Megasquirt and shunting.
Discussion
Well to say I was disappointed would be an understatement! Particularly since the car had never shunted before.
So for anyone researching the issue, as I was, I thought it worth posting.
I am grateful to previous posters on this subject.
Since the engine has had some changes, it may be that these have initiated the shunt.
As a background, it's a 5ltr, 52,000 mls and was using oil, had uneven compression test figures and a worn camshaft. On stripdown, the oil use was clearly due to the valve stem seals, the cam was badly worn on a number of lobes and the exhaust valves were badly coked, presumably due to the oil, and looked to be the cause of the uneven compression test. The bores were barely worn, still showing a good honing pattern and the pistons only slightly scuffed. All bearings showed no more than expected wear except that the surfaces of two camshaft bearings were starting to break up.
On John Eales advice, the rotating assembly was internally balanced, the block pressure tested then bored to 94.5mm and new Omega pistons fitted. I had the valve guides k-lined, new oil seals and the existing valves had the seats and faces re-cut. I chose a Stealth cam, after positive reports on here, but with according to Rob, a slightly tighter LCA. The squish is 40thou and the compression ratio 10.8:1. Cometic gasket and ARP studs. 72mm throttle, standard 500 stacks, opened up inlet manifold tracts and all ports matched. 5" air intake to an enclosed cone filter in place of the AFM and straight on to a smooth bore elbow. TPS replaced by a Colvern TPS. 240cc Vectra injectors. Remaining pre-cats removed. All in all, nothing out of the ordinary apart from perhaps the CR.
So to the point.....
I started off with a fairly standard sort of map using speed density, 17.7ms required fuel, batch fire, 2 squirts per cycle alternating. VAG wasted spark logic coils. 14 deg ignition advance gave a nice smooth idle with 13.5AFR. nothing unusual in the ignition or fuel maps.
Shunted like a pig, otherwise drove well.
The various cures appeared to be richer fuel, more advance, less advance and a mixture of these but, strangely, no one cure for every engine. However, all the remedies aim to kill torque in the shunting zone.
I found some difference using a 12.5AFR mixture but not a cure and I didn't really want to run this rich anyway. Likewise, advancing the ignition in this area provided little improvement and made the car feel jittery, particularly at junctions etc.
The cure in my case was to significantly retard the ignition in the shunting zone. My kpa at idle is around 60. At the load sites from 46kpa to 66kpa and up to 1100rpm I have 5 deg advance. At the same load sights up to 1500 rpm it's 6deg and averages 13deg (varies a bit according to load) by 1900rpm and rises quite sharply thereafter.
Thankfully, the car is once more a pleasure to drive.
I hope this information may help others, although I appreciate that it is not a cure-all, just a cure for this particular engine.
Dougal.
So for anyone researching the issue, as I was, I thought it worth posting.
I am grateful to previous posters on this subject.
Since the engine has had some changes, it may be that these have initiated the shunt.
As a background, it's a 5ltr, 52,000 mls and was using oil, had uneven compression test figures and a worn camshaft. On stripdown, the oil use was clearly due to the valve stem seals, the cam was badly worn on a number of lobes and the exhaust valves were badly coked, presumably due to the oil, and looked to be the cause of the uneven compression test. The bores were barely worn, still showing a good honing pattern and the pistons only slightly scuffed. All bearings showed no more than expected wear except that the surfaces of two camshaft bearings were starting to break up.
On John Eales advice, the rotating assembly was internally balanced, the block pressure tested then bored to 94.5mm and new Omega pistons fitted. I had the valve guides k-lined, new oil seals and the existing valves had the seats and faces re-cut. I chose a Stealth cam, after positive reports on here, but with according to Rob, a slightly tighter LCA. The squish is 40thou and the compression ratio 10.8:1. Cometic gasket and ARP studs. 72mm throttle, standard 500 stacks, opened up inlet manifold tracts and all ports matched. 5" air intake to an enclosed cone filter in place of the AFM and straight on to a smooth bore elbow. TPS replaced by a Colvern TPS. 240cc Vectra injectors. Remaining pre-cats removed. All in all, nothing out of the ordinary apart from perhaps the CR.
So to the point.....
I started off with a fairly standard sort of map using speed density, 17.7ms required fuel, batch fire, 2 squirts per cycle alternating. VAG wasted spark logic coils. 14 deg ignition advance gave a nice smooth idle with 13.5AFR. nothing unusual in the ignition or fuel maps.
Shunted like a pig, otherwise drove well.
