Fitting the Crank Trigger
There is no appropriate trigger setup on a TVR Griffith. As standard, it uses a 4 tooth wheel and hall sensor in the distributor.
I chose a 36-1 trigger wheel with a standard Ford VR sensor. A kit of parts specifically for my engine was available at a reasonable price. Since I don't have time to fabricate such things and can't take the car off the road for that long (it's a daily driver), this was the best option.
Note that this is the problematic type of wheel (for VR sensor) that has much bigger signal amplitude at the missing tooth. Adaptive hysteresis threshold (adjusted with R181) must be lower (=>worse noise immunity) than with a normal wheel, and not be close to the limits. Schedule a test to check that even +5% amplitude variation (than the actual wheel has) is handled right. See ElectronicDesign/TriggerSignalGenerator
Fitting the trigger wheel took a whole day, as the crank pulley was held together with Imperial bolts in very awkward positions! (They were supposed to be M8 according to the literature) A tool to remove them was fabricated and they were replaced with M8 cap screws.
Apart from this, the physical installation was straight forward. I set the VR sensor to be 1.5mm from the toothed wheel using a feeler gauge. No science here - just an educated guess. Now on to setting the ECU up and testing.
Setting Trigger parameters
Here's the parameters I used in the ECU:
| Edge: | Rising |
| Type: | Multitooth |
| Filtering: | Disabled |
| TDC after trigger (deg): | 62.5 (measured) |
| No teeth on wheel: | 35 |
| Trigger tooth: | 1 |
| Next trigger tooth: | 9 |
| Crank min. period: | 1875 |
| Angular width of tooth: | 3 |
| Angular width of missing tooth: | 6 |
Parameters to calculate above:
| Rev limit: | 7000 |
| Imaginary teeth: | 36 |
| Missing teeth: | 1 |
| Cylinders: | 8 |
| dec: | 3 |
| Engine Phase in "funny" degrees: | 216 |
Testing the Crank Trigger
Bench testing has already been done, so it's time to put the ECU in the car, but only so it can monitor what's going on.
The first time things were connected, the ECU picked up crank rotation and gave a stable speed. The problem is that it is only half the real engine speed. Still, not bad for a first attempt.
Trigger works properly now the config is actually set in the ECU and not just on the PC. Doh!
- Just need to test the polarity on the VR sensor - it doesn't seem to make much difference to how it runs now.
- it's possible that it runs quite well with reverse polarity, or only counts wheel-err sometimes or above certain RPM. Actually, it is possible to configure multitooth "advanced filter" to work properly with reverse polarity VR for all RPM. But avoid it, and go with the right polarity. Publish InputTrigger/TriggerLog if uncertain.
I managed to get a clean "mde40" log file after a lot of bother with the old ECU and doing some maths. The idle was never very good as the old ECU is very crude. Polarity has been established now. As you say, it didn't make much difference - the change in missing tooth length was not much bigger than the RPM change in the engine.
Now it's all installed properly. I have a lot of extra noise now, I think it's due to the wires being closer to the old loom. Must check the shielding connection is OK tomorrow and put the sensor ground in properly.
Shielding problem has been fixed. I replaced the trigger wire with a better quality wire, and now I get 0 wheel errors. At least it was simple.
Finding TDC
The ECU has been used in a "fueling only" setup for the initial period. This means I could get away with only an estimate of the TDC position from counting teeth. (Injector phasing isn't yet important to me) Now I'm doing the ignition phase of the project I need accurate TDC so that the ignition tables are in real degrees.
To find "exact" TDC you need a few tools and some basic knowledge. I found some help on the Internet somewhere:
The fastest way is to use a piston stop and the degree wheel you use to degree in your camshaft. Here's how:
Mount the degree wheel on the balancer with the correct turning tool to be able turn the engine by hand. Mount a pointer on the block that will identify the numbers on the degree wheel.
