And relax. I updated the firmware again in the ECU, this time to V73, and I updated the software on the Laptop to v74, and then corrected a couple of errors in the shiftcut map. The previous version of DTASwin didn't have error checking on the Shiftcut screen, and that had allowed me to enter some incorrect values in the map, specifically, Shiftcut start point and Shiftcut end point were both the same values. With these corrected, and the engine running, it is now detecting gear changes, and cutting the ignition accordingly. I also ran it for 25 minutes and it didn't stall once. So the Fiesta is now on the trailer, with the new wheels, and it does only just fit on. All ready for Sundays trip to Santa Pod. I will not be running Flatshift, as it still needs setting up carefully, and the drag strip, in front of several thousand people, is not the place to be testing stuff out.
I rewired the gear position sensor again last night, to use the Sensor Ground connection instead of the ECU case as a ground, and it has totally cured the noise on ana3. So with the fault fixed, and the engine warming up, and me not sitting in the car or touching the gear lever, the engine just cut out on three separate occasions. No warning, it just died. I can see nothing in the ECU error logs that would indicate a fault. I don't think its the Shiftcut thinking there was a gear change, as the data log doesn't show a gear change event occurring when the engine stalls. Very odd. I just happened to be filming when it occurred, and the video is below. It almost sounds as though its run out of fuel. But it restarts and runs fine. I have been advised to upgrade the ECU firmware from V71 to 73, to help the Shiftcut algorithm record more accurate cut times. So first job tonight is to flash the firmware to v73, and see if the problem persists. Not happy that this has started happening. Until I upgraded the firmware from v63 to v71 this never happened before.
Ran the engine last night, with the laptop attached to the ECU, and the Ana3 signal from the gearbox is far cleaner. There is still variance of approx 40mV, but, gone are the 1-200mV jumps that I had before, when the engine was running. I re-calibrated the ECU for the new voltages from the pot so it knows which voltage represents which gear selection. Now I need to try the shift cut again, to see if the reduction of noise allows the ECU to actually see a gear change in progress.
I have rewired the gear position sensor, to use the 5V supply from ECU, instead of the 5V from the Geartronics display. Testing with the ignition on, the display still shows the correct gear from the sensor. I just need to check the calibration of Ana3 on the ECU and then run the engine to see if the interference has disappeared. The 5V supply, pin 9 on the ECU, was feeding just the throttle position sensor, but it now feeds both sensors.
Ran the engine, after refitting the exhaust and the engine steady bar has made the engine rock steady again. I fitted new rod ends on the turnbuckle, and a M10 shoulder bolt, and its eliminated all of the play. I also checked the Ana3 signal from the ECU, and it is still noisey with the engine running so I am now rewiring the gear pot to take the 5V supply from the ECU, which will cure the problem. Ana1 in the video is the traction control signal, which is 4995mV, and very steady.
Diagnosing the interference on Ana3
New engine steady bar fitted
I removed the whole exhaust system inc. manifold last night, so I could then work on the torque link C bracket that fits where the rear of the original engine cradle would bolt to the chassis floor. The C bracket utilises the two original tubes that the engine cradle bolts through, and the bolt hole nearest the link has gone oval, allowing the C bracket to rock back and forth as the gearbox tries to moves. So I've fitted a couple of additional bolts, through the bracket in to the floor, to try to disuade it from moving. I'll refit the exhaust tonight so I can work on the flatshift over the weekend.
I dont know where the months go, but I have another article to write for Retro Ford. I think I'll write about the new wheels and the trip to Apex Festival.
Connected the gear position sensor ground to the ECU ground last night. With the power on, I measured 11mVDC between the ECU case and the 0V from the sensor, so this could well have been the issue with the interference. I also bought a set of male M10 rod ends (LH and RH threads) and a female M10, and some 10mm shoulder bolts, to make the gearbox torque link more robust.
Last night I removed the torque link turnbuckle on the rear of the gearbox, after noticing on Sunday that there was excessive movement of the gearbox when I revved the engine. One of the M10 rod ends is shot, its quite loose, so I'll exchange it for a higher performance one, with PTFE lining which should prevent the problem reoccuring. I've bought a RC car to muck around with. Its from Apex Models, FTX5505, and comes preassembled, and the last time I ran an RC model was over 30 years ago. They've certainly come on a long way. Now it is 2.4GHz instead of 27MHz, the speed controller is integrated in to the servo. A quick charge and we'll see how it goes.
