Fitting LPG - hey, it's a gas, man!

[Laird_Scooby]

44 minutes ago, hughezee said:

Good progress buddy, looking forward to an update

No sooner said than done Stu!

As promised, some more pretty pictures and some videos! The first pic shows the "static" state of the LPG ECU and the default value of the actuator/stepper-motor as 84 which does suggest that the engine now needs more fuel since removing the cat and sorting the exhaust leaks.

If you look at the rev counter, you'll see it's idling at about 1100rpm so it is still cold. It's also still on petrol. If you have a look at this first video (ignore the shouty tappets! ), you'll see that between 1500-2000rpm it's going very lean and the actuator is trying to compensate by opening up a lot. I opened the idle bypass screw on the vapouriser by about 1/2 a flat on the hex adjuster and ended up with this :

http://vid567.photobucket.com/albums/ss113/scoobyh123/LPG/VID_20160704_134323646.mp4

Keep your eyes peeled at about 1:15 as i let go of the loud pedal and the actuator dives for a position of 20 (i could set this lower e.g. 0 but the recovery time to default would be longer and may cause problems) for the fuel cut-off.

Having driven the car after this, it certainly drives well, power delivery is smooth and consistent and the changeover is seemless once the engine is warm. If i let it changeover while it's still cold, it does cough a little at first. I mentioned during the install i wasn't going to connect the temperature sender connection - i may rethink this and invoke the temperature controlled changeover. This would mean that it would still change to gas on a trailing throttle above 1800rpm but not until it had reached a certain temperature. I think part of the reason it currently coughs a bit when cold is because the Lambda sensors are single wire jobbies - i.e. non-heated so take time to warm up before giving the correct response.

On a slightly different note, i've come to the conclusion the switch in the coin tray is a bad idea. On fastback models there is a slot just above the Climate Pack for the rear wash/wipe switches. Sadly this is a bit too small for the LPG switch, unless i make a panel to fit inside the hole then another to hold it there and fit the switch in that, my only other option is to fit it in the shelf/instrument panel i'm fitting in place of the passenger airbag. I'm still undecided on which would be best, the Sterling is different again as the fog light switches are in the hole above the Climate Pack but i have a custom switch planned for that anyway.

More to come when i do something else and/or the next fill-up!

[Laird_Scooby]

Another little update - i looked at connecting the temperature sensor to the Leo ECU but forgot that the pin isn't connected in the multiplug that goes in the bottom of the ECU so that idea has fallen by the wayside.

However, the Leo can be set to ignore the Lambda input for a different amount of time so i'm going to try that instead.

Also filled up for the 4th time today and got a bit of a shock. I'd driven more gently, less short runs to the shops, walkies etc and steadier driving in general so wasn't really expecting 19mpg as the figure it returned! Used a smidge under 27L (26.29L i think it was) to do 111.7 miles which is about 19mpg give or take a gnats wedding tackle.

Something that has niggled me for a long time now is how sharp the handbrake is on the coupe. I'ts a long travel to the first notch anyway but i sometimes struggle to do that as it's getting pretty damned tight by then. If i really yank on it i can just about get it onto the second nothc but really struggle to release it. Thinking maybe the rear brakes are dragging just enough to cause a problem in the mpg (and maybe the performance) department which would explain why i've always felt the Sterling is livelier, more economical etc.

The other thing about driving more gently is that it's spent more time gently accelerating in lower gears to above 3000rpm where it tends to get quite a bit richer. That may be pointing to the venturi being a bit small in the mixer but i'm going to investigate the brakes first - probably the handbrake has been over-adjusted by that numpty of a mechanic that has done all the other bodges on the car! As i don't use the handbrake much, it could be why i've not thought of this before. Maybe i can neutralise several parrots with one missile on this one!

More to come once i've done a bit of probing........

Might even get round to fitting the air horns or as i like to call them, the "sc-air horns"

[hughezee]

Hand brake ding ding ding, I mentioned that a while back when comparing the two cars, I hope it's that simple buddy

[Laird_Scooby]

You did indeed, yet the coupé rolls easily enough when it's off but perhaps not easily enough when up to speed. I'll have a play in a little while.

[Laird_Scooby]

Something has just crossed my mind while reading my post back - some days it rolls much easier than others and goes better on those days, others it doesn't and it struggles more and the brakes seem a lot sharper on those days.