The various cures appeared to be richer fuel, more advance, less advance and a mixture of these but, strangely, no one cure for every engine. However, all the remedies aim to kill torque in the shunting zone.
I found some difference using a 12.5AFR mixture but not a cure and I didn't really want to run this rich anyway. Likewise, advancing the ignition in this area provided little improvement and made the car feel jittery, particularly at junctions etc.
The cure in my case was to significantly retard the ignition in the shunting zone. My kpa at idle is around 60. At the load sites from 46kpa to 66kpa and up to 1100rpm I have 5 deg advance. At the same load sights up to 1500 rpm it's 6deg and averages 13deg (varies a bit according to load) by 1900rpm and rises quite sharply thereafter.
Thankfully, the car is once more a pleasure to drive.
I hope this information may help others, although I appreciate that it is not a cure-all, just a cure for this particular engine.
Dougal.
Interesting post D, I’m surprised retarding the ignition made that much difference. Folk usually add a bit more fuel in the shunting zone and some have advanced the timiming but I’m no expert. Great it’s worked for you. Mine is going ok but I’m always interested in ways to make the cars smoother below 2k rpm.
I'm glad you found a solution. That required fuel sounds quite high, those injectors must be quite small. Mine is 7.71 using Astra VXR injectors 470cc. I'm currently upgrading mine to an MS3 Pro to get more injector outputs as I've always wondered if the batch firing 4 injectors at a time was contributing to the issue.
Have you got any pictures on the VAG coils? I might be using the same ones, if they are the block of 4 units.
Also do you find you smell of fumes after driving with those low AFR?
Have you got any pictures on the VAG coils? I might be using the same ones, if they are the block of 4 units.
Also do you find you smell of fumes after driving with those low AFR?
BoostedChim said:
I'm glad you found a solution. That required fuel sounds quite high, those injectors must be quite small. Mine is 7.71 using Astra VXR injectors 470cc. I'm currently upgrading mine to an MS3 Pro to get more injector outputs as I've always wondered if the batch firing 4 injectors at a time was contributing to the issue.
Have you got any pictures on the VAG coils? I might be using the same ones, if they are the block of 4 units.
Also do you find you smell of fumes after driving with those low AFR?
I wasn't sure whether the OP meant paired injectors when using the word batch (i.e. 4 channels) rather than banked injection (2 channels, either left and right bank or double-U). Have you got any pictures on the VAG coils? I might be using the same ones, if they are the block of 4 units.
Also do you find you smell of fumes after driving with those low AFR?
The original injectors are only around 185cc and have a req_fuel of about 21, so the 240's look about right at 17. There's plenty of range in the fuel maps to accommodate a range of req_fuel numbers anyway; it doesn't have to be exact.
I'm also using VAG coils - 032905106B - logic level so built in igniter.
I found switching from the 2 original bank based channels (running aftermarket ECU) to 4 paired channels on the same ECU produces an absolutely massive difference. It runs a lot smoother, is more powerful, and the idle quality is a lot better. I'm not intending to switch to full sequential though.
lancelin said:
Interesting post D, I’m surprised retarding the ignition made that much difference. Folk usually add a bit more fuel in the shunting zone and some have advanced the timiming but I’m no expert. Great it’s worked for you. Mine is going ok but I’m always interested in ways to make the cars smoother below 2k rpm.
I've never (that I can think of) had any success advancing the ignition to rid the shunting, retarding always works in my experience.When you start to think about what's going on it makes more sense too.
Retarded timing forces you into using larger throttle openings, this reduces the vacuum in the plenum so reducing the tendency for exhaust to come into the plenum on overlap.
Larger throttles also means more airflow and air in the plenum, so any exhaust reverse flow is trying to contaminate a larger volume of fresh air.
Larger throttles also give higher airflows in the ports, so improving the fuel mixing, and air speed into the chamber.
A richer mixture, and retarding the timing, always works to improve things for me.
Boostedchim.
I believe you should aim for a max injector duty of around 80%, this should ensure a reasonable open time at a low duty for better control. My max duty is 86%, so perhaps the 240cc injectors are very slightly small. I would however prefer to be on the small side for low down tuneability.
Here's a pic of the coilpacks. (I'll answer the anticipated comments now!! -No trouble with proximity to exhaust.....so far!!)
ric355
No smell of fuel but I don't see an AFR of 13.5 particularly low at idle and just above for these engines???
I believe you should aim for a max injector duty of around 80%, this should ensure a reasonable open time at a low duty for better control. My max duty is 86%, so perhaps the 240cc injectors are very slightly small. I would however prefer to be on the small side for low down tuneability.