My entry for Prescott Hillclimb on June 14th has been accepted :D
100% sure now that the reason the flatshift isnt working is due to the interference seen on the Ana3 channel. With the engine stopped, everything appears fine. It is not until the engine is running, that the 200mV noise is introduced on to the signal. Allan at DTA suggests that this is a Ground Loop. The dual channel gear position sensor is powered from the Geartronics Display, and not the ECU. So if there are any fluctuations in the 5V rail provided by the Display, then the ECU is going to see this as a fluctuation in the signal. I shall rewire the pot to use the 5V Sensor power rail from the ECU, bypassing the Display.
If I had to hazard a guess, I would suspect that, as the Display is almost certainly strobed, probably at 10Hz, then this accounts for the 10Hz noise. Either way, rewiring the Pot to use the 5V Sensor supply should solve this problem. I know I dont see noise on any of the other analogue channels. Notably the throttle position sensor, which is Analogue. Those signals are rock steady, because they use the 5V Sensor rail, not an external rail.
The split rims are on the car. I took them over to Jeff Fowkes in Ibstock, to fit the 8mm BBS valves and the Dunlop slicks on Saturday, and an hour and a half later I had a phone call saying that they were ready. On Sunday I took the plunge and fitted all four, and with the different ET26 offset, they give the car a far more aggressive stance at the rear. But she wont go on the trailer with the wheels on, so I'll be replacing the 20mm spacers on the axle with 10mm spacers, to bring the wheels back inside the arches. I also bought a new battery from Halfords, which when connected to the trailer winch, does a much better job of running the cable in and out of the reel. I had another go at testing the flatshift feature out, but try as I might, I cannot get the ECU to detect a gearchange. I set the throttle limit to 0%, so theoretically the flat shift should work at tickover. But it does nothing when I pull the lever with the clutch in, even with the engine revs raised, zero effect. The theory is that the signal from the pot is too noisy for the ECU to rely up, so I shall focus on eradicating the noise before doing anything else. The position sensor is fed from the Geartronics display, not from the ECU 5V supply, so thats the first area to work on.
I've stuck a last minute entry in for LDMC Prescott Hillclimb on June 14th. Fingers crossed. Places were running out, but with luck I'll be squeezed in. Sunday is already full, but Saturday had a handful of spaces midweek.
I've altered the stylesheet on the website to try to make the fonts appear correctly on Android Chrome. I'm not using anything fancy, but when viewing on Chrome the fonts do seem to take on a life of their own. I've now passed two years working for Rolls Royce, and the work has almost dried up. If you know of any contracting opportunities, please get in touch. I saw the company Spitfire being put through its paces over Derby yesterday, as Rolls celebrated the shipment of the first Trent XWB engine to Airbus. I've seen these engines being built, and they are collosal. In fact, I've been to all the plants in the UK and seen all the parts being made. Its been quite a privilege working for them over the past couple of years.
Connected the laptop up to read the data logs from the ECU but of course when I loaded the 2013 map on to it after experiencing issues after the run at Lydden, the data logs were cleared! Doh! Not to worry. I've put the new map back on and I shall be testing the shiftcut again this week.
The videos from Lydden are now up on Youtube and Facebook.com. I've uploaded one at x20 speed, and one at real time. Oddly, you can see the refelection in the windscreen of the gear position indicator, albeit reversed, so you can see the gears I'm using.
AllanW at DTA has recommended that I try flat shift again, but this time with the 'No Shift Cut below Throttle' set ot 0% and the shift cut time delay set to 500mS, with the engine running in the garage, and it should show then if the flat shift is working. There are no reported issues with the Reverse Pot Voltage feature set to on. I need to take the car off the trailer anyway, to fit the new wheels, so I'll have to try that out.
|Apex Festival May 2014|
Had a mixed day at the Apex Festival yesterday. I couldnt get the flat shift to work, and I had what I thought was a misfire. The misfire turned out to be the 'Low temperature RPM limiter' working, but for some reason the ECU thought the water temp was below 20C when it wasn't, which is why it kicked in. After a few laps, the rpm limit went, and I could then use the full revs. (I've turned this feature off for now) And I'll check the logs in the ECU to see what temperature it thought was being registered. I know the ECU water temperature sensor works, because the cooling fans come on at 75C. And the dashboard was showing 50-60C when I went out, so the engine was definitely warmed up.
