PI lean spike

[Peter Cobbold]

Summarised:

http://supertrarged.wordpress.com/2013/06/12/the-lucas-pi-lean-spike/

 

Peter

[Steves_TR6]

Very interesting !

 

i am unable to think of anything i could do with this new knowledge though

 

steve

[Nick Jones]

Interesting. What you don't mention though is the effect of the changing manifold pressure affecting the evaporation rate of the pooled fuel and fuel on the wall which some would argue is the largest effect. End result is the same though.

 

Interesting write up here

http://www.msextra.com/forums/viewtopic.php?f=4&t=16087

which discusses this in the context of adjusting the "accelerator pump effect" for Megasquirt EFI.

 

In fact PI is probably about as good as you can get, short of direct petrol injector into the combustion chamber, at minimising these effects but still not quite good enough to eliminate the lean spike. Setting it slightly rich (which many are anyway) mitigates the effect at the expense of fuel usage.

 

Nick

[Peter Cobbold]

Interesting. What you don't mention though is the effect of the changing manifold pressure affecting the evaporation rate of the pooled fuel and fuel on the wall which some would argue is the largest effect. End result is the same though.

 

Interesting write up here

http://www.msextra.com/forums/viewtopic.php?f=4&t=16087

which discusses this in the context of adjusting the "accelerator pump effect" for Megasquirt EFI.

 

In fact PI is probably about as good as you can get, short of direct petrol injector into the combustion chamber, at minimising these effects but still not quite good enough to eliminate the lean spike. Setting it slightly rich (which many are anyway) mitigates the effect at the expense of fuel usage.

 

Nick

Nick,

Yes, evaporation is faster at lower pressures, so slows when the manifold is near atmospheric- but I had to keep it simple.

 

Someone might have tried moving the Lucas injectors closer to the inlet valves to reduce fuel pooling- we'll see, but apart from that the manifold cant be improved AFAICS. Maybe they got too hot closer in. Would have liked to be a fly on the wall at Triumph when they were working on it.

I suspect these engines on carbs, or efi, will run at AFRs around 15-16 at cruise, given a tad of vac advance. That would reduce consumption by around 20%. What I dont like is running at AFR 13 at cruise just to avoid a second or so pinking. Or having to use 98RON fuel.

Probably the easiest route is to get rid of the pinking and live with the lean spike and transiently retard the spark for a second after each sudden throttle opening. Once thats gone then tuning the PI for leaner cruise could be done following that link by Tim Walker.

I'd do it myself, but the PI has long been shelved in favour of one SU.

Peter

Edited June 12, 2013 by Peter Cobbold

[Richard71]

That indeed is an interesting article Peter.....drawing on your mechanical experience of the TR6. Upon cylinder head removals & inspection over your years of driving, how does carbon build up compare between Pi and carbs with similar driving conditions?

 

Richard.

[Tim D.]

Very interesting article.

I spent some time last year with my wideband O2 trying to eliminate the transient snap throttle knock. It now seems only to occur with cheap petrol which is good, and a recent trip to Spa returned around thirty mpg

I like the idea of dragon teeth by the way.

I guess conceptually the issue is that the MU has to deliver a transient overfueling in response to sudden increases on thottle.

I could imagine a mechanical system the allows a sudden drop in manifold vacuum to cause an overshoot in the MU which returns to the required injection size. Perhaps a combination of spring and damper?

Just a thought ?

Tim

[Nick Jones]

Nick,

 

I suspect these engines on carbs, or efi, will run at AFRs around 15-16 at cruise, given a tad of vac advance. That would reduce consumption by around 20%.

 

Peter

 

My 2.0L Vitesse does happily cruise at 15 - 16 AFR and has been known to top 40mpg though upper are 30s more typical due to lack of restraint from driver.

 

So far I've not been able to get my 2.5 PI to match it though - it needs more fuel to prevent hitching. Same cam and similar induction/exhaust. Don't know why it's different unless it's the larger chamber and reduced squish. Haven't got above 35mpg from it yet in spite of more favourable gearing.

 

Nick

[Peter Cobbold]

That indeed is an interesting article Peter.....drawing on your mechanical experience of the TR6. Upon cylinder head removals & inspection over your years of driving, how does carbon build up compare between Pi and carbs with similar driving conditions?

 

Richard.

Richard,

Not sure that I'd expect the PI rich mixture to leave more carbon behind, the mixture should all burn but not to completion. In the rich mixture more carbon monoxide does not get oxidised to CO2 so consumption suffers.

 

Unfair comparison - PI vs SU+blower ! But I'll report when the head is off.

