Feature request: better resolution in boost bias table

[elaw]

I've recently starting fooling around with boost control using a 4-port solenoid, and it's working pretty well but the MAP is very sensitive to boost duty changes... at one spot in the log a 5% change took me from 180 kPa to 230!

So I'm wondering if there's a way the resolution could be increased in the bias table? I note that it's stored as a U08 and since the PWM value can't ever go above 100, even doubling the values to give 0.5% resolution instead of 1% would be helpful and the values still could be stored in 1 byte.

[jsmcortina]

I've recently starting fooling around with boost control using a 4-port solenoid, and it's working pretty well but the MAP is very sensitive to boost duty changes... at one spot in the log a 5% change took me from 180 kPa to 230!

How do you have this wired?
In some testing with my idle valve, I found that just using an MS3X output gave "grainy" response. Adding the 1N4001 diode across the valve made it far more responsive.

James

[elaw]

Pretty much the conventional way... one side of solenoid to +12 and the other to the MS3X boost output.

I suppose I can try adding the diode, but it controls quite well just as it is... it's just that a 1% change in PWM makes a big difference in boost. 4-port setups are known for that problem, as they have a lot more authority than 2- and 3-port setups.

Edit: I was about to write that to some degree this is a red herrring, as I thought internally there was more resolution (because MLV displays boost duty with 1 digit after the decimal) and that closed-loop would rectify some small errors. But I just realized that's not the case, so increasing the resolution may not be as easy as I thought?

[dontz125]

This fellow replaced the normal 4N4001 diode with a Schottky diode, and got better response. Th blog is getting old, and the code used here is also beyond obsolete, but I have to imagine the premise is still valid.