Megasquirt injector dead time guide, pic heavy
 

Right why am i bothering to mess around with my megasquirt settings?

Answer: I have read lots of posts on this forum and others about megasquirt not being able to idle smoothly at AFR's much higher than about 13-13.5. It seems that some people have no problems at all whilst others (me included) have been. At first i thought this was due to people using EFI heads and others using spacer plates but actually i believe it is due to people using different injectors, some with dead times closer the MS default 1ms.

My car idled fairly well at 1000rpm and 12.8:1, i could get it to 13.5:1 but the idle would be very weak casing the car to stall easily, it also over fuelled on light lift off (im yet to see if this has been improved but this is another common symptom).

What is injector dead time?
Theres lots of blurb about this in the interweb. But heres and brief and i hope simple explanation:

Injector dead time is the time it takes for fuel to start squirting from the injector AFTER the injector has been energised. Like a tap, water doesn't squirt out the second you turn the knob, it can sometimes take 1-2 turns before water comes out, the time between you first turning the tap and water coming out could be called tap dead time.....

Injectors also don't stop injecting fuel instantly (although they stop much faster than they start).

Injector dead time is heavily effected by fuel pressure and voltage, fuel pressure should be taken care of by your pressure reg, but voltage (which changes a fair bit) needs to be taken care of by a look up table or a constant value.

Megasquirt uses 0.2ms per volt as its standard constant. So for each volt the opening time is reduced by 0.2ms...

(and seeing as we love examples!)

An injector dead time of 1.1ms @13.2V and a voltage correction value of 0.2ms/V would change to 1.3ms @12.2V and 0.9ms @ 14.2 volts....

Whys it important?
Well, your ecu needs to know an injector dead time because it needs to know how long to open the injector for. Its not overly important at higher pulse widths as and variation is easily sorted by your VE table, but at lower pulse widths where the injector isn't injecting very much fuel at all a small change in pulse width (0.1ms) can be the difference between a squirt and no squirt. The injector dead time is added to the PW.

This would explain why my idle had to be rich, as a leaner idle meant small pulse widths and not very much fuel, hard to control and difficult to achieve a lean smooth idle.

How do you find dead time?
The easiest way is to look for a manufacturers figure... however i really couldn't find mine (Bosch 360cc), Siemens Dekas are fairly well published. But it never hurts to double check.

Another method uses an Oscilloscope, and another mean playing with number of squirts and some other jargon... these methods are at best approximations.

I measured the actual dead time of mine using my MS ecu, and some jam jars.....

This was only a first attempt as i will be changing injectors soon and will do it all again properly.

How i did it...

I wanted to use my megasquirt in out put test mode to gradually increase the pulse width until fuel JUST started coming out this pulse width is my dead time.

I did this in the car, i would have been better doing it with a power supply so i could alter the voltage more and get a more accurate voltage correction value.

I removed my fuel wiring and injectors, stuck the injectors back into the fuel rail and then shoved the free ends into some jam jars. adjusting my fuel reg to 3bar (running pressure) as i was getting a value of nearly 4bar with no inlet vac attached.

Heres some pics

Attachment 137147

Wiring removed

Attachment 137148

Attachment 137149

Injectors in a jar using the heater box as a handy stand

Attachment 137150

and a pic of my Ohhhh so professional set up

Attachment 137151


And thats about it.

I ran the fuel pump and tested the injectors, starting at the standard setting of 1ms... no fuel, and slowley ramped up to 1.3... fuel JUST started coming out at 1.4ms, so thats my dead time.

Now the voltage correction.

In an ideal world this would be done using a power supply to give the base figure of 13.2V MS wants. However this was just a trial. So i turned every thing on, lights, fans etc. and got my voltage down to about 11.4v from 11.8 this was just enough to give a dead time variation of 0.05ms (1.45 instead of 1.4) and from that i now have a voltage correction value of 0.125ms/V.

From this i have a dead time of roughly 1.3 at 13.2V (this is still a rough figure, but its good enough to see if its improved anything).