Peter

[Peter Cobbold]

Very interesting article.

I spent some time last year with my wideband O2 trying to eliminate the transient snap throttle knock. It now seems only to occur with cheap petrol which is good, and a recent trip to Spa returned around thirty mpg

I like the idea of dragon teeth by the way.

I guess conceptually the issue is that the MU has to deliver a transient overfueling in response to sudden increases on thottle.

I could imagine a mechanical system the allows a sudden drop in manifold vacuum to cause an overshoot in the MU which returns to the required injection size. Perhaps a combination of spring and damper?

Just a thought ?

Tim

Tim,

 

Do you see the lean spike on the wideband?

 

Are you thinking of a way of keeping the MU diaphragm at atmospheric for a second or so after shutting the throttle? You could do that with a microswitch that makes when the throttle is shut**, a 1to2 second timer circuit that open a 3-way solenoid valve tee'd into the manifold-MU hose. When inactive and closed the valve links the MU to the manifold as normal, when active for a second or so it blocks off the manifold connection and links the MU to atmosphere. It would however not only work in gear changes, but that might not matter.

Tune the amount of extra fuel by adjusting the 'on' time and watching your UEGO.

Peter

 

** clutch actuation might be easier. and it keeps the enrichment specific to gearchanges

Edited June 13, 2013 by Peter Cobbold

[Peter Cobbold]

 

My 2.0L Vitesse does happily cruise at 15 - 16 AFR and has been known to top 40mpg though upper are 30s more typical due to lack of restraint from driver.

 

So far I've not been able to get my 2.5 PI to match it though - it needs more fuel to prevent hitching. Same cam and similar induction/exhaust. Don't know why it's different unless it's the larger chamber and reduced squish. Haven't got above 35mpg from it yet in spite of more favourable gearing.

 

Nick

Nick,

Those are mileage figures that I suspect most TRs never approach. We could learn from your tuning methods. So can I invite you to post an article on the blog as to how you do it? PM me.

Peter

Edited June 13, 2013 by Peter Cobbold

[Tim D.]

Have attached a trace showing a gear change as measured using an AFR (I didn't have an RPM trace on this run)

Cruising AFR is between 12-14 depending upon speed. tends to lean out a little at higher speeds..

Tim

 

[Peter Cobbold]

Have attached a trace showing a gear change as measured using an AFR (I didn't have an RPM trace on this run)

Cruising AFR is between 12-14 depending upon speed. tends to lean out a little at higher speeds..

Tim

 

Tim

OK if I add the trace to the blog post?

Peter

[Tim D.]

Of course.

I have a load of data. At some point o had been planning on spending a day mapping the effect of turning the mixture screw on the MU on the AFR at various manifold pressures.

Tim

[Peter Cobbold]

Of course.

I have a load of data. At some point o had been planning on spending a day mapping the effect of turning the mixture screw on the MU on the AFR at various manifold pressures.

Tim

Excellent - keep it coming!

Peter

[Nick Jones]

Nick,

Those are mileage figures that I suspect most TRs never approach. We could learn from your tuning methods. So can I invite you to post an article on the blog as to how you do it? PM me.

Peter

 

I cheat I'm afraid. Both cars are Megasquirted with wideband O2 sensors. This means that I can drive around with a laptop and datalog what is happening or get someone to drive me around while I fiddle with it on the fly. There is even the ability to put in a target table of AFRs wanted against particular rpm/load points and set it to tune itself - which also works pretty damn well - you drive to cover as much of the range as possible and it sorts itself out - only as good as your target table of course. I usually set cruise areas up by mapping on the fly with someone else, familiar with the car, driving and keep taking fuel out until it either "goes flat" or starts hitching and fidgeting. At this point I know I've gone too far and put a bit back in. You don't even really need the wideband for that part but it, does come in handy when doing the higher load areas.

 

Nick

[Peter Cobbold]

 

I cheat I'm afraid. Both cars are Megasquirted with wideband O2 sensors. This means that I can drive around with a laptop and datalog what is happening or get someone to drive me around while I fiddle with it on the fly. There is even the ability to put in a target table of AFRs wanted against particular rpm/load points and set it to tune itself - which also works pretty damn well - you drive to cover as much of the range as possible and it sorts itself out - only as good as your target table of course. I usually set cruise areas up by mapping on the fly with someone else, familiar with the car, driving and keep taking fuel out until it either "goes flat" or starts hitching and fidgeting. At this point I know I've gone too far and put a bit back in. You don't even really need the wideband for that part but it, does come in handy when doing the higher load areas.

 

Nick

Nick

The Megasquirt info would be valuable to know. It sets a standard that the Lucas could aspire to.