Would maybe explain the variable performance and economy if there isn't enough clearance for the brakes to work properly including the auto adjuster which does ratchet up noticeably on both of mine - not so "stepless" as later models.

[Laird_Scooby]

Well now! This has shocked me! Been on this planet the best part of 5 decades and been tinkering with mechanical and electrical stuff for 4 of those and genuinely not come across one like this before!

Start with the scientific stuff - i got a stopwatch app set up on my phone so i could time how long it took to roll from where i park it to the pavement opposite. This would give me my benchmark which after the first run of 12.67s, the next two came out at 10.7 +/- 0.1s at which poiint i drove it round the block to get a "live" feel for how it was running today.

Returning home, i removed the upper rear centre console trim/handbrake cover to reveal the adjuster nut that acts on both cables via an equaliser to operate both handbrakes (left and right) with the same force. Then i undid the adjuster nut on the handbrake cable, initially by 1.5 turns (6 off 1/4 turns) and did the roll test again - no difference. Something i noticed as i drove round the block with the window down, heater fan and stereo off was the sound of brake pads scuffing discs - normally the stereo is on as is the Climate Control/heater fan and the window is up so i can be forgiven for not noticing before, although i should have done.i suppose.

Anyway, i undid the adjuster nut another 4 quarter turns (1 full turn) and tested again. This time it was 8.7s so i thought i was going in the right direction. I tried it twice more and got 10.7s give or take which kind of bears out what i said earlier about being loose enough to pass the MoT (i.e. not binding) but tight enough to cause drag once up to speed. Taking it round the block once more the difference was tangible - the car was much livelier, smoother and felt lighter somehow as if not bogged down by a heavy weight in the back! Interestingly i felt the auto adjusters operate on the rear brakes at least twice during that journey and i still have very sharp service brakes but the handbrake needs at least two clicks to stop/hold the car now with 3 being just about possible if i use a lot of force.

While on this second tour of the village, i lifted the handbrake to the first notch and the car felt exactly how it did yesterday and on my first "benchmark" run today. I have at times felt the wheels after a long run and neither rear wheel has been warm, never mind hot so presumably whatever heat was generated by the drag was dissipated by the airflow around the wheel and it was admittedly rare that i came to a complete stop without slowing down for a few minutes or more first - this could have given me a false impression.

Time and the next fill-up will tell if this has cured it but it certainly feels very different!

[hughezee]

I hope that simple tweak solves it buddy

My Dad never has much luck with Honda handbrakes, his current motor had a rear axle imbalance like clockwork just before and MOT it all starts with him rolling at traffic lights so its the opposite of your potential issue. Almost three years of niggles and it was driving me mad and I think I've finally sorted it after discovering the N.S that was always excellent and strong was actually the the whole problem, difficult to imagine but you had to be there 

[Laird_Scooby]

I hope so too! One particular junction that i have to take a bit wide and tight left hand lock, t never crept there before unless cold and on fast idle. Now it creeps when hot there so it's certainly pointing at that being a problem.

Fingers crossed both mine and your dads are both sorted!

[hughezee]

Fingers crossed indeed  when I reluctantly ask he's says its been great, but only 6 weeks/ 1300 miles has past

[Laird_Scooby]

Well that won't do my fuel economy any good whatsoever!

Nipped to the shops and when i got back in the car, i hoofed it to see how it went - and it did! The difference, even in first is amazing! It's a lot quieter as well which i suppose it would be if it's not working it's hexagonal fasteners off just to overcome the brakes!

Think i need a run to the filling station again, partly to see if there's any obvious improvement in the economy (which might not show because of my lead foot! ) and also to give myself a "clean sheet" without having hoofed it etc to check it longer term.

[Laird_Scooby]

Had my run to the filling station, 42 miles on 10L which works out as 19mpg again. However, bear in mind that 30 miles of that was with the handbrake problem and short runs to check i'd got some progress by slackening the handbrake cable.

I'm not even going to try and summise what it may or may not have been doing round the village giving it a few beans while doing the test runs, the next fill will tell all - good, bad or indifferent!

