18psi

iTrader: (76)

Thanks for the link, I totally forgot about that thread. (and apologize for re-posting about the calibrations)

And yes, they are NOT shutting off with speed, no matter what speed. I've tested it to 95MPH, and they still didn't shut off. I posted up datalogs that matt looked over, and he was the one that suggested adjusting the programable on/off outputs

Ben

iTrader: (33)

Yeah, I meant fat-fingered on our end.

No, I'm saying to leave them at their defaults and not worry. The firmware options for ECU controlled fuel pressure were added more recently than development on this ECU began. Either way, this is a complete non-issue.

Ben

iTrader: (33)

It's not.

It's there due to customer demand because people (mostly Miata people) specifically requested a dedicated cooling fan control mode, with ability to restrict fan activation under defined VSS and TPS, and with a feed forward idle control bump.

You can activate a fan as a general purpose output, but you have less control over multiple conditional shut off strategies and no feed forward idle.

Ben

iTrader: (33)

Go to your project properties and make sure that port status is enabled. Then send an email with msq and datalog showing it. You can reduce the VSS trigger to something low like 30 mph so you don't need to speed.

Q: How do you *know* that the fan doesn't shut off at speed?

Also, your car has a 3 stage fan system. The cooling fan control would only shut off stage 1, and the other triggers (such as a/c) would keep fans activated.

richyvrlimited

iTrader: (1)

Isn't it?

It's sold as a PNP device for a Miata is it not? Or do you mean the PWM fuel pump isn't enabled?

18psi

iTrader: (76)

Here are the logs posted earlier in the thread. (post #101 if it doesn't go there directly) . I can email as well if need be.

https://www.miataturbo.net/megasquir...2/#post1227833

The green "fan" button stays illuminated. Also according to what Matt said, 1 of them shuts off and the other doesn't.

Ben

iTrader: (33)

PWM fuel pump is not enabled in the MSM base tune. However the option is there if someone wanted to get crazy.

Email please. Much easier for me to track.

arghx7

Question about the MS3 software: Why doesn't the feed forward portion of closed loop boost control have an air temperature compensation table? Same could be asked about open loop control. Boost control can vary drastically with air temp, and that's asking a lot from a set of global PID gains (scalar, no air temp compensation either).

aidandj

iTrader: (5)

Because theoretically closed loop PID should be able to take care of that.

Braineack

iTrader: (62)

theoretically? it DOES take care of temp.

open loop would require a temp correction table.

arghx7

That's asking a lot from simple PID feedback control. Air temp's affect on a turbo is very much a non linear system. When you use the same set of feedback gains in the winter as in the summer, you will inevitably have spikes or lazy response in one season or another. Look at stock PID control on a Subaru and they have feedforward IAT compensation, and air temp compensation for PI gains, and Gain vs Error tables as opposed to scalars. We don't need that level of complexity but a simple air temp table would help a lot.

I am trying to make the car as "set it and forget it" as possible and am running into the lack of air temp compensation here. Also, idle air control initial duty table can only be water temp axis or air temp axis. It really should have both. I had to lose my water temp compensation by flipping the table to an air temp axis. This isn't a big deal but if you want to idle at stock speed without any hot start surging it matters a lot.

Similar can be said for air density compensation. Air density compensation does not have a load component, when any old Hondata system has that. The Megasquirt only uses 1 IAT sensor, so when the IAT sensor heatsoaks you have to start making compromises. Either it runs lean on hot start and surges, or you richen it up on hot start and then go rich under boost. You can rely on closed loop but you're stuck with only 1 set of gains for idle, cruising, and boosting. You're always trading off stability vs response of the system. It requires lots of compromises.

Sorry if it sounds like I am complaining/trying to be difficult... I am working with Ryephile on tuning his turbo build. https://www.miataturbo.net/build-thr...d-80222/page5/

aidandj

iTrader: (5)

Megasquirt Support Forum (MSEXTRA) ? MS3 Development (View forum)

Where you should probably be bringing all this up. We just use their stuff. Well DIYAutotune develops, but most of us just use what they give us.

arghx7

Thank you for the link. I didn't mean to intrude on your thread.

Braineack

iTrader: (62)

If you cant tune the MS3 PID boost control (or idle) because of ambient temp, it's you, not the temps.

read heavily into that.

arghx7

it works fine, it would just benefit from more flexibility.

Braineack

iTrader: (62)

I'll give you that. :P


IF youre using an intial values table for CL, then a temp correction comp table wouldnt hurt, but it gets you close enough in most conditions that it's really not needed.

18psi

iTrader: (76)

By all means, ask any questions you want.

I think I remember you from Nasioc, and also have and tune multiple subaru's

Zaphod

@arghx7 - Please post all those suggestions and knowledge at the msextra.com link, I think we all could benefit from those things! Very good points you brought up...

richyvrlimited

iTrader: (1)

+1

arghx7

Alright, I'm digging up the screenshots I need to make a case for these feature updates.

First, the feedback gains:

The values aren't filled out, but you can see on this AEM infinity tuner default map that the boost control gains are tables of gain vs rpm, rather than global scalars:



The lambda gain is also based on rpm.



The screenshot below shows stock 02-05 16 bit ECU WRX boost controls that are relevent. The first two are initial duty cycle compensation tables for barometric pressure and air temperature. The bottom shows the proportional gain--it is based on error (difference between target and actual boost), so that you can adjust the sensitivity of the feedback depending on how far you are away from the target at any given time.



I'm not saying that it has to be one exact way, as I don't know the effort in developing the code, but I do know that plenty of alternative standalone ECUs and 10+ year old stock ECUs are not relying on global gains for lambda or boost control.

The last thing is the air temperature compensation table for a Haltech Platinum Sprint RE for an Rx-7. This particular code doesn't use VE look up tables (although some other Haltechs do), but the air temp compensation is clearly a 3D table with a load axis. This helps account for the heatsoak issue I mentioned earlier.