Lucas EPIC DSA1191 / DSA1096 ECU Hacking on a '99 Smiley

[marcrbarker]

I've just got a smiley-front LWB hi-top to play around with, I think it's a MK5 or maybe MK4, someone can maybe correct me on that. So I've decided to delve right in to the engine management system and also do some serious hacking too with the drive-by-wire system (which I've had great fun with before in the past). If I cock things up it's no big deal as the van is almost falling apart now. What I'll do is publish here what I've been up to and I'm sure whatever I come up with will be useful.

A bit of history on the van: LHD, registered in and lives in Spain. Last owner really pleased with it for 5 years, reliable, quick, economical and great for doing home moves. What went wrong is slight leak injector pump and then different mechanics breaking things fixing it. Then all the leaks went crazy after timing belt changed (probably because the pump shaft seal gets pulled about in the process). At this point things went from bad to worse, a mechanic ended up with the pump in myriad of pieces spread all over a dirty bench for nearly a year and then that was the end of the original DSA1191 pump. So I got a DSA1096 off a scrapped MK(3?) and the mechanic was very keen to fit it - so he could charge for 'fixing it'. The owner then got the van back and it barely went faster than 40 km/hr , the check engine light on and then shortly after that pump started leaking just as bad as the original. This is when the van came to me to play around with. I think it's a shame to scrap the van and I decided it would be fun to get it back just as it was before mechanics got hold of it, and going better than before. Also there's friendly rivalry between the last one and me.

So this is what I've got up to so far: DSA1096 still in place, no leaks at all now, I've done the seals that were leaking from an original kit covered in dust I'd found locally. The engine still runs on 10 s after the key is removed and the check engine light still on. I've just looked at the ESOS shut off valve plunger and there's no rubber tip at all all on the end, which seems very odd. It's just bare metal with a ridge for retaining a rubber. In the kit of seals there's no replacement for it which is also odd. I'm going to put a end on it anyway, because I can't see how it would work otherwise. So this leads me to my first question:

1. When the ESOS fails to stop the engine within 10 seconds I presume this sets 'limp mode' of low power. Is this limp mode also programmed to make the van slightly 'kangaroo' along (to call attention)?

2. I don't have access to a Laser 2000. Is there a 'paperclip method' (shorting diag socket pins) for getting codes?

I think I can get by without a scanner. I can look at the 0-5V analogue signals and the PWM and directly see when things are getting out of range, and without waiting for trouble code. What I can see so far with this Lucas EPIC system is it's quite primitive. It doesn't have a needle lift sensor for injection timing. This EPIC system seems to be 'open-loop', that is all the timing is referenced to the Crankshaft sensor. This means the advance must be carefully calibrated at manufacture. I guess that might be what the Calibration Resistor might be: residual error after factory calibration. If so, that will mean just moving the pump position back and forth on the timing belt won't directly affect the timing (because injection timing is referenced to crank sensor). That will also mean that the timing stays put as the belt stretches with age, until out of accomadation range, then it'll suddenly throw a fault code more often as the belt ages.

This is what I have planned later in week: I'll mess about with the throttle and timing advance servos, by adding some custom hacking circuitry. The objective is to fine trim the static injection timing on-the-fly while driving (and listening carefully). The other servo I'll play with is the fly-by-wire throttle which I'll see how far I can tweak things before the ECU realises. I'll post here the final circuits I use.

[MinorMatt]

Firstly DSA1096 and DSA1190/1191 pumps are interchangeable

The esos plunger should have a rubber end on it, and if it doesn't the engine will run on for 10seconds before the management shuts it down another way.

As you turn the ignition off, stall the engine. This tricks the management in to thinking everything works properly and it should run at full power next time...

[marcrbarker]

Good idea about stall the engine after turn off. I'll try that tommorow. What about the limp mode? Does it also include gentle but annoying kangarooing?
That's really odd about the end of the plunger missing and also not a new one included in the seals kit. If it had disintegrated I would had expected to had found debris caught in one of the internal filters but nothing. Maybe a mechanic removed the rubber tip completely because it was "stopping the engine". They do crazy things like that here, motorcycles here often have bypassed vacuum-operated fuel valves - because the first thing they do when a no-start is check the fuel flow and they find a 'blockage' which they bypass. You often see motorcycles with pools of gasoline under them. Another thing I've seen done here is big holes drilled along cooling ducting to try and fix overheating.

