Page 1 of 1

Injectors to large for decent idle?

Posted: Mon Apr 16, 2007 7:37 am
by SierraV8
Hello,

I've been mapping an MSI with EDIS on 2ltr Turbo Zetec,

The engine spec should make 300bhp so i have used a standard set of Cosworth 4x4 injectors (330cc) running at 3bar fuel pressure.My problem is the "resolution" of the VE table and/or minimum on time of the injectors means the idle CO is around 4%, if i tune it for less at idle the engine stumbles and runs rough. (CO goes up and down)
If i fit a standard set of 2ltr zetec injectors and re-tune the engine runs better and idles around 1.5% with no problems. (just runs out of fuel on more than 7psi of boost!)

Im running open loop at all times and 4.7ohm injector resistors when using cosworth injectors (imp 2.5ohm)

I've had the same problem with big HP cosworths running Weber ECU and large injectors - anyone got any comments or suggestions?

Cheers Dale.

Posted: Wed Apr 18, 2007 7:24 am
by Dave_Coxon
I did some testing on the Cossie injectors, and found the response of them was very poor - by swithching to more modern design of injectors (Zetec) you are proving this...

I'd find a set of injectors from a late Rover turbo with the dizzyless engine.
Despite the Rover moniker, these are actually fantastic injectors with approximately a 450cc rating. The response time on them is 0.3ms better than the cosworth injectors at 3-bar rail... And at idle, that 0.3ms makes a BIG difference!

The injector code you want is MJY 100 460. They are high impedance btw.

Posted: Wed Apr 18, 2007 7:46 am
by SierraV8
Dave_Coxon wrote:I did some testing on the Cossie injectors, and found the response of them was very poor - by swithching to more modern design of injectors (Zetec) you are proving this...

I'd find a set of injectors from a late Rover turbo with the dizzyless engine.
Despite the Rover moniker, these are actually fantastic injectors with approximately a 450cc rating. The response time on them is 0.3ms better than the cosworth injectors at 3-bar rail... And at idle, that 0.3ms makes a BIG difference!

The injector code you want is MJY 100 460. They are high impedance btw.
Thanks Dave,

I did try a set of Rover Turbo injectors and found the same problem, did you tune the injector response time in Megatune?

The cosworth injectors are low imp, so i have used 4.7R resistors and left the MS injectors settings in Megatune as default.

The resion is used cosworth injectors was a standard cosworth will produce 270bhp and idle perfectly, this maybe because the weber ecu is sequental?

Cheers Dale.

Posted: Wed Apr 18, 2007 8:01 am
by Dave_Coxon
Did you try the later, not the earlier injectors?

The early Rover MJY 100 230 I recall?) are 'old school' stuff, and nothing like as good on the resolution. check the numbers you used.

Posted: Wed Apr 18, 2007 8:15 am
by SierraV8
Don't know, the car they came from was around 15 years old - Rover tomcat Turbo.

I understand wot you are saying with injector reponse time (some are better than others) but the smaller the injector the less affect the response time has on fuel flow at low RPM, normally idle problems (over rich) happen with very large injectors on small engines - cosworth 2ltr idles OK upto green injectors (around 470cc) after that thay always seem to run rich at idle, if you try to cut it back you get a very rough idle. (missfire - lean/rich/lean/rich)

I still think it should be possable to tune a zetec on standard cosworth injectors with no problems, did you tune any injector settings in your system? eg the injector response time - i've always left mine at 1.0ms, have you ever tried tuning this? I tuned in once but all it seems to do is add or remove on time.

Your feedback is appricated! Cheers Dale.

Posted: Wed Apr 18, 2007 8:42 am
by Dave_Coxon
My testing wasn't on a vehicle, but on a test rig I made specifically to find out what the response time of injectors were.

If you have a big-lazy injector of older technology, it'll always have a big flow rate at 100% duty - it'l also have an erratic, poor resolution at the smallest duty cycles. 100% duty is nigh-on useless as a figure anyway; 50% duty is more usefull.

More modern units (like the MJY 100 460) have both a high flow, and a VERY good resolution and response. whilst all injectors may look the same and just have different flowrates, the characteristics of the units are hugely different.

I suspect you tried the MJY 100 230 rover, not the 460

Posted: Wed Apr 18, 2007 8:46 am
by Dave_Coxon
FYI - This was information I submitted on another website following my tests - copied and pasted here for your pleasure. :)


Injector response of the Rover MJY100460 injector (as found on MPi mini / Dizzyless [DIS] Rover 2000cc 16v turbo).
This is a very VERY good injector. It has a fantastic spray pattern, and is a high-impedance unit.

