zps2004

So I was really trying to figure this out by finding information on here and being able to figure it out, but I haven't been able to yet so I figure I would ask here.

I have a GT2560 on my 97 Miata with a DIYPNP and the Megasquirt boost control solenoid. Nothing really different or fancy. Before I try to get closed loop boost control to work I figure I would get open loop to work correctly. (I have messed with closed look but with no success yet).

So I can get the car to make more boost than the wastegate, but the quicker I get on the gas from being off boost the more it overshoots. Even if I have the entire duty cycle table on the same duty, it still overshoots the quicker I get on the gas. If I take it easy it will hold 10psi no problem. As the turbo spools in the lower RPM's it will hit 10PSI no problem, maybe 0.5psi overshoot, but if I get on it quickly it hits the cut at 14psi. I would like to be able to get this fixed before moving on to higher boost and closed loop.

I have the table set up so 100% duty is wastegate pressure. The only other thing I have noticed is that the overshoot is worse on cold mornings compared to the afternoon. On a cold morning I can't get it not to overshoot while getting on the gas or it spooling. Is that a double negative?

Any thoughts would be appreciated. If you have questions feel free to ask.

concealer404

iTrader: (3)

Is there a setting that would be similar to "gain" on a "traditional" electronic boost controller? Sounds exactly what happens when you have the "gain" too high.

zps2004

Not as far as I know. In open loop when you are defining the duty cycle there isn't a gain since the duty is defined. That is why am so confused as to why it is acting like it is.

soviet

iTrader: (10)

uh, post your boost duty table?

Braineack

iTrader: (62)

why arent you using closed loop?

FoundSoul

iTrader: (5)

'Gain' would be equivalent to adjusting PID parameters in closed loop mode. In Open loop, it's all in the tune. You have to tune the table to command the proper duty cycle to bring the boost on as quickly as possible, but to back off the duty cycle commanded as the boost comes on so that it reaches your target without overshooting. Then you modulate the DC as the RPMs increase to keep it on target.

The main negative to open loop is it will boost higher when it's cold out, as you're command a duty cycle and not a boost pressure target.

I'd run 100% DC at TPS above maybe 30% and your leftmost RPM column as well as the one next to it to the right. That second column should have it's RPM set to just before boost climbs at WOT. Then place your next rpm column at the rpm where the turbo is fully spooled up, and in that column you'll want to quickly decrease the commanded DC (if you're inverted where 100% is wastegate pressure, you'll increase the number, to decrease the boost controllers attempt to increase boost).

Once that is set right, you shouldn't overshoot any more, and you can tune the rest of the columns to behave as you wish, in most cases the goal being steady boost pressure all the way to redline. Some people like to let it taper off to reduce impact on the engine, others like to taper it up towards redline to make the torque curve and RPM continue to climb as long as possible...

zps2004

I can get steady boost if I get on it slowly, and I have good spool. 100% duty is wastegate pressure, and to get 10PSI at 4K I am running about 65%. I spool to 10 at about 3200. Now at 4K if I get on it slow it is fine, but if I get on it quickly I overshoots a lot. I have tried having the different throttle positions below 100% be 100%DC, 65%DC, and less. It all seems to overshoot if you put your foot in it quickly. I will mess with it more tomorrow probably, or I might have time tonight. I will post up my current duty table as well once I get home from work.

I have tried closed loop, but I haven't had much luck there either. I started with P=100 I=0 and D=0, and slowly moved P down until I started getting over wastegate pressure. At P=8 I would still have wastegate pressure, and at P=7 It would overshoot right to 14 right where the turbo would spool. I am guessing that I might need more I in that area to see what happens. I might mess with that more this weekend as well.

Also, for closed loop, does it still work well if you have creep? Over 4000 I start to creep to 11psi at redline. I am planning on running 12PSI though so I am not worried about the creep.

FoundSoul

iTrader: (5)

Don't tune it for part throttle and then note it's overshooting when at WOT. Tune it for WOT based on the suggestions above and it'll be fine at partial and if you want to see it different you can tune that out later.

While tuning WOT I'd just put the same numbers in each column and not do anything based on TPS, then once that's right you can adjust based on TPS for lower throttle inputs.

Boost creep shouldn't matter so long as it doesn't creep above the amount of boost you intend to run. It will likely effect what your table looks like as you'll need to drive the EBC less and less as it starts to creep and the EBC will barely be giving it anything over wastegate pressure at redline.

zps2004

I have done what you mentioned and at WOT, it spools great and holds well. I was messing with partial throttle since the overshoot is dependent on how fast I go from 0% to WOT, so I was seeing if different partial throttle positions would change the overshoot. That was my thinking at least.

Braineack

iTrader: (62)

im still trying to figure out why we are futzing with open loop.

FoundSoul

iTrader: (5)

When you tune it as I described for WOT, and then got WOT from a stop, it puts the boost where you want it and doesn't overshoot?

The answer is in that table. The EBC valve is going to respond exactly how that table is setup. It helps to visualize the conditions that are occurring and anticipate what the EBC will need to be told to do in different conditions to accomplish the goals you're after.

zps2004

For FoundSoul, WOT from a stop it is fine. I had the car on the dyno on Tuesday and on pulls from 2000rpm it was fine. On the street shifting from lets say 2nd to 3rd at redline when I get back on the gas it will overshoot.

So should I just work on getting closed loop to work not open loop? From reading a number of threads people say to get open loop working correct before going to closed loop. I know that I have the duty set correctly where you lower the duty to get more boost. I also wasn't sure how well closed loop could work with creep, but since that isn't an issue I would like to work on it more.

zps2004

So we will have to come back to this at a later date. Yesterday my car started loosing sync for some reason, so I need to look into that first.

zps2004

So for some reason when I have boost control turned on it looses sync randomly. Both in open loop and in closed loop. If I have it turned off the car runs fine. Any ideas? I have had it on for about a week before this happened with no problems and I have had the megasquirt for over a year. It has to be warmed up, and sometimes it just randomly looses sync. Sometimes it is as a particular RPM.

JasonC SBB

Overshoot happens because the wastegate can takes time to fill up and move the swing valve to the position your are expecting it to go to.

The only way around this with your setup is to run closed loop, and the 'D' term will reduce overshoot.

Braineack

iTrader: (62)

whoa whoa whoa, why would we ever want to use closed loop?!?!

zps2004

I am getting there with closed loop, I just need to fix the sync thing first. Maybe I need to adjust the VR conditioner? I am also going to look through the boost control circuit to see if something looks odd with my soldering.