Here's a pic of the coilpacks. (I'll answer the anticipated comments now!! -No trouble with proximity to exhaust.....so far!!)
ric355
No smell of fuel but I don't see an AFR of 13.5 particularly low at idle and just above for these engines???
Sorry ric355, you were asking about injector banks. Batched is megaspeak for anything other than semi or full sequential.
I only have 2 injector channels so it must be 4 at a time. I could choose which 4 but I just stuck to alternate banks - for no good reason! Interesting regarding pairs of injectors but surely the difference here is that rather than whether the injectors fire in pairs or fours, or whether it's by bank or 2 from each bank, you must squirt 4 times per cycle whereas I only squirt twice. I will try four times per cycle and see if there's a difference.
I only have 2 injector channels so it must be 4 at a time. I could choose which 4 but I just stuck to alternate banks - for no good reason! Interesting regarding pairs of injectors but surely the difference here is that rather than whether the injectors fire in pairs or fours, or whether it's by bank or 2 from each bank, you must squirt 4 times per cycle whereas I only squirt twice. I will try four times per cycle and see if there's a difference.
Dougal9887 said:
Sorry ric355, you were asking about injector banks. Batched is megaspeak for anything other than semi or full sequential.
I only have 2 injector channels so it must be 4 at a time. I could choose which 4 but I just stuck to alternate banks - for no good reason! Interesting regarding pairs of injectors but surely the difference here is that rather than whether the injectors fire in pairs or fours, or whether it's by bank or 2 from each bank, you must squirt 4 times per cycle whereas I only squirt twice. I will try four times per cycle and see if there's a difference.
I think it's a little more subtle than that. With only 2 channels you can't achieve a consistent injection point as your four cylinders are in quite different points in the cycle when the injection event occurs. Especially if it's left bank and right bank channels. I think this is what's behind the "double-u" pattern, as it is a more consistent approach.I only have 2 injector channels so it must be 4 at a time. I could choose which 4 but I just stuck to alternate banks - for no good reason! Interesting regarding pairs of injectors but surely the difference here is that rather than whether the injectors fire in pairs or fours, or whether it's by bank or 2 from each bank, you must squirt 4 times per cycle whereas I only squirt twice. I will try four times per cycle and see if there's a difference.
"Paired" is better still - it's the same order as the wasted spark ignition so has a better consistency of cycle point at the moment of injection. I have 2 squirts, alternating.
Did you get good dead time / voltage correction info for your Vectra injectors? I was thinking of changing to them but could find no technical details for them.
I originally went with 2 banks and used only 2 channels so that I didn't have to create a new injector loom, but in the end I went ahead and created one and used all four channels, and the result was like night and day.
ric355 said:
I think it's a little more subtle than that. With only 2 channels you can't achieve a consistent injection point as your four cylinders are in quite different points in the cycle when the injection event occurs. Especially if it's left bank and right bank channels. I think this is what's behind the "double-u" pattern, as it is a more consistent approach.
"Paired" is better still - it's the same order as the wasted spark ignition so has a better consistency of cycle point at the moment of injection. I have 2 squirts, alternating.
Did you get good dead time / voltage correction info for your Vectra injectors? I was thinking of changing to them but could find no technical details for them.
I originally went with 2 banks and used only 2 channels so that I didn't have to create a new injector loom, but in the end I went ahead and created one and used all four channels, and the result was like night and day.
Interesting. To to get the pairs timed are you using fully sequential but for four cylinders just like wasted spark? I see MS2 does support 4 injector channels, with internal mods, but I also would have to create a four channel loom. If it's as good as you say, well worth the effort."Paired" is better still - it's the same order as the wasted spark ignition so has a better consistency of cycle point at the moment of injection. I have 2 squirts, alternating.
Did you get good dead time / voltage correction info for your Vectra injectors? I was thinking of changing to them but could find no technical details for them.
I originally went with 2 banks and used only 2 channels so that I didn't have to create a new injector loom, but in the end I went ahead and created one and used all four channels, and the result was like night and day.
I haven't any info on the Vectra injectors except that they are high resistance, can't remember the figure, so use a dead time of 8ms.
Edited by Dougal9887 on Tuesday 11th June 08:01
Dougal9887 said:
Interesting. To to get the pairs timed are you using fully sequential but for four cylinders just like wasted spark? I see MS2 does support 4 injector channels, with internal mods, but I also would have to create a four channel loom. If it's as good as you say, well worth the effort.
I haven't any info on the Vectra injectors except that they are high resistance, can't remember the figure, so use a dead time of 8ms.
