john banks

As title - any comments? If the EBC is not throttle position aware how can you still get good performance without a shed load of part throttle boost? Most EBCs except the HKS EVC Pro and the AVC-R do not seem to read throttle position do they?

Blackscooby

As you are aware the APEXI certainly is...

dfullerton

Blackscooby how hard was this to setup. i have thought of buying one but have been told the are a pain to setup? did you fit it?

Moles Dad

John Banks, I have a HKS EVC IV...I dont have any issues with part throttle boost, they do not have throttle sensor but I presume that due to the "vacuum" sensing the system is aware of the power requirement. (if you know what I mean).

Quite often I drive in "Shopping mode" and see -readings (boost) when burbling along, without (I think) part throttle issues.

Is this what you are talking about?

MD.

john banks

Trying to think through the vacuum issue - say at 4000 RPM you go for 1/3 throttle you will come out of vacuum so how does it then know the difference between 1/3 and WOT? If it cannot tell then presumably it would either give you suboptimal WOT gain and hence suboptimal spool up or it would go to full boost at 1/3 throttle by aiming for target boost? Any user of an unbled MBC on a small turbo will tell you you can reach maximum boost at 1/3 throttle at 4000 RPM, which presumably works the turbo harder as well as giving about 5% CO at say 17-18 PSI, which some would consider a little lean (unless you are called Cosie Convert who thrives on that sort of thing and burns us all away).

I suppose the way of finding out is if you use more gain does it give you a quicker spool up but not overshoot target at WOT? If so then there are more benefits to be had by sensing throttle position?

Or have these HKS boys been incredibly clever in some way?

NITO

That'll be the fuzzzzzy logic!

john banks

How does fuzzy logic help it to know throttle position? All it knows is boost - how can it derive TPS? (I have a sneaky suspicion it can't hence the extra features in the Pro).

NITO

It doesn't detect throttle position, but it reads vacuum and boost which is directly related to your right foot hence throttle position

With the evc when you accelerate on light throttle the boost builds immediately the moment you push the pedal to accelerate and go into boost...if you turn the boost controller off you see that boost drop by about 0.3bar at any constant boost.

eg. if I go from vacuum to light throttle and hit say 0.5bar pretty much instantly if I then turn off the controller it drops to about 0.2bar.

Its able to hold the actuator shut until it needs to open it at the pre programmed boost limit as it can vary how much air its bleeding as opposed to it being a constant like with a bleed valve.

You'll have to come out in mine sometime John! You'll be able to monitor pretty much anything from my cockpit Latest additions include techtom lol!

Rgds
Nito

Pooder

What is WOT

mutant_matt

Wide Open Throttle.....

john banks

I'm still not sure how reading vacuum and boost will tell you throttle position when you are *on boost* ? Surely it doesn't matter how clever the logic is - you can't derive information on a variable that is masked by so many others reliably? With different gears and inclines can't see how it would detect rate of change and convert it meaningfully to a throttle position? Given that a small turbo can give full boost on 1/3 throttle how can it then tell where the throttle is for the other 2/3 throttle travel?

Blackscooby

dfullerton,

Very easy, just cost me thats all

All the work in setting up the fuel and boost controllers was performed by DP Motorsports in Warrington.

I wouldn't know where to start !

Cheers

James_PowerMad

Oh, dear...

John, something I learned about a month ago is that the classical form of closed-loop control just doesn't work for boost control. There is simply too much transport delay in the system, which will form an oscillator really easily.

AFAICT, the original Subaru system works by learning what duty-ratio is needed to give the desired boost over a long (relatively speaking) period of time.

This value would be learned only when the RPM was above say 3500, with WOT for a second, so the boost has stabilised.

Then, when RPM is below say 2500, and any small throttle is applied, the duty ratio is almost full on (spool stage).

Then as the boost is getting towards target / RPM increasing, the duty ratio is lowered down to the learned value (transition stage).

Then the learned value is used till 4500 or so, when it tapers down the duty ratio gradually.

Original issue - part throttle... the duty ratio will either be very high, or locked at the learned value (depending on RPM).

James_PowerMad

BTW. The original Subaru boost control algorithm is actually pretty damn good in my opinion. ;-)

+ I have lots of graphs to back up my story if anyone is interested?

EvoRSX

James,
That sounds like the same way in which my Blitz boost controller works on my Evo.