[Laird_Scooby]

Well it's had the next fill - it took 23L which is a smidge over 5 gallons in old money. On previous use that would have given me about 90-95 miles if i was lucky. This time? It stretched all the way to 110 miles which equates to 22mpg so an improvement of about 16% just from adjusting the handbrake cable. Once again, most of my running was short runs locally so probably not the best test but at least it shows it's going in the right direction. If i add on the 15% to get the petrol equivalent mpg it works out about 25-26mpg which is about right for that sort of running. In terms of cost per mile with LPG @ 50.9p/L and Ultimate 124.9p/L it's roughly equivalent to getting about 55mpg on petrol.

Still hoping to get a bit more mpg from it, i've come to the conclusion the mixer is on the small side causing a rich mixture from about 2500rpm upwards. The actuator/steppermotor does a lot to curtail the rich running but doesn't cope with all of it. I have another mixer to try with a bigger hole in the middle (venturi) which should improve volumetric efficiency overall and once set up, should give a more balanced mixture across the rev range.

Fingers crossed!

On a slightly different note, i've found a suitable SGi (Sequential Gas injection) system for sale at a sensible price which if things go according to plan will be finding its way onto the Sterling. By using injection i'm hoping for better economy, performance and driveability - time will tell! Expect another (slightly shorter!) thread for that!

[hughezee]

Excellent progress and fingers crossed for an improvement buddy, really enjoying this thread

[Laird_Scooby]

Thanks Stu - i'll see how this tankful fairs before changing the mixer. Got a few other options to try if the bigger mixer doesn't improve things enough. First is a diffuser mixer and second is a universal mixer jet.

Both mount in a 12mm hole in the intake, the diffuser mixer is a glorified universal mixer jet which is basically a 12mm OD tube with the end cut at an angle. In other words, a very simple fixed jet!

[Laird_Scooby]

Coming back from my nearest town the other day, there's a downhill/uphill bit on double double-bend and on the village end of these bends, the 30mph limit starts.

Doing my usual "hoon it through the bends and lose everything behind me" trick, i reached the uphill section approaching the village. Still doing 65-70mph, i let go of the loud pedal at the point where i would normally need to let go so the incline slowed me down to 30mph or thereabouts for the speed limit.

Let me rephrase that - i let go of the loud pedal at the point where i always have done and have usually dropped to 30mph ish in time for the speed limit signs. This was the first time since adjusting the handbrake cable that i've done this trick, the traffic has been too heavy other times. I was shocked to still be doing 50-55mph as i reached the speed limit signs! A bit more proof the handbrake was over-adjusted i think!

By comparison, on the Sterling i normally let go at the bottom of the dip when doing a similar speed and just about drop to 30mph in time for the speed limit. I'll give that point a try next time i'm out in the coupé and see if that works.

[Laird_Scooby]

Well my snipe went in successfully on that LPG SGi system. A friend is picking it up tomorrow for me as he is going to the place where the kit is and i should have it sometime in the next 7-8 days.

The fun part will happen when it comes to fitting and setting up!

[Laird_Scooby]

The injection LPG system is now safely back in Cambridge with a very jealous friend who can't believe the price i got it for! Think he paid a lot more for his!

On another note, i removed the mixer this morning to measure the venturi in it :

The other one i have here is 42mm diameter so that's a difference of 9mm in diameter. Doesn't sound much but when you consider the area of a circle is PI x R x R where R is the radius, that's a big difference. At 33m diameter the area is 855.3 sq.mm and at 42mm it is 1385.4 sq.mm so 530.1 sq.mm bigger!

I'm sure the extra breathing afforded by the bigger venturi will improve volumetric efficiency (it had better after the grief i had removing the original mixer!), power and the economy and reduce the richness in the mix above 3000rpm - it's basically dumping gas into the inlet manifold without having enough air to burn it effectively.

Will report back later on any improvement if any )

[Shaun]

Digital vernier callipers Dave, A bit easier to read than the ones I use to use as a young apprentice.

[Laird_Scooby]

They certainly are Shaun - can't beat them especially as they're under a tenner these days! Even digital micrometers are pretty cheap too! Found a couple on fleabay for £8 from China and a whole load for about £14 (they look better than the £8 jobbies) and then the sky is the limit.

http://www.ebay.co.uk/sch/i.html?_from=R40&_sacat=0&_nkw=digital micrometer&_dcat=147807&Item%20Type=Micrometers&rt=nc&_trksid=p2045573.m1684

I find verniers easier to use for most of what i need to measure so as yet i don't have a digital micrometer. Somewhere i still have an "analogue" one but i can't remember the last time i needed (or wanted!) to use it.

as for the bigger mixer, on petrol the car is a lot livelier but on gas it's not running right because the Lambda sensor isn't being read correctly. Whether that's my laptop, the Leonardo ECU or yet another faulty Lambda i don't yet know. Going to hook the laptop up to the Jeep soon and see what that gives, i also have a spare Leo ECU so i might just plug that in and try it.I have a feeling i know the problem and what caused it and if so it's my own fault for not making a heatshield for the Leo ECU - but i don't want to jump the gun just yet!

