Issues with boost control - very unstable

[oneten]

Hi Guys,

Our Toyota MARK 2 with a 1.5JZ engine went on the dyno for full power tune last night. For the most part everything went very well, the EMU tuned up great and car made fantastic power. We did have a fair few dramas with boost control however hoping the guys here can give some advice so I am not barking up the wrong trees.

The engine is a 1.5JZ using a 3L 2jz bottom end with 1jz head. Turbo setup is a Garrett GTX3582 on 6boost manifold & Turbosmart 50mm gate.
Wastegate currently has 15psi of springs in it. Our aim was 20psi of boost using the MAC valve with the EMU to control boost.

First attempt:
We had the gate plumped up to the boost solenoid in 'external gate" fashion, where both the bottom nipple (under wastegate diaphragm) and top nipple (above wastegate diaphragm) as utilized. Pressure is always fed to the bottom diaphragm (to open the gate) and the MAC valve controlled by the EMU can feed pressure into the top part of the gate diaphragm to make the gate harder to open, therefore increasing boost. I hope that makes sense (for those who dont often use this setup).

In this configuration, the boost curve was extremely stable on gate pressure and adding in boost in Open Loop mode by increasing duty cycle did work... but only to a degree. Boost curve was extremely sensitive to the point where 1% increase in duty would see a massive 5psi in boost difference, and was not stable at all.

Next we tried closed loop, which was far worse and almost always overboosted to the point it would hit boost cut (25psi).

Second Attempt:
Changed the wastegate plumbing so the MAC valve was in "Bleed off" mode, using only the bottom part of the diaphragm but using the MAC valve to bleed off air pressure causing and increase in boost pressure. This is the way most people plumb up the gate.

In this mode, open loop was immediately more stable but still not nearly good enough. We would still see variances in the boost curve each run without even changing anything.
Closed loop again was horrible and we must have played around with it for a good hour, chaning PID settings & other things before giving up on closed loop.

We got open loop to a point it was hovering around 20psi, overshooting boost slightly at first then dropping back down again.
As soon as we got it off the dyno, it was ok for a short while on the street but now has started to over boost yet again and we are back on running just wastegate spring boost pressure.

I monitored the duty cycle and it appears to be doing what we tell it to but boost just isnt right.
Could this be a firmware issue? we are running 1.172exp1.
It is almost as if the EMU is doing say 38% duty but the solenoid is actually doing something else.

Anyone have any ideas or pointers as to what is happening here? The manual is somehwat devoid of information about the various boost tables and how they affect the boost controller.
I will collect the current tune and a picture of the final dyno graph to show what is happening as soon as I have a chance, but I guess I would just like to know if anyone else has had similar issues and if I should expect the EMU boost controller to be a lot more stable than this.

Thanks,
Hayden

[oneten]

Hi,

Attached is a log from a pull in 3rd gear showing the unstable boost, with 34% valve DC.

Also attached is a log with wastegate pressure only, you can easily see how much it varies with boost control kicking it and how much more stable it is without.

Also have attached the current tune file.

If anyone has any ideas it would be much appreciated. For now we have had to turn the boost down because trying to target only 140kpa sees us hitting boost cut at 260kpa constantly.

Thanks,
Hayden

[Swift turbo]

Hi..in your boost target table #1 you have 2.4 bar as your target and i think that's what you are hitting.Might be worth switching on DC error correction too and adjusting the DC table to make the error correction zero?

[ren]

You need your target to be closer to the actual boost if you want the control to be effective.

Looks like you need some fuel cut. You're going to cause engine problems throwing that much fuel at the engine.

[oneten]

Thanks for the responses guys.

Hi..in your boost target table #1 you have 2.4 bar as your target and i think that's what you are hitting.Might be worth switching on DC error correction too and adjusting the DC table to make the error correction zero?

Assuming the boost map is in absolute pressure (like all other MAP values in the ECU) we are targeting 140kpa (1.4 bar) boost above atmospherics which is 240kpa absolute, so this is the correct target boost, as we are trying to hold 140kpa (20psi).

We are also using Open Loop control at the moment and my understanding is that the target boost map is only used for Closed Loop. Is this correct?

As for DC error correction, it is not enabled at all at the moment, we are simply using a solenoid duty cycle value to raise the boost levels.

You need your target to be closer to the actual boost if you want the control to be effective.

Looks like you need some fuel cut. You're going to cause engine problems throwing that much fuel at the engine.

As I mentioned above we are currently only using Open Loop, and we are targeting 140kpa boost (above atmo, 240kpa absolute) so I am not sure this should be an issue?
To be clear, we obviously had the target boost map more accurate when using closed loop, but if it has nothing to do with operation of boost control in open loop it becomes a non-issue.
Every other brand of ECU I have ever tuned has operated in this same fashion, unless I am missing something here?

As for the fuel cut, I am not sure what you mean. The AFR's are pretty much spot on, we are seeing 590hp on 20psi with pretty much spot on 12.2 AFR to redline. Starts to raise a little in the last 1000rpm but mainly because we are out of fuel pump. We are using overrun fuel cut etc.

[Karel]

Sorry, but what i read = open loop unstable / switched to closed loop = still unstable.

