EFI University

[sam@tdi]

[StockieXR6T]

Not so good in 4x4 racing. I have seen large inpacts cause throttle faults in stock reprogrammed ecu's (LS1 1meg, E38, E67) and complete failure of throttle when dunked under water. Last thing you want when powering up a rocky 45 degree incline is the throttle to close due to a large bump. Have seen it happen, going backwards fast!

[sam@tdi]

Thats a very good point.

You have to strike a fine balance here, the temptation is to remove almost any tolerance for positioning errors,

The first fail throttle protocol I programmed dictated that if TPS is outside +/- 1% of DBW target for greater than 1ms or TPS/PPS a/b validation is outside by the same margin then the H'bridge controller would let go of the throttle motor and the throttle flap would return to a 10% tps sprung stop.

After a while of suffering similar systemic failures to the one in your example, ie throttle failed when driver contacted or throttle failed when driver smacked an inside curb with a touch too much enthusiasm, I decided to simulate some real throttle failures in the safety of the workshop. Without going into every single detail what I found was that whenever I hobbled the system to any meaningful extent, the kind of extent you might be worried about, then the failure in terms of positioning and time was massive and very obvious. Based on that research I made my systems much more tolerant to minor positioning errors, even major positioning errors for very short periods, this so far has led to 100% throttle system reliability, fingers crossed.

Regarding throttle failures due to dunking in water, frankly use a water proof wiring loom and I'd be willing bet that problem would disappear.
_________________
Sam Borgman
www.facebook.com/torquedevelopments
www.tdi-plc.com
www.rototest.com

[Turboivo]

I have red somewhere that in modern OEM's DBW controls as a safety feedback for throttle position is used the mass airflow and engine rpm. Anyway, at some point of the article it was mentioned that if one put a "hotter" cam on such engine (without proper software change) it could lead the engine to limp home mode due to the increased flow.
Just my 2c.

[Gadgeroonie]

can a DBW throttle stick open ?

what would cause this to happen. is there hardware built into the throttle body to stop this happening ?
_________________
Dyno Developments
2WD and 4WD Chassis Dynamometers



www.DynoDevelopments.co.uk

[bing]

[sam@tdi]

Absolutely right, and yeah it has the ability to stick open the same as any other throttle system. But unlike with the average mechanical throttle the engine management is watching this process like a hawk and as such it is far far more likely to identify a fault situation and step in to save you than a in a similar situation with a mechanical throttle system.

The basic safety features present in the typical system are,

2x throttle pots (normally one high>low the other low>high) parity checking each other, these are normally referred to as TPSa and TPSb, hence what I meant above by a/b validation.

2x pedal position pots (again one normally high>low the other low>high) working with the same a/b validation as the throttle position sensors

As ascertaining these two sensor feeds accurately is so crucial the threshold for a/b validation is normally fairly narrow, something like a/b must not be outside 5% of each other for more than 2ms.

The other level of safety comes in the form of a simple process comparing the final DBW target position and the actual (a/b validated) TPS signal. I personally aim to reduce the DBW position error to below 1% within 5ms opening and 2ms closing.

But as I said above when you actually do see a failing throttle body due to seizing motor bearings or reduction gears or carbon fouling on the plates etc, you find that the DBW err starts to generate some very large numbers momentarily and this is the cue to start thinking about shutting the thing down.

I don't know what failure protocols other people are writing but I tend to deliberately default the H'bridge controller on an a/b validation error which as a Bing says allows the throttle to safely ride down powered by a closing spring until it hits a sprung stop high enough to give a high idle. The reason being that whilst you'll have no throttle control you will still have power brakes and power steering where applicable.

On a positioning error I operate the same failure technique but with an added twist of a watch dog program that dictates that if the throttle hasn't returned to the sprung stop within 150ms then the ecu should cut Ign and Inj outputs killing the engine. Slightly scary when it happens but no worse than running out of fuel, and certainly better than the alternative of a stuck open throttle.

From an electronic standpoint you'll also find that the current drawn by a stuck motor thats generating a large PID duty will often be 2 to 3 times higher than under normal operation, so careful fuse selection on the H'bridge power feed can be helpful and a pure "belt and braces" fail safe.
_________________
Sam Borgman
www.facebook.com/torquedevelopments
www.tdi-plc.com
www.rototest.com

[gregpe]

I agree with all previous posts but I'm still concern about reliability..
one friend of mine is racing mini cooper ( new mini) with OEM throttle end M800. he sad that need to replace throttle every 2 -3 races because reliability problem. I think that race car vibrations causing the problem, engine is probably fitted without any rubber mounts...
What about you guys do you have similar problems.
From mine memory a lot volvos had problem with DBW as well (road cars)....
some Ferrari 360 prepared by Michelotto instead using DBW still use cable -reliability problem as well....

[Gadgeroonie]

on a different note

how easy is it to convert a fly by wire to a cable operation

and

can you get dbw throttle bodies that are cheap and reliable
_________________
Dyno Developments
2WD and 4WD Chassis Dynamometers



www.DynoDevelopments.co.uk

[StockieXR6T]

Our appllication have been stock E38 ecu's so the option of changing fault parameters is not there. To convert one of these means winding back the loom and ecu to an earlier LS1 PCM and using an interface or changing the cam gear and relutctor ring on the crack. Or going aftermarket ecu.