35psi Rail Pressure.
0.5ms: Nothing
0.6ms: Extremely unstable / barely opening
0.7ms: Erratic
0.8ms: Good
0.9ms: Perfect

70psi Rail Pressure
0.6ms: Nothing
0.7ms: Nothing
0.8ms: Erratic
0.9ms: Good
1.0ms: Perfect


XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Now - here is something VERY interesting... A cosworth 'yellow' injector response was tested aswell. These are 'low impedance' units; supposedly superior response time to high-imedance units.

35psi Rail Pressure.
0.8ms: Nothing - injector not even opening
0.9ms: injector making a noise, but no fluid
1.0ms: Erratic
1.1ms: Perfect


70psi Rail Pressure
0.9ms: Nothing - injector not even opening
1.0ms: injector making a noise, but no fluid
1.1ms: Erratic
1.2ms: Perfect


XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Finally - the 'high impedance' vauxhall injectors (from an XE) circa 1988.
The reason i wanted to test these, was because EVERYONE says that low-impedance injectors have a faster opening than High impedance. It's written in enough textbooks and websites, so it must be right huh?
Well - as i reported in the first post - the low impedance 'cossie yellow' unit was some 0.2 to 0.3 seconds slower response than the (high impedance) Rover MJY100460 unit. This was throwing the 'oft stated facts' into the air. So where did these 'facts' originate from? Easy test - try some injectors of the same generation as the 'cossie' units.
These injectors are the ones used in the Vauxhall XE engine, right until it went dizzyless (whenever that was).
And what a load of crap they are!

35psi Rail Pressure.
1.2ms: Nothing
1.3ms: Faintest clicking- nothing coming out
1.4ms: Faintest clicking- nothing coming out
1.5ms: quiet clicking- nothing coming out
1.6ms: Loud clicking- nothing coming out
1.7ms: Erratic
1.8ms: Good
1.9ms: Perfect


70psi Rail Pressure.

1.3ms: Faintest clicking- nothing coming out
1.4ms: Faintest clicking- nothing coming out
1.5ms: quiet clicking- nothing coming out
1.6ms: quiet clicking- nothing coming out
1.7ms: Loud clicking- nothing coming out
1.8ms: Loud clicking- nothing coming out
1.9ms: Erratic
2.0ms: Perfect



So to summarise:
Twenty years ago - Low impedance (peak and hold) injectors were indeed faster responding than high-impedance injectors.
Times change however - as proven by an injector from almost a decade ago having a response time superior to that of a low-impedance unit of times (thankfully) long gone.
I don't feel it's unreasonable to assume that modern high-impedance injectors are probably down as low as 0.6ms stable response. I don't believe they'll get much faster than that however...
You cannot assume anything of course - if you are unsure, get your injectors tested.

Posted: Thu Apr 19, 2007 12:39 am
by SierraV8
Thanks for the info, Dave.

I think i'll get some differant injectors to play with but i'm still confused why a standard 2ltr cosworth YB does not appear to have these idle problems with the standard injectors.

Is you say the trend in recent years has been moving away from low imp injectors, for example the first EFI version 3.5L Rover V8 engine used low imp injectors, when the engine went from the old style "flapper" flow meter to the hot wire the injectors remained around the same flow rate but switched to high imp. These same injectors are used on the 3.9 to the last 4.6L.

Thanks again for the very helpfull info,

Cheers Dale. :D

Posted: Thu Apr 19, 2007 4:16 am
by detcollector
Actually your test data is inline with expectations. Its merely the terminology that is incorrect.
The time taken for the injector to open is the incorrect value for Megasquirt (MS) "injector opening time".
The value you put in MS is a compensation factor to get the injector to provide the amount of fuel calculated from the injector flow rate.
Low impedance injectors may take more time to open but they take less time to close. This results in the "injector opening time" (MS term) being less.
And new generation injectors have better magnetic characteristics and don't need to be low inductance to get the same performance.

Posted: Sat May 05, 2007 2:25 pm
by Mini Sprocket
Interesting stuff there Dave, I must have missed it some where. I am confused, as you know only too well thats easy to do.

I have some MJY10029 injectors (orange band) of this generation from a late model Rover 820 N/A with DIS these flow 230cc and were used batch fire, half that of the MJY100460's i had tested at 460cc sequential fire.

Im after some known Injector open/close times for these injectors to set me on my way :P Should I or should I not use the info Dave, that you have posted up, as the data to go for?

If any one can help out :?

This is the only topic that turns up with a serch for MJY100460 :( there used to be more

Nice facts, Dave.

Posted: Sat May 05, 2007 9:59 pm
by Heribert
Hi Dave ,
thanks for the time/performance tables at different pressure. Great job.