If you set it in auto and set it to hit a certain value of say 1.2bar it then 'learns' your driving style and adjusts its response accordingly. i.e. first time you boot it after setting it up, it may well give a overshoot of 0.1 bar or more but then the following time you boot it, it will only overshoot by 0.8 etc etc. It then learns the response that it needs to get the boost to the desired setpoint. It also repeats this process in each gear so that it gets the desired response under different load conditions etc.

I haven't tried adjusting the gain to see if this provides a change in the rate of spool up but this is what it's supposed to do and its only supposed to change the rate at which the boost hits the desired level and not to change how much overboost you generate. The amount of overboost is done using the scramble feature (on the Blitz)

Andy

john banks

James I am very puzzled.

Firstly classic closed loop boost control works incredibly well on my car just using PI. It hits target MAP voltage time after time in every gear, and is the best boost control imaginable in my book.

Secondly, from discussion with Stephen Done and looking at the Ecutek site, there seem to be static duty cycles by RPM and load, with a boost target. This is on MY99/00. http://www.ecutek.com/tuning/oemcompare

Thirdly, if the ECU does adjust the duty cycles over time I have seen no practical evidence of it on my MY00 with and without PPP. If the actuator, restrictor or breathing are changed it goes completely haywire and the boost is far more variable day to day than a closed loop EBC. It seems open loop during spool up with a fixed duty cycle, and only if the boost goes too high does the ECU butt in, and if it is more than 1 PSI over held it oscillates. Peaks and offsets seem to occur with only minor system changes and it just falls over on my car. On the EBC I can set a duty cycle that will produce up to 3 PSI away from target boost and it will find it with little drama without oscillations or ridiculous peaks. It practically seems far more effective than the standard ECU boost control.

I gather MY01 and after are completely different and seem to adjust parameters over time to even reverse the effect of a decat or restrictor change.

I thought the whole point of closed loop boost control is that it will cope with delays in systems by appropriate setting of the parameters. It certainly behaves practically like that from what I have seen.

[Edited by john banks - 3/28/2002 2:34:16 PM]

dowser

James

Stephen Done's/EcuTek's DeltaDash software actually logs the solenoid duty (Primary Wastegate Duty Cycle). If you download the demo program and the maps from real cars you'll see how Stephen's my99 ecu (modified by him/PE, of course!) is controlling the solenoid.

On the 'drag' log file you'll see that between 31 and 32 seconds the duty is dropped from 86 down to nothing (Stephen - I assume this figure is %?) - this correlates to engine revs of 4500 through to the gear change into 2nd.

I too understood that the ecu guessed a duty rate based on other parameters and that it can learn. Looking at Steve's map, I'd say it does pretty well too I'd like to see the log from an oscillating PPP's boost control affected car to compare the results, especially duty compared to revs and MAP

Richard

dowser

John - posted together

Only anomaly I've seen on my my00 PPP was when going from Dawes (been on 4 weeks) back to a standard solenoid/restrictor combination....on a cold day....guaranteed to give problems in the past. It didn't - perfect boost control (albeit a bit slower)!

I should have reset the ecu to rule out any learning it may have done, but didn't think of it at the time. And *if* it had learned, from what? Dawes was controlling boost - not the bloody solenoid....me confused.

Richard

john banks

I am looking at the drag.csv file in Excel now.

The striking thing is that the same duty cycles crop up at the same RPM and throttle positions in each gear and appear to show no compensating behaviour with respect to the actual manifold pressure seen. This would support it being an open loop map which only has an integral component to correct boost if it goes over target for that load/RPM point.

My PPP ECU just cannot get it right (even after a reset or being out of the car and powered down) - you can adjust the restrictor and adjust held boost all the way up to the target point in the map at which point the ECU intervenes and drops the boost. It appears to make no effective attempt to compensate and it doesn't improve with time - earlier in the week I was running a restrictor that was too small for the PPP and it just did not improve at all over a few days. It just oscillates and never learns. The OEM boost control is fine if you have the right parameters in its map, but for those of us that are not able to access these tables for each modification we do, we have to resort to restrictors to sort it out.

[Edited by john banks - 3/28/2002 2:55:11 PM]

john banks

Richard did you have a fairly large bore restrictor in the pipe you put back in? What would happen if you put a 1.2mm one in with PPP and DP?