Each channel fires 2 injectors so it's not fully sequential (that would be 8 channels). I'm not on a megasquirt so the terminology may be slightly different. The pairing is exactly the same as wasted spark: (1,6), (8,5), (4,7), (3,2) . And the ECU simply fires the channels 1-2-3-4 over the cycle. I'm on 2 squirts so each is actually fired twice evenly spaced. I'm pretty sure this is how most of the aftermarket systems are configured. I haven't any info on the Vectra injectors except that they are high resistance, can't remember the figure, so use a dead time of 8ms.
Note that getting the injector close angle correct makes a big difference on smoothness too.
I spent a bit of time looking at the original lucas firmware and could see it does 2 squirts per cycle (once every 4 ignition events) but I was not able to work out exactly how it timed the fuel injection event in relation to the ignition event. The reason I was looking at that was that on 2 channel left/right bank injection I could not get my ECU to idle any better than the 14CUX despite having better control over ignition and fuel. Which is ironic considering how lumpy it is normally! However with 4 channels it is as I said before like night and day. Clearly the VE and ignition maps have a lot to do with it but without any changes it was noticeably stronger in the lower ranges and smoother at idle. My car is entirely road mapped by myself, so I'm not pushing the absolute limits on advance.
Obviously you mean 0.8ms for dead time ;-)
ric355 said:
Each channel fires 2 injectors so it's not fully sequential (that would be 8 channels). I'm not on a megasquirt so the terminology may be slightly different. The pairing is exactly the same as wasted spark: (1,6), (8,5), (4,7), (3,2) . And the ECU simply fires the channels 1-2-3-4 over the cycle. I'm on 2 squirts so each is actually fired twice evenly spaced. I'm pretty sure this is how most of the aftermarket systems are configured.
Note that getting the injector close angle correct makes a big difference on smoothness too.
I spent a bit of time looking at the original lucas firmware and could see it does 2 squirts per cycle (once every 4 ignition events) but I was not able to work out exactly how it timed the fuel injection event in relation to the ignition event. The reason I was looking at that was that on 2 channel left/right bank injection I could not get my ECU to idle any better than the 14CUX despite having better control over ignition and fuel. Which is ironic considering how lumpy it is normally! However with 4 channels it is as I said before like night and day. Clearly the VE and ignition maps have a lot to do with it but without any changes it was noticeably stronger in the lower ranges and smoother at idle. My car is entirely road mapped by myself, so I'm not pushing the absolute limits on advance.
Obviously you mean 0.8ms for dead time ;-)
Yes 0.8ms!Note that getting the injector close angle correct makes a big difference on smoothness too.
I spent a bit of time looking at the original lucas firmware and could see it does 2 squirts per cycle (once every 4 ignition events) but I was not able to work out exactly how it timed the fuel injection event in relation to the ignition event. The reason I was looking at that was that on 2 channel left/right bank injection I could not get my ECU to idle any better than the 14CUX despite having better control over ignition and fuel. Which is ironic considering how lumpy it is normally! However with 4 channels it is as I said before like night and day. Clearly the VE and ignition maps have a lot to do with it but without any changes it was noticeably stronger in the lower ranges and smoother at idle. My car is entirely road mapped by myself, so I'm not pushing the absolute limits on advance.
Obviously you mean 0.8ms for dead time ;-)
So each of your pairs squirts twice per cycle; at each squirt (assuming correct phasing), one injector is firing at an intake valve which is just about to open on the intake cycle, and the other is firing at a valve which is closed, as that cylinder is in the compression/expansion phase, and then on the next squirt, the opposite? So just like wasted spark.
At present my injection is untimed, 2 channels merely firing left and right banks alternately.
Dougal.
Dougal9887 said:
Yes 0.8ms!
So each of your pairs squirts twice per cycle; at each squirt (assuming correct phasing), one injector is firing at an intake valve which is just about to open on the intake cycle, and the other is firing at a valve which is closed, as that cylinder is in the compression/expansion phase, and then on the next squirt, the opposite? So just like wasted spark.
At present my injection is untimed, 2 channels merely firing left and right banks alternately.
Dougal.
Yes, that's it.So each of your pairs squirts twice per cycle; at each squirt (assuming correct phasing), one injector is firing at an intake valve which is just about to open on the intake cycle, and the other is firing at a valve which is closed, as that cylinder is in the compression/expansion phase, and then on the next squirt, the opposite? So just like wasted spark.
At present my injection is untimed, 2 channels merely firing left and right banks alternately.
Dougal.
