'79 CIS TO EFI

[Bruce M]

Fabulous.

Note: if you get a misfire the WBO2 will read super lean because none of the O2 gets burnt up.

[PeterK]

Thx all, exciting times for sure

[PeterK]

After the high of yesterday, confusion today. I’m trying to balance my ITBs and can get 5 roughly the same, but cylinder 2 is not really changing.

I cleaned the spark plug and checked that it was sparking (outside of the engine). When I pull off the injector connector, the revs dropped. It did seem to be spitting back out of the ITB a lot, so I removed them, and used my inspection camera down the manifold. The valve is opening and closing (I did wind the engine over with a spanner, not the starter !) and there doesn’t appear to be any foreign bodies down there.

Car starts fine, but does sound like it’s missing a cylinder, but as it is probably so far out of tune (rhs showing almost 20% O2, lhs only a bit), who can tell.


I have noted that where the ITB throttle actuator (the bit I took off to weld an extension to) attaches , there is a plastic washer, then a metal one, then the nut. The rhs washers all look concentric, whereas the lhs do not. Just pondering (any straw to clutch), whether I might have an air leak here. A job for tomorrow.

Any thoughts ????

[Bruce M]

Swap the injector for another one on the opposite side, just to eliminate that. Air leak is a good bet. Does it even out with a bit of throttle/ revs (when a small leak makes less difference).

[PeterK]

Tried with a synchronometer which showed massive airflow. So my thinking at the moment is a bent butterfly. I’ll test with some smoke (not from wires) later Wednesday.

[PeterK]

What a difference a day makes – we’re back to ups rather than downs.

I finished yesterday on a bit of a low, so I cracked on and fixed my battery isolation / master switch issuette – the escaping smoke. This was the damage done


And this is what it looks like now


Yesterday, before my tuning downer, I also modified my throttle swivel to give more throttle action. I engaged the assistance of SWMBO to press the keyboard button to accept full throttle, as I opened the throttle in the engine bay, checking to see when the butterfly(s) were vertical. With closed / zero also set, I could use my tablet, bluetoothed from the ECU, to check what throttle opening I was getting when either under the engine lid or in the driver’s seat. I get 97% and a little bit now, so that’ll do for me.

To modify the swivel, I ground away a slot to allow the throttle actuating rod to be closer to the pivot, thus giving more swing (does that even make any sense ?)




Onto today, and having given my issue some pondering overnight, I Set to with the magic smoke. It was very revealing. This was cylinder 2 with the throttles closed – not ideal


Some fettling followed, and thankfully, I now have vacuum, and so balancing of the ITBS ensued. Next up was tickover, which was about 500rpm, so somewhat lumpy. I did try to adjust the stops, which are grub screws, located under the black plates (just above the upside down 5)


The one really didn’t want to play and was starting to round itself out, so time for another Plan B. I remove the plates / bars, then could more comfortably encourage the grub screws out


To replace them with M5 bolts and locking nuts


This made adjustment a lot easier, and now with a tickover of about 900-1000 rpm, the engine is a lot smoother. I tweaked the linkages to each bank from the swivel, and the linkage coming from the pedal. Hopefully it is close enough for now.

I’ve still got a problem with my left bank lambda sensor (too lean, so it shuts itself off), so tomorrow I might try swapping left for right and see if the issue moves with the sensor. I also checked the fuel pressure with the engine running and it was spot on, and with the timing locked at 10’ and the timing light set to 10, I reset the static timing.


So, the car starts on the button, stumbles a little if I give it too much throttle too quickly, although I do have acceleration enrichment turned off (as suggested by Aaron), ready for test driving. I’ve been tasked with gently driving at fairy constant revs, with logging on, then stopping to check what lambda adjustment the ECU suggests, then modify the VE Table, and repeat.

As of now, the fuse and relay area look like this (bare)


And like this put to bed


Finally, with the trumpets and air filters on, I give you a 911SC 3.2SS with ITBs



The eagle-eyed will notice that
- I still have the lead to allow the timing light to be attached. Unless I have issues with it, it will stay until the engine has been dyno-tuned
- the fuel return pipe loops across the back. Once I am confident that the fuel pressure is where it needs to be, I will remove the gauge and feed the returns from each bank to opposite sides of the FPR, allowing the pipework to be neatified (my blog, I can make up words if I want to)
- the vacuum hose follows the fuel return. It too will be tidied up once the fuel return is sorted.

[sladey]

Nice work - engine is certainly looking cool

[911hillclimber]

Mega project.
When I contemplated all this for the hillclimb Lola engine years ago I realised I would not have the stamina to complete it so went for a set of lightly used PMO carbs which i don't regret for the car.
Fab job almost completed now, v interesting to watch but outside my skill base!

I'll be particularly interested in your take on road drivability when all done Peter.

[Darren65]

Great work Peter, looking forward to hearing how it drives.

[PeterK]

Sorry that it’s been a while since my last update, but I have been riding the ups and downs of tuning. When driving, my left bank was very lean, spiking to a (max) 2.4 lambda, while the right was very rich.
I tried logging the car, so that I could auto-tune sections back at home. That was another learning curve – apparently you can’t use autolearn with dual VE maps, so set to a single lambda for tuning. It also doesn’t help if you have lambda control left on – the car is doing its head in trying to correct. So out for another test drive, but then autotune, because the left O2 sensor (the very lean reading) is the one the controller is looking at, the map correction is about 400% at 2000 rpm. Not likely.
To check, I took the car to my friendly local MOT man, we stuck a probe up the exhaust (they don’t like it up ‘em, Captain Mainwaring), and just as I could smell, his machine confirmed that it was very rich – high 0.7 lambda.