Some pics

Attachment 137152

Attachment 137153

Attachment 137154

Attachment 137155

The result!..!

Fantastic, reducing my VE table around idle now gives a smooth idle at around 890-920rpm (need to turn this up!) and at an AFR of 14.5(ish) doesn't stall as easily and seems much smother:ban-split:

My injectors aren't amazing, they don't appear to atomise the fuel very well at small opening times and one injector seems to have a slightly lower dead time... but hey.

nearly back together

Attachment 137156

and a good figure at a steady idle :)

Attachment 137157

Some different injectors and some fine tuning should see the idle become even better.

Obviously my VE table now needs some adjusting :)


I've just upgraded to sequential 3.1.0 code so will be trying semi sequenial (and fully sequential when i have a cam sensor) as soon as my maps a bit more refined. This should help make for a smoother idle and better fuel consumption etcetc.

Cheers for reading and i hope it helps someone! if any ones got any suggestions/corrections please post them up as you'll be helping me for next time.

Good work rob, I will have another read when i've not had a beer, can only help when I switch to m/s from weber

Good read rob... Maybe you've stumbled on the reason for the varying idle afr probs people are having!?
Im gonna re-read and learn this ready for my MS!

Well done rob, these are the kind of threads that i like reading.

Excellent thread Rob, very helpful and something I have not played with before! :king:

i had a play with this recently, more to do with the voltage correction/injector setting than anything though.
the dead time did make a slight difference, but your ve tables can sort that out, just need a fraction more adding to the ve values really.
i will be going back to the dead times as i have figures for my injectors off someone else who had them flow tested etc.
good post though, may help some out.

Try it on MS 3 where there is a non linear correction map for voltage for each injector!

The problem is a lot of injectors, and dekas are really bad for it, give inconsistent results at small pulse widths so no amount of tuning wll help. If you're in the Market for new injectors, get Bosch ev14's

Rick

good point there rick, deka's are crap at low pulsewidths, but if i raise my idle to 950rpm, it is as smooth as you like at 14.7-15.1.
it's ok at 850rpm on 14.3-14.5, but higher idle is just nicer IMO.

xr2wishy - wot injectors u using?

630cc Deka IV same as Karl's

Good point about the VE table, however i didn't find that worked for me, i just couldn't get a good idle. Now with the correct dead time (or near to the correct dead time) it idles a million times better,

---

Cheers guys, hope it helps some one, or at least gives you somthing to fiddle with!

Rick i'll have a google on the ev14's :king:

Rob,

Also Rick, the MS3's independent dead time although slightly over kill does sound really handy, one of my four injectors had a different dead time by 0.1-0.15ms which ms3 would have taken care of.

The new 3.1.0 code seems to sort trigger wheels out slightly differently too as i don't have trigger wizard or trigger offset any more.

Rob,

EDIT*** are these the ones Rick? http://www.ebay.co.uk/itm/Authentic-...item4cf96ee587

Hello Rob!
Great progress.

There are a few web places where injector lag or dead time data is available.
Mostly North American data, but some international injector types.
As you can see from the chart (go to site below) there are vast differences between injectors. Very useful information for getting good efi results.

http://injector-rehab.com/shop/lag.html

i think the dead time works best at low resolution as the ve table can't seem to get it accurate enough to drop out 0.1ms or less at that level.
ms3 however can sort this out as it has 0.1 resolution rather than the ms2 which can only use whole figures, one reason i'm planning a change before too long.

Rob yea they are a 55lb one, perfect for your CVH

Lovely thanks Rick :king:

XR2wishy - yea MS3 does look very good, but MS2 does well enough :) the only place i would want such small step changes would be at very low PW values and frankly the dead time takes care of this very well.

Other than cam timing control, lots or inputs/outputs and a few other trixy bits the MS3 isn't a million miles better than an MS2. The new 3.1.0 code for MS2 takes care of full sequential for example.