Once changed over I had to take a bit of fuel out due to the increased efficiency. And I had to take a bit more out after having got the injector close angle tuned. It fixed an inconsistency in the fueling I'd had at cruise, which was the main reason for the change. Something I was not expecting was that reapplication of the throttle after a gear change is now much smoother also.
Dougal9887 said:
You've got me hooked, I can feel the next mod coming on 
How did you go about tuning the injector timing? I would need to have a closer look but I think MS uses an angle v. rpm table, then takes into account fuel and ignition load calcs.
My ECU uses a single value for close angle across the entire RPM range. It uses that together with the current crank angle and speed to work out when the injection event should start in order to ensure it finishes at the specified angle. Not sure if MS is the same in that sense but there are a couple of ways to work out the close angle given the way mine works. How did you go about tuning the injector timing? I would need to have a closer look but I think MS uses an angle v. rpm table, then takes into account fuel and ignition load calcs.
The first is to use the camshaft specification to work out the angle at which the inlet begins to open, and set the angle to a few degrees before that.
I used the other method which, I guess would need to be done at multiple RPMs if you have a table to populate, is to run the engine and monitor the AFR whilst changing the close angle. Look for the point where it goes richest, as this is the point where you are getting the most fuel burned. In my case I worked backwards from 359 degrees and ended up with 295 degrees but the numbers are going to depend exactly on how MS works. I have a standard 450 cam, and looking at the specs this is pretty much bang on where it should be (I did not know the spec when I tuned it).
Dougal9887 said:
The cure in my case was to significantly retard the ignition in the shunting zone. My kpa at idle is around 60. At the load sites from 46kpa to 66kpa and up to 1100rpm I have 5 deg advance. At the same load sights up to 1500 rpm it's 6deg and averages 13deg (varies a bit according to load) by 1900rpm and rises quite sharply thereafter.
Thankfully, the car is once more a pleasure to drive.
Dougal.Thankfully, the car is once more a pleasure to drive.
Thanks for sharing your ignition regarded timings, what’s AFR now at idle and in the shunting zone?
Joolz,
Thanks for you explanation, I see.
Dougal9887 said:
Boostedchim.
I believe you should aim for a max injector duty of around 80%, this should ensure a reasonable open time at a low duty for better control. My max duty is 86%, so perhaps the 240cc injectors are very slightly small. I would however prefer to be on the small side for low down tuneability.
Does it really matter if you max out your injectors duty cycle if the AFR is ok on full load? Does a longer squirt time low down improve idle or shunting or just low down tuneability? i.e. more accurate AFR.I believe you should aim for a max injector duty of around 80%, this should ensure a reasonable open time at a low duty for better control. My max duty is 86%, so perhaps the 240cc injectors are very slightly small. I would however prefer to be on the small side for low down tuneability.
ric355 said:
My ECU uses a single value for close angle across the entire RPM range. It uses that together with the current crank angle and speed to work out when the injection event should start in order to ensure it finishes at the specified angle. Not sure if MS is the same in that sense but there are a couple of ways to work out the close angle given the way mine works.
The first is to use the camshaft specification to work out the angle at which the inlet begins to open, and set the angle to a few degrees before that.
I used the other method which, I guess would need to be done at multiple RPMs if you have a table to populate, is to run the engine and monitor the AFR whilst changing the close angle. Look for the point where it goes richest, as this is the point where you are getting the most fuel burned. In my case I worked backwards from 359 degrees and ended up with 295 degrees but the numbers are going to depend exactly on how MS works. I have a standard 450 cam, and looking at the specs this is pretty much bang on where it should be (I did not know the spec when I tuned it).
Thanks for explaining how you tune your injector close crank angle.The first is to use the camshaft specification to work out the angle at which the inlet begins to open, and set the angle to a few degrees before that.
I used the other method which, I guess would need to be done at multiple RPMs if you have a table to populate, is to run the engine and monitor the AFR whilst changing the close angle. Look for the point where it goes richest, as this is the point where you are getting the most fuel burned. In my case I worked backwards from 359 degrees and ended up with 295 degrees but the numbers are going to depend exactly on how MS works. I have a standard 450 cam, and looking at the specs this is pretty much bang on where it should be (I did not know the spec when I tuned it).
So you have, 2 squirts 2 cylinders at a time and you can set the injector close crank angle which means every cylinder gets half a squirt just before the inlet opens? That’s a far cry from the 14CUX that relies on the spark to trigger the injectors which varies with mechanical and vacuum advance and has no idea of the crank angle.
Gassing Station | Chimaera | Top of Page | What's New | My Stuff