[Laird_Scooby]

Victor Meldrew time - i don't believe it!!!

Yet another rear bank Lambda FUBAR on the coupé! Come to the conclusion the Leo ECU is faulty as if it was down to over-fuelling (due to the tiny venturi in the old mixer) then both Lambdas would go.

The plan is to fit the near-new Leo from the Jeep to the coupé with a new Lambda and an OMVL Millenium (a Leo by another name) to the Jeep as long as it works.

On the good side, after a (very) quick bit of tweaking by ear, i've got the new bigger mixer running reasonably - it won't run properly at the moment due tho the stepper-motor/actuator not moving because there is no Lambda input but is much more like an 827 should be.

At least it's progress, even though it's 3 steps forward and 2 back!

[Laird_Scooby]

Minor update - rechecked the fault code and because the yellow LED was flashing as well as the red one yesterday  misread a single red flash as two thanks to the reflected light from the yellow LED flashing. Not the first time and probably won't be the last but at least no harm done! Also i was hurrying because i was grumpy that a second Lambda had failed.

This puts things in a whole different light as it turns out the Lambda that failed is the other one of the pair i bought in April that failed in short order. Reckon they were a bad batch seeing as the listing disappeared soon after. I've ordered another from a UK seller and it should (hopefully) be with me tomorrow. Then i can reset the ECU and set the new mixer up properly.

[Shaun]

Those digital micrometers look really good.

I did my apprenticeship as a miller many many years ago just before everything went to CNC, worked on some pretty interesting stuff over the years.

[hughezee]

Very frustrating with these lambda issues buddy, hope you get it sorted and it is just a simple case of a bad batch of sensors

[Laird_Scooby]

Thanks Stu - got an extremely weird one with the rear Lambda now except it's nothing to do with the sensor, so far my investigations are pointing towards a wiring problem which isn't good. Narrowed it down to a 2ft section though. Front bank sensor replaced anyway after fighting with the radiator, cooling fans, compressor and hoses etc.

Also found (and ordered) a neat little tool to replace the service port valves on A/C without evacuating the system - more on those in a separate thread when the tool and valves land!

You must be a similar age to me Shaun - CNC was just coming into the mainstream when i was doing my apprenticeship on the black art of electrical & electronic engineering with mechanical engineering thrown in for fun (no paperwork on my mechanical side, just the learning/experience!) and i got to work on lots of interesting stuff too, most of which i can't mention but the one i can is the CD system - that was really interesting and my contribution saved a lot of money and time for the company. Never got the bonus though but hey, i know it and that's what counts for me!

[Laird_Scooby]

Finally got to the bottom of this one after fitting a new front Lambda and trying several different Leo/Millenium ECUs, mutilating my nice, neat wiring and copious amounts of blood, sweat, tears and curses!

Although the rear Lambda is working, for some unknown reason the output has gone weak. In other words, while it still provides a voltage reading into a multimeter, it hasn't got the power behind it to hold that voltage when connected to the Lambda input on the Leo and/or the OE ECU resulting in what is effectively a lost signal.

I ended up stripping the loom back to its wires to check for any wiring faults, none found. While in this state of stripped loom, i decided to make up a jumper lead to link the front Lambda to the Leo which restored normal operation. Now i have a lash up consisting of a wiring adaptor into the front Lambda plug/socket, a wire running to the Leo input and also to the rear OE ECU (PGM-Fi) input to fool the PGM-Fi into thinking it has 2 working Lambdas, although it never showed a fault on the rear Lambda.

At least now i can get on with the final clean/polish/vacuum ready for the show tomorrow and when that's done, i'll tidy up the wiring lash-up and tune the vap to suit the larger mixer which thankfully has cured the excessively rich running at the top end. IT's also less restrictive on the engine so the power output is very much improved!