Also I hope to get the SU to run precisely lean so knowing how lean is feasible with your set-up will show me where I have to go.

I think you posted a AFR table for MAP vs rpm on here a year or so back - that would be very good for the blog !! So would the spark timing for the same map.

Peter

[Tim D.]

Peter,

Happy to send the data, not sure of the best way to send it?

My aim is to get the RPM feed set up ( I have a techedge WBO2 that is a bit more basic ( read cheap) than the innovate) and ideally manifold pressure. I have picked up a MAP sensor just need to wire it in.

Then I could generate a full dataset.

Would be great to compare with the megasquirt.

I part built a megasquirt system on a lotus twincam. Only got it as far as running the ignition. Nice piece of kit.

Was one of the reasons I bought the TR 6 as I was intrigued by the idea of mechanical FI.

When I got the car it was running really lean. Would only run with the enrichment lever part engaged. Interestingly advice was not to tweak the MU. I Had to agree that without an AFR meter adjustments are difficult, but with one I had the car working in the 12-14:1 range in 15 mins. I also thinkit is running a little leaner than most. I was surprised at Spa how rich most PI exhausts smelt. Mind you mine was bad on the way home due to a stuck injector!

Regarding your suggestion of using electronics to get transient enrichment, nice idea. However what would be really cool would be to perfect a mechanical solution. Now there's a real challenge.

Cheers

Tim

[Peter Cobbold]

Peter,

Happy to send the data, not sure of the best way to send it?

My aim is to get the RPM feed set up ( I have a techedge WBO2 that is a bit more basic ( read cheap) than the innovate) and ideally manifold pressure. I have picked up a MAP sensor just need to wire it in.

Then I could generate a full dataset.

Would be great to compare with the megasquirt.

I part built a megasquirt system on a lotus twincam. Only got it as far as running the ignition. Nice piece of kit.

Was one of the reasons I bought the TR 6 as I was intrigued by the idea of mechanical FI.

When I got the car it was running really lean. Would only run with the enrichment lever part engaged. Interestingly advice was not to tweak the MU. I Had to agree that without an AFR meter adjustments are difficult, but with one I had the car working in the 12-14:1 range in 15 mins. I also thinkit is running a little leaner than most. I was surprised at Spa how rich most PI exhausts smelt. Mind you mine was bad on the way home due to a stuck injector!

Regarding your suggestion of using electronics to get transient enrichment, nice idea. However what would be really cool would be to perfect a mechanical solution. Now there's a real challenge.

Cheers

Tim

Tim

If you send me a PM I'll send you my email address - I dont like putting emails in plain text on fora, too many spammers trawling for them!

Mechanical cure...hmm, it would have to mimic a throttle pump to be fast enough. Maybe a Bowden cable from clutch cylinder lever to excess fuel lever on MU...Depress the clutch to activate extra-fuelling.????? *** That might work for gear changes, but maybe there's also a problem with pinking simply when opening the throttle from cruise? I never had that, but pinking does seem to crop up often on the forum.

cheers,

Peter

 

*** edit. That would be a real pain in traffic - ditch the idea I think.

Edited June 14, 2013 by Peter Cobbold

[Smeggie]

Copied from ignition timing thread, I am at the bottom of the thought chain on mapping but agree basic adjustments are easy and my car doesnt stink of unburnt fuel

 

Thanks Peter

For advice on checking TDC, it took two reads to understand why anybody would use a Mini wheel bolt on a Triumph but a simple solution! I fitted a new timing chain earlier this year and was not 100% confident with the results using the punch marks on the wheels so will give this method a try. But first I will use this to see how close the marks are on the pulley. Your article on Lean Spike is very informative. I have fitted an AFR sensor but not a map sensor, that said I have without much effort got similar readings to those on the accompanying article. Mostly around 13:1 which leans out momentarily on hard acceleration or gear changes.

Cheers

Smeggy

 

[Tim D.]

Finding a cheap MAP sensor that works in my system is going to be the challenge.

I thnk using the enrichment level is probably the way to go.

Tim

[Nick Jones]

Finding a cheap MAP sensor that works in my system is going to be the challenge.

I thnk using the enrichment level is probably the way to go.

Tim

Scrapyards are full of them these days quite a few of the older EFI cars have them external to the ECU. Alternatively Megasquirt ECUs mostly use these

http://uk.farnell.com/jsp/displayProduct.jsp?sku=1457153&CMP=KNC-GUK-FUK-GEN-SKU-G12&mckv=sNWTzNat5|pcrid|13359409269|kword|mpx4250ap|match|p|plid|

 

Nick