With the cylinder heads on the engine you will need to use a piston stop that screws in the spark plug hole. It does not matter exactly how far in the cylinder you go just so long as the stop is secure and will contact a flat spot on the piston, try to stay away from any machining edges. In most cases you should not have a problem. If the heads are off you can use a
plate-style stop or a dial indicator. [My piston stop was made from an old spark plug and length of threaded bar. With a vent hole so the cylinder couldn't compress]
Spin the engine slowly "by hand" clockwise until you contact the
piston against the stop. Write down the number you see the pointer aimed at on the degree wheel.
Spin the engine counter-clockwise until you hit the stop again. Again, write down the number the pointer is aimed at.
Find the center of these two marks. (it is usually pretty easy to just split the difference on the wheel). Remove the stop and spin the engine to the correct TDC point on the degree wheel.
Mark your balancer as "0" on your balancer at the zero mark on your timing tab. (This would also be the place to set zero on your timing tape if you decide to use it.)
Secondary (cam) trigger discussion
VEMS has a really good set of individual cylinder features now, so let's use them! Step 1 is a working cam trigger.
Building the cam trigger - new better version!
Looking at the timing cover on my engine during the rebuild, I noticed it was cast to take a cam trigger sensor, but it was never fitted because my cam pulley didn't have trigger teeth.
"But how hard can it be to add them?" I asked...
The answer would appear to be "Not very". I sketched a cam trigger in a CAD tool and had it water jet cut:
Then bolted it on to the engine and fitted a sensor so I get a single pulse per cam rotation.
Configuration
The ECU config is here:
https://dl.dropboxusercontent.com/u/8968424/vems/v3.3_u004371-A-2014.07.04-09.26.05.vemscfg
Note that the config is for a zero mile engine that has run for a grand total of 3 minutes and 45 seconds, so all of the calibration parameters are all over the place and mostly wrong. However the trigger settings should be right. (I know it's still using dual-out ignition - I wanted it to run first time...)
Trigger log
I made a trigger log using the ECU in the above configuration. I think I see cam trigger pulses where I expect to, but would be very pleased to get a second pair of eyes on it because the trigger logging bit of VEMStune is new to me.
https://dl.dropboxusercontent.com/u/8968424/vems/v3.3_u004371-2014-07-04-19.18.03.triggerlog
I hope I have interpreted the trigger log correctly... This chart shows the primary trigger tooth count and the secondary pulse reliably appearing just after tooth #24 on the primary trigger. Not ideally placed - would be better sitting after cyl 2 TDC, but should be OK if I rotate the firing orders to suit.
I assume I configure the firing order to be 4-3-6-5-7-2-1-8 so that the cam trigger resets the cycle at the end?
Back to MembersPage/DavidBlades
Building the "cam" trigger - OLD VERSION
Since the cam is difficult to get at, I thought I'd use the (now redundant) distributor body and a hall sensor instead of the camshaft wheel. They both turn at the same speed (more or less) so I see no problems.
Some spare dizzy parts have generously been donated to the cause. They are for an older Rover V8, so need some modification to fit...
Spare Dizzy parts. Note the original on the top was different to the donated one until it was attacked in the workshop...
With no vac advance needed, the middle plate can be used to mount the hall sensor to look at the remaining spring retainer pin from the remains of the mech. advance mechanism, so I get one pulse every 720 crank degrees.
First camsync trial. Hall sensor is G-clamped to the case!
Unfortunately, this didn't work reliably when tested - I think there was too much ferrous metal spinning round - , so my new idea was to remove the plate completely and make a thin tapered blade with a M6 bolt in the end to fit on the dizzy shaft. This was much more successful and worked well on the test bed. (bench power supply, volt meter and pull-up resistor)
It's actually better balanced than the old distributor was. (Tested unscientifically by spinning it with a drill)
It's back in the engine now, and I just need to align it and do the config.
The nice thing about using the dizzy is that the camsync pulse can be positioned anywhere in the cycle to suit both hardware and firmware.