The track action was pretty frenetic. The first session, I had to wait for the low temperature rpm limiter to clear. So only after the first couple of laps, was I able to use full revs. With the tyres as well as the engine warmed up, I set off after the brace of VW Golfs, and had a good few laps following them round. Once past I then caught and overtook the 2.0 Nova and Porsche, which both held me up for a while. All the time, the blue Mk1 and Mk2 Escorts were trying to keep up, and everytime I caught traffic my rear view mirror was full of blue Fords :D And then the session was over, so I never got to do a single flat out lap, which was disappointing.
A long tow there and back again, for one session. If the weather had been as good as last year, I think it would have made for a better show. The Superwinch worked, just, because the battery that I had charged for four days, was flat, but it had just enough juice left to drag the car on to the trailer. As for why flat shift wouldnt work, there are so many parameters to tweak, and it was not the right environment to wander out on to the track for testing. I'd have been in the way.
The BBS slim tyre valves turned up last night, but I doubt I'll have time to get them fitted for the weekend. The forecast is quite poor for Sunday, with rain expected. So I'd be better off running the Kumho's instead of the new slicks, which I'll now keep for the Classic Ford event at the beginning of June.
Below is a picture of the Superwinch fitted to the trailer. I have added an Anderson connector to the cable, so it can be plugged in to the jump start battery, which is now on charge ready for Sunday. The 10mm axle spacers have turned up, but no sign of the slim tyre valves as yet.
The June 2014 issue (#99) of Retro Ford magazine is now out, and I get a great two page spread with my Rockingham report. I've also been in touch with Omex regarding the new engine build, and am now waiting to hear their thoughts on the components I've bought, and the possibility of assembling the engine for me.
So a couple of snags with the fitment of the new wheels. One, they didnt come with tyre valves, and when I turned up at National Tyres on Saturday morning, I was greeted with blank looks as they'd never see a rim with an 8mm hole in it before, and they wouldnt even attempt to fit the modern 11mm valves. Probably just as well. A lot of phone calls later, I secured a set of 8mm BBS valves on Ebay which should hopefully arrive this week, and sort that issue out.
The second snag, is that the split wheels arent ET34 as advertised, but are ET26. The issue that causes me, is that I run Ford Racing Puma 20mm spacers on the Focus rear stub axles, and no spacers on the wheels themselves. I havent physically got the room on the trailer to run wheels that would be effectively 22mm wider than the overall width of the rear axle as it stands at the moment. So I'd have to remove the 20mm spacers from the stub axles, and replace them with 10mm spacers. So I'm going to have to accumulate a few bits before I take the plunge and fit these rims. At the front, the problem isnt so bad. I run 10mm hubcentric spacers with the Superleggera's (ET34) so I can fit the splits, and run without spacers, no problem.
With that disappointment out of the way, I loaded the new map on to the ECU, and calibrated the ECU for flat shifts. This entails connecting to laptop to the ECU, running the DTASwin.exe software, and selecting the Shift Cut screen. Now, with the gear position sensor 2nd channel wired up to Analogue3 (ANA3 on the ECU), I could see again the voltages from the sensor. Then, each gear is selected in turn, and the voltage from the sensor is programmed in to the ECU memory. This process is repeated several times, until the voltages are repeatable. Once setup, F4 is pressed and the ECU map updated. So now I can try flat shifting again at Lydden on Sunday. I have updated the DTA S80 page with the screen shots from the ECU software.
I then fitted the Superwinch 3000 to the trailer, to replace the manual winch, which is now over 10 years old. The Superwinch runs from the spare battery I take with me, and should make loading and unloading a lot simpler, and a lot less strenous, than before.
I updated the firmware in the DTA S80 from v63 to v71 last night, and filmed a quick Youtube video showing the process.