Once you noticed your boost control in open loop is going to be unstable, you need to solve all problems to make it stable as hell. Once you achieve your goal, switch on the closed loop and allow just small changes for compensating altitude, atpospheric pressuer diferences etc..... no electronic can solve harware problem.

This is just my point of view.

I dont use MAC valves, because i found them to have too big holes in it. If you will soften the response, even OEM factory (for example Mitsu EVO) uses restrictor in lines, to slow down boost control response. 1mm will be the start point.

[oneten]

Sorry, but what i read = open loop unstable / switched to closed loop = still unstable.

Once you noticed your boost control in open loop is going to be unstable, you need to solve all problems to make it stable as hell. Once you achieve your goal, switch on the closed loop and allow just small changes for compensating altitude, atpospheric pressuer diferences etc..... no electronic can solve harware problem.

This is just my point of view.

I dont use MAC valves, because i found them to have too big holes in it. If you will soften the response, even OEM factory (for example Mitsu EVO) uses restrictor in lines, to slow down boost control response. 1mm will be the start point.

Thanks Karel.

This is exactly the approach we took with it.
Open Loop was unstable as hell with the gate plumbed up with the top&bottom diaphragm's, and was only midly better with a bleed off setup connected to bottom of the diaphragm only. Still not anywhere near as stable as I would like. We switched to closed loop only to see if it was any better.

We did around 30 power runs trying to get open loop to stabilize changing everything we could think of but wavering between 19psi, spiking up to 21psi and coming back down to 19psi again was as good as we could get it. Then it almost immediately began over boosting again with no changes to the map.

With previous ECU, a Haltech PS2000, this setup sat perfectly on 19psi to redline using the same type MAC valve and no restrictor, but I see the sense in what you are saying, and to be honest my next step was to try a whole new valve. I've never had to use a restrictor previously however out of many similar setups, the main difference with this car is the EMU (first one I have tuned) so I assumed it was the anomaly.

Thanks

[Karel]

Again, its not ECU related. In open look the ECu just uses fix Duty cycle and it cant make somtehing wrong. Try to read out your previous setup like frequency, duty cycle etc and use the same values on EMU. Should work the same. You have for sure flyback diode on this output. I found 2 times in last 6 moths trouble on boost control ... 2 times MAC/Apexi boost solenoid , solution = put in new Pierburg valve and solved.

IF you have it stable on open loop, than go closed loop. If you have it stable on open loop and unstable on closed loop, than you need to lower dont integral part of PID.

[gregpe]

HI
what frequency of valve you running?
are you using flyback diodes?

[RHD]

DC error correction works in both open and closed loop so that will help even it out. You can set quite an aggressive slope if you need to.

It sounds like the valve is not coping with the DC requirement. Try playing with the frequency sometimes dropping it down a bit helps give the valve a chance to hit the full open and full closed position without getting too close to the dead time ... if that makes sense

you can find the max operating Hz using the user PWM table... for example set the DC to say 10% then play with the frequency until you find the point where the valve starts or stops clicking loudly and that will give you a maximum frequency that will allow DC range of 10-90% etc.

[oneten]

Thank you for all the input guys.

We were using a MAC 35A-AAA-DDBA-1BA with 33hz frequency, which is identical setup to Haltech. Yes we have the flyback diode wired in.

This is a brand new valve and I am begging to suspect the valve might be faulty. After some more testing is seems like the valve is "sticking" during operation, which would explain my issues perfectly.

Will try playing with the frequency and another soldenoid and see how that goes.

Thanks,
Hayden

[silikone]

you dont need flyback diode at this frequency

[KirillZl]

The PCV Valve is a check valve. You shouldnt need two oil catch cans or two "check valves".

you do when youre running 12psi and leaking boost

[WHPZach]

One thing to be aware of, DC correction is always enabled, whether you're in open or closed loop. Check your logs and verify that the duty cycle you're commanding is what the ECU is actually outputting.

[PSI-Motor]

Set 19hz
Map filter 3

[A.Dutsov]

I found this on internet ...
maybe will help somebody
approx Duty range %
3port MAC-5.4 watt
HZ - Duty%
10 7-90
15 10-89
20 15-86
25 17-80
31 20-80
40 26-68

[gips]

Whats the catch with teh Map Filter 3 ? I understand that it will filter the signal more aggresively and possibly eliminate wrong map readings under high boost. I have the same problem mentioned here, i just can't make the closed loop work, the spikes are too big, for some reason , so i am using the openloop, which is ok, but can benefit from small corrections from PID. Unfortunately ,if i set the PID settings to like 3:1:0 it does not correct fast enough. If i set to like 6:2:0 it overcorrects, so i was wondering ,if this Map Filter will help resolve this.

[RHD]

you dont need closed loop if you set up the ref and target tables well you can use some quite aggressive DC error correction slopes (assuming the valve is working corectly!)

this is a little ca18 we did last year running a 3 port MAC valve and a very low pressure spring in the wastegate, look at the manifold pressure!

[gips]

This looks rock stable. I think i understand waht you mean. I will try it as soon as i get rid of the slipping clutch

[Zaxos]

Hello from a new member. Anyone that has a target pressure setup working carres to sear the pid setup with us? Along with the frequency part???