Perry, cheers for the link :)

Rob,

So you can run fully sequential injection on MS2?
Whats the advantage on the Bosch injectors if the Dekas can idle at 950rpm at 15.1 afr?

Although I don't have Megasquirt, thats a good read. Love stuff like this, good effort, Rob.

Cheers Muz,

Roger, MS2 can run full sequential fuel and spark with the new code but you need a cam shaft position sensor too. I will be trying semi sequential as soon as my maps a little better.

Main advantages are (apparently!) better fuel consumption, better emissions and a better idle (its especially good at lower rpm's) with larger injectors.

BUT

Im doing because you can alter the injection time (when you inject fuel in the combustion cycle) and because you can set individual cylinder trim values. This means if one cylinder is slightly lean but the others are slightly rich (giving an average AFR from your single lambda) you can tune it out. Im going to do this with 4 EGT probes and a datalogger.

But generally batch injection is just fine.

Rob,

How are you going to measure the afr ratio of each cylinder with egt probes?

At a guess the leaner cylinders will be hotter?

Sounds like a good idea Rob will be nice to see how it works i think it will by quite a challange.

Good luck and keep us updated

fantastic thread!!

i thing stu has some competition with his technical guides lol

great stuff

All sounds good rob. As said many times... An excellent thread, good work m8!

I guess the semi sequential is 2 injectors at a time?

that's just batch firing. semi sequential tends to be sequential spark and batch firing injectors.

Yes mate, an EGT probe on each cylinder. Bit of a faff and takes a while to sort but it should be worth while in the end. My aim is to have a car that drives/starts/idles/cruises like an OEM mapped would. I'll do it on some rollers i think.

Yes mate, its a bit weird but the cylinder will get hotter the leaner it gets up until a certain point when it gets cooler. I'll never get each cylinder exactly the same but i'll probably get them within 50deg C. Eventually i want to run pretty high boost so i think this will be important to me.

Lol cheers mate, don't think stu has anything to worry about though :)

If people are interested i'll do the same when i sort the EGT/sequential?


Roger, yep semi sequential still uses 2 banks of injectors (on a 4cyl) but it injects half the fuel twice as often. So at the moment 1&2 get a full squirt, cylinder 1 can use it, but cylinder 2 is on its compression stroke so can't.... so it has to sit in the port until the valve opens, similar to wasted spark.

But with semi sequential half the fuel is injected into a port that can't use it, where it stays until the exhaust stroke (hopefully nicely atomised) when the other half is injected as a timed event. Some engines like fuel to be injected at a different time to others but generally we will inject on a closed inlet valve. TBH i may not see any improvement with semi or even full sequential, but im up for a good trial :)

Rob,

Nope, i'll be semi sequential fuel and wasted spark. I can't see many advantages to sequential spark tbh mate. I might go coil on plug at some point but i will still use wasted spark.

Rob,

All sounds good rob, some wicked plans there... Even without all that can the MS still be mapped to drive like an OEM ecu? IE smooth power delivery at part throttle and steady at cruise etc etc

Definitely mate, it just takes time and some patience. IMO anyone can map for WOT but the refinement takes time. My car is cruising really well, smoothy no missfires no jurkey power delivery etc, could be a little leaner but its getting there. It idles well without an idle control valve now should be even better with one.

Once the maps sorted i'll probably change injectors (and steal Karlos's base map :))and start all over again :cry: but then its time to sort out the starting. The car always starts but i haven't bothered to make it start well (this should be done last any how) as i don't have an idle valve yet which will help with cold starting with no throttle.

I found mapping the car to its current state really easy, a good base map really helps, the final tuning will however be done on a dyno.

But yes, as long as you spend the time the reward will be a car that drives really well, and really reliably. :king:

r u hinting that i havn't sent you my iacv yet lol?

Sounds good m8, can't wait to see your car and how the MS operates!

Hahah, lol not at all mate, the cars tucked away in the garage at the moment i go up there occasionally when im feeling enthusiastic (and the girl friends gone out!).