A bit on injectors and how they really work. The determining factor for the opening time is the drive voltage and the L/R of the injector magnetic winding.
The drive voltage is for MS appr and not 13,5 Volts and not subject to much variation once the engine runs and the alternator kicks in.
The L/R is what gives the absolute majority of the delay at opening. It limits the rate of increase of the injector current , since the injector is a current operated device ( force proportional to ampereturns). It can be measured quite easily with the aid of a shunt, a test generator and an oscilloscope . As test generator the MS box will do fine.
Drive the injectors with a 30 Hz square wave and measure the injector current on the scope.
You will be able to see when the pintle flies off the closing seat and travels to the open position. There will be a little s-curve on the exponentially rising current flank , and for most of the hiZ injectors I have tested it comes after 1,15 to 1,2 ms. I have not tested many loZ, and what testing I have done was done on antiquated Bosch stuff from the 70-s.
Fuel pressure makes only a quite small contribution to the opening time.

The closing time is a direct function of how high the flyback current voltage clamp level is . Have made experiments with special drive MOSFETs and no flyback clamp at all . HiZ closes in 50µs!.
With a simple diode+ zener diode 36 V clamp closing is 0,12 ms .
Anyone interested in the double diode clamp for hiZ, PM me with your email and Ill send you the schematic.

As to the difference between loZ and hiZ in opening time , it is quite easy ( if you have the correct instruments) to determine the difference by measuring the injector inductance ( L, in Henry) and resistance ( R in ohm ) and that will tell you the difference in performance .

A word of warning. Do not try to run injectors in real life without a flyback voltage clamp to find that "extra 0,1 ms ". The electrical disturbance created by 800 V flyback peaks is quite substantial and will create havoc !

Best regards

Heribert

Re: Nice facts, Dave.

Posted: Wed May 09, 2007 8:55 am
by SierraV8
Heribert wrote:Hi Dave ,
thanks for the time/performance tables at different pressure. Great job.

A bit on injectors and how they really work. The determining factor for the opening time is the drive voltage and the L/R of the injector magnetic winding.
The drive voltage is for MS appr and not 13,5 Volts and not subject to much variation once the engine runs and the alternator kicks in.
The L/R is what gives the absolute majority of the delay at opening. It limits the rate of increase of the injector current , since the injector is a current operated device ( force proportional to ampereturns). It can be measured quite easily with the aid of a shunt, a test generator and an oscilloscope . As test generator the MS box will do fine.
Drive the injectors with a 30 Hz square wave and measure the injector current on the scope.
You will be able to see when the pintle flies off the closing seat and travels to the open position. There will be a little s-curve on the exponentially rising current flank , and for most of the hiZ injectors I have tested it comes after 1,15 to 1,2 ms. I have not tested many loZ, and what testing I have done was done on antiquated Bosch stuff from the 70-s.
Fuel pressure makes only a quite small contribution to the opening time.

The closing time is a direct function of how high the flyback current voltage clamp level is . Have made experiments with special drive MOSFETs and no flyback clamp at all . HiZ closes in 50µs!.
With a simple diode+ zener diode 36 V clamp closing is 0,12 ms .
Anyone interested in the double diode clamp for hiZ, PM me with your email and Ill send you the schematic.

As to the difference between loZ and hiZ in opening time , it is quite easy ( if you have the correct instruments) to determine the difference by measuring the injector inductance ( L, in Henry) and resistance ( R in ohm ) and that will tell you the difference in performance .

A word of warning. Do not try to run injectors in real life without a flyback voltage clamp to find that "extra 0,1 ms ". The electrical disturbance created by 800 V flyback peaks is quite substantial and will create havoc !

Best regards

Heribert
Hi Heribert,

I did some very simple experiments a wile ago on a Rover V8 3.9 with slightly larger hiZ injectors fitted running the standard hotwire efi.
The standard ecu uses a pair of power transistors to switch the injectors, one transistor per bank of 4 injectors. The "flyback" protection as standard was a 100ohm power (2watt?) resistor in series with a 1microF 400v cap connected from each transistors collector to ground. There was also a 47v zener in parallel with the RC network. (collector to emitter/GND)

When i disconnected the RC network (lifted the 100ohm res) the engine ran slightly richer, do you think the RC networks maybe improve the injectors close time?

The megasquirt V2.2 uses an flyback circuit with a transistor, does this really help or is a zener and/or an RC network a better bet?

All this still does not explain why a stadard 2ltr cosworth idles and drives perfect with it's standard injectors and a 2ltr zetec with MS fitted using the same injectors does not idle with a CO less than 3-3.5%!

Maybe because the cosworth's ecu is seqential and the MS is batch fired?

Any suggestions?? Dale.