[Interested]

What Software are you using on the E38 ECU's ? I have seen the drop down box to disable DBW outputs in the software I use.

The Ford SCT gear has the ability to disable/ignore the DBW section aswell.

[StockieXR6T]

Mostly EFI live, sometimes HP tuners, depends on cost of license and what mods.

[sam@tdi]

[gregpe]

So which one is the most powerfull, quickest then? Which would you recommended?

[sam@tdi]

I haven't got part numbers to hand at the moment as I'm in China but I can grab some part numbers from some of the better units I've used and post them up when I get back.

From memory though the Bosch unit that comes with the GM Z20 engines (Astra/Corsa VXR. VX220) has a 10amp motor and is a strong little unit. I got some 100mm units from a US model 4x4 as well which turned out to be very good and quite reasonably priced.
_________________
Sam Borgman
www.facebook.com/torquedevelopments
www.tdi-plc.com
www.rototest.com

[figgie]

In the end,

it is not the performance of the DBW that sets it appart from the other but instead the fail-safe mechanisms.

I come from the F-16/f-15 B-1b, b-2 bomber world.

FBW on ALL jet engine aircraft for power output % and on the bombers, ONLY way to use the 4 x 4000psi hydraulic outputs to move surfaces at speed. On those, we are looking at quad redundancy channels with quad controllers.

On the cars, they only have dual channels but only one active electrical driver, that being the ECU/ECM. the Throttle and the DBW TB are passive without that signal.

Personally, I have always liked DBW. Mapping the TP% to TB opening% is always great.

[DKM.Motorsport]

Im a fan of DBW for a street car.

Essentially you are mapping a torque curve via the pedal. When the driver wants more torque - they push the pedal. The only issue is shaping the ratio of input form the pedal to the Throttle and getting it right.

Mapping DBW is also excellent as you can control VE across the range. Specifically for N/A its fantastic as excessive throttle opening can slow port velocity and drop VE. Again you sort out your shape of pedal input to TO, then scale the TO back form there on the dyno to suit maximum torque.

Excellent benefits for fuel efficiency!

[Hugh]

Have decided to take the plunge and retro fit an M3 throttle pedal and fly by wire motor to my engine.

The M3 setup has a throttle Motor



A throttle potentiometer.




And a throttle pedal with a potentiometer.

The ECU input needs to look like this with potentiometer voltages working in the opposite direction.



I am looking for a wiring diagram for the Motor and the potentiometers.

Can anyone help?

Thanks

Hugh
_________________
R32 GTR Nissan Skyline Race Car+ twin HKS 3037S
OS Giken 3 litre block + OS Giken sequential
Hopefully Motec M1 ECU

[CMW]

Hi Hugh, funnily enough I have been asking about DBW for multiple throttle bodies on the Motec forums, as I too want to add this to my Nissan RB26 engine, mainly as a means of adding traction control and paddle gear shift for the Tractive sequential that's behind this engine.. Motec were a bit sniffy about DBW on the good old `elf `n' safety grounds, and were talking about not unlocking the ecu unless running an OE DBW set up, not a home brew add on... That aside, I too have been looking at the BMW M3 motor as it looks quite a flexible package to mount. Does anyone know how much grunt it has? Is it a reliable unit? Are there others that would lend themselves well to multiple throttle body set ups? Finally, is there a requirement or overwhelming reason to use two separate TPS sensors, or can you get a combined TPS sensor with both resistive tracks in one housing? Thanks.
_________________
Best regards,
CMW

[sam@tdi]

Guys for getting creative with DBW the Motec is the wrong tool for the job, I would recommend you look at the Life F88.

Overall I feel that DBW is safer than a cable throttle as long as the A/B validation protocol is programmed properly. Yes you do absolutely need two TPS tracks, otherwise it could read wrong and you wouldn't know, if you have two you can tell your ecu to cut the engine if they ever differ by more than say 5% for 100ms or more.

I believe the BMW throttle actuators are a DC motor rated at around 10amp, plenty strong enough. You'll find that the trick to making DBW work well and work reliably is properly controlling the motor with a good H' bridge and then protecting the mechanical aspect of the system with the software calibration.

On occasion I've been asked to "look at" some drive by wire systems built and set-up by others due to continued unreliability, and I am starting to understand why the guys at Motec are reluctant to open up the control loops because some of the work being done is shockingly poor.

Most of the DBW hardware unreliability I see is simply caused by bad H'bridge control, or dumb programming mistakes, like for instance giving the system the ability to define a target throttle plate angle of say 90deg when the max physically achievable is 89.5deg's thus causing the system to hit the stop and then wind up on the integrators constantly driving the motor against the mechanical stop, the motors do well to last a couple of minutes like this.

I've also seen multiple cases of drive gear damage even on really robust OEM hardware caused by rubbish aftermarket PID control programming, people get in there and crank up the gains without bothering to calculate the final compound results of the numbers they're using and can inadvertently end up with viscous bouts of positive feedback, and as you can imagine the system just loves cycling hard from stop to stop at say 100hz
_________________
Sam Borgman
www.facebook.com/torquedevelopments
www.tdi-plc.com
www.rototest.com