sure if that is coincidence or not. |
|
I'm also playing with having the pump turn off at a stop and the brake_stopping_target (put less pressure on the brake when the car comes to a stop). OP puts a lot of pressure on the brake when the car comes to a stop causing the pump to activate and also is a bit jarring when the car starts up again. We'll see how my experiments go..... I'll report back. |
|
Really really excited to see your findings...
…On Oct 31, 2018, 8:07 PM -0400, kegman ***@***.***>, wrote:
I'm also playing with having the pump turn off at a stop and the brake_stopping_target (put less pressure on the brake when the car comes to a stop). OP puts a lot of pressure on the brake when the car comes to a stop causing the pump to activate and also is a bit jarring when the car starts up again. We'll see how my experiments go..... I'll report back.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.
|
|
curious as to why you removed method documentation in this PR.. |
|
@dekerr not sure what you mean? |
|
docstrings in hondacan methods were removed. I don't think they should have been. |
|
@dekerr: code changes came from Comma.... I just submitted the PR as per @rbiasini suggestion as I tested their patch. |
|
merged into devel in 0.5.6 |
ghost
deleted the
Prior to this fix, supplied by @rbiasini at comma.ai, the brake pump would always be set to ON any time the braking value was non-zero while OP was active. This created more braking pump noise than necessary during braking and possibly overworked the pump. This fix mimics Honda's stock ACC behaviour, where the brake pump only turns on during times of increasing braking pressure. The result is decreased pump activity (and noise) during braking.