I checked for a cylinder not working, and gave all of the coil and injector connectors a good tug. A bit too hard on cylinder 3, as I pulled the trigger wire out. Re-crimped. Still high lambdas on the left and low on the right. Played around, and somewhere in the 10 to 15% less fuel than Phill / Lee’s 911 on Triumph ITBs seems least bad.

More pondering needed. To distract myself, I fiddled with the throttle swivel thingy. To increase the throttle opening, I tried making the lower arm (actuated by the throttle) shorter, so that the ratio of butterfly opening to ITB throttle movement would increase.


Unfortunately that was less successful than expected, as the throttle couldn’t pull the shortened lower arm enough – it kind of locked out in an almost straight-line. Time for another Plan B. I extended the top to give more leverage, and gave myself a number of options for the lower lever, but ended up removing this extension.


So now I have full throttle travel (the pedal could go slightly further, but is limited by the ITBs. I checked the ECU TPS zero and maximum settings, although they didn’t really change much.

Still had lean left & rich right, so time for another ITB balance. Not sure whether I mentioned this before, but as I’m too lazy to go back & check, here it is anyway (again). I use a Morgan Carbtune, that has 4 sections. I start by connecting ports 1,2 & 3 to cylinders 4, 5 & 6 and port 4 to the cylinder 2. I linked cylinders 1 & 3 so that they weren’t pulling air from the MAP manifold.


I could then balance cylinders 4, 5 & 6 to one another (I did this side first, as it has the TPS) and to cylinder 2. I then swapped the Carbtune to cylinders 1, 2,3 & 5, to repeat for the other side.

Test drive and still very lean left and rich right.

New plan – replace spark plugs, just in case they are getting too fouled. Fitted resistor versions (NGK BR9ES) of the ’79 SC recommended plug
Guess what – no, still lean left, rich right.



New plan – replace the lean O2 sensor. With a new to me but used one. Less error messages, but still lean left, rich right. Maybe time to give up and let the experts sort it on the dyno. Unfortunately it’s a bad time of the year - earliest possible date is 25th Feb.



Hey ho. Yet another Plan B (must be on about Plan Q by now). I swapped the rich sensor (CANbus connected, on the right side) to the left side (internal ECU controller) and refitted the original suspected ‘dodgy’ sensor on the right. I checked for air leaks in my welding but couldn’t see any. I did a tug test on all of the O2 sensor connections, again maybe a little aggressively. Re-crimped the IP connection.
Still lean left, rich right.

OK, maybe I have a missing cylinder – lo & behold, cylinder 3 is not totally at the party. Check the connectors. I must learn to be more gentle – re-crimped the trigger wire. Pulled the spark plug and cleaned it. Checked that I had sparks – all good there. What else – maybe an air leak elsewhere. There was a smattering of blackness under the car both sides – the SSI to silencer bolts took an extra quarter turn of tightening. Cylinder 3 (and 6) have the brake booster take-offs. Can’t be a booster problem, or both 3&6 would have issues. However, whilst poking and pondering, I did notice that the take-off from cylinder 3 was not properly tightened. Tighter now.

The car was now running much better, left still lean and rich still rich, but significantly closer. So what about left to right balance now. Rather than getting the Carbtune out, I just loosened the linkage to the left cylinders – buy tweaking it I could get the left and right lambda readings to get close, and even swap, so that left was richer and right leaner. Houston, I think we may have cracked this. A bit of playing, and left and right banks are broadly the same lambda, at both tick-over and at higher revs. Result. Reset tick-over down a smidge.


So, relative to Phill / Lee’s SC, I have 15% less fuel over all of the map, with an additional 12.5% less in the range 1-3% throttle, upto 2000 rpm, giving a very close to lambda 1 at tick-over and at 2k revs, on both banks.
Just need to test drive now, but in my life of Efi ups and downs, we’re currently at an up, hoping that its now good enough to pass the MOT (due by the end of the month – yes I do know that it’s exempt, but my friendly testers give me a useful validation that nothing is about to fall off – and they have a brake testing machine).

[The Garage Bitch]

Well done mate

With so many variables it would be easy to keep chasing your tail trying to get everything spot on

Temp (at least) good enough readings to go for the run on the dyno is definitely an "up" day

[impmad2000]

Setting mine up, I always used this type of synchrometer. I tried the vacuum based ones and the needles just jumped around too much.
Great to hear you are getting to the root of the issues. I did spend ages making sure both sides were opening evenly and maintaining a balance.
Great work

[PeterK]

I started with my Synchrometer, but once I had the trumpets & air filters fitted, the Carbtunen was easier. I'll probably remove the trumpets, or at least the air filters once I have an MOT out of the way & go back to the Syncrometer. I do need to ensure that my throttle swivel sides are exactly the same length to avoid changing balance with % throttle opening (I think the lengths are within a mm, but that is not necessarily accurate enough) - the top plate is just a 'try it' for now. Today I changed the fueling map to Phil/Lee's less 30% all over. Still a little rich, but don't want to go too lean, as I have no acceleration enrichment turned on currently.

[impmad2000]

The two reasons for me choosing to use the cables as per the original Triumph bike is
A) it keeps the pull even over the actuating arc, i.e. no non linearity at the extremes of throttle travel
B) by using a slide to pull two same length cables, both should remain synchronised as pull is even on both.
I used a wooden board which I bolted the TBs onto simulate the engine on the bench. I could pull the throttles and measure the throttle angles on both banks with the whole setup on a test bench. It helped me but you are up and running already.

[PeterK]

Will you send me some details please Tim
I’ve got the cable pieces that bolt onto the ITBs in my box of bits, so could still go that way