I've just put an end on the plunger. I'm trying to paste an image into here but doesn't appear so I'll describe it, it's just a blob of fuelproof high temperature automotive silicone, allowed to dry then shaped. It just so happens to be red colour too. I would try buy a new plunger but I'd decided I wouldn't spend any money if possible. Tomorrow after trying the 'stall and restart test' and see it that clears limp mode then after do that a few times then I'll refit the re-ended plunger.

[marcrbarker]

By the way I'm still not quite sure which two pins to short together on the diagnostic connector to get these 'blink codes'. I presume then this Ford it's the standard pins typical for late 90s

[MinorMatt]

Believe it or not, the rubber breaks down in the diesel and disperses in to the pump - I have never seen one have issues associated with it once the plunger has been changed.

You cannot get blink codes from a 1999 vehicle - that stopped in 1994

[marcrbarker]

The 'paperclip trick' finished that long ago 1994? THought it was later. Wow. PLunger rubber end completely dissolved away with nothing trapped in a filter? That's really neat how they do that, if the bits are small enough to get past a filter I expect it'll pass through an injector just as well, probably give the injector a good clean on its way out. What I STILL find peculiar is there's no replacement plunger rubber at all in the kit of seals. They include all the outside plastic blanking caps for lazy pump rebuild shops, but the kit omits the very part that dissolves away the most.

I haven't been able to get across town to the van to try the stall test yet. I am still curious most of all to know if this gentle but annoying 'kangarooing' is a part of the limp mode. If the ESOS failure limp mode can be easily cleared by stalling (how ever many times needed to clear that DTC) then I might not even bother putting the plunger back in the pump, just leave it out completely while I mess about with the injection timing advance 'on-the-fly' I've planned later for this week. I'd rather not start any more mechanical interventions just yet until I've investigated enough first.

[MinorMatt]

The 'paperclip trick' finished that long ago 1994? THought it was later. Wow. PLunger rubber end completely dissolved away with nothing trapped in a filter? That's really neat how they do that, if the bits are small enough to get past a filter I expect it'll pass through an injector just as well, probably give the injector a good clean on its way out. What I STILL find peculiar is there's no replacement plunger rubber at all in the kit of seals. They include all the outside plastic blanking caps for lazy pump rebuild shops, but the kit omits the very part that dissolves away the most.

I haven't been able to get across town to the van to try the stall test yet. I am still curious most of all to know if this gentle but annoying 'kangarooing' is a part of the limp mode. If the ESOS failure limp mode can be easily cleared by stalling (how ever many times needed to clear that DTC) then I might not even bother putting the plunger back in the pump, just leave it out completely while I mess about with the injection timing advance 'on-the-fly' I've planned later for this week. I'd rather not start any more mechanical interventions just yet until I've investigated enough first.

The plunger comes as a single part, you don't get a separate seal as they are bonded together. Every time you switch the van off you will need to stall it and that can't be good for the clutch/mechanics on a long term basis? The ESOS plunger is also not an off/on its operated on a PWM basis to help regulate flow through the pump - so could well be the cause of the subtle surging.

[marcrbarker]

There's no plunger (rubber end missing) fitted at all and it hadn't made any difference to anything it being there or not. Clearly because there's no rubber end on it to do anything useful.

This Lucas EPIC system is incredibly primitive. The ESOS isn't clever at all, it is just ON/OFF, not variable. The ECU just delivers three fixed PWM modes: 1-circuit test/ 2-solenoid pull in/ 3-solenoid retention and that's it, about 1% / 100% / 30% fixed. The ECU can't tell the plunger's missing after it's stopped sealing, wouldn't be that clever.

That's exactly what I've planned, stall the van every time after turning off while I'm 'chipping' the management system. I don't plan doing that forever, and anyway the van could likely get scrapped before it's damaged.

[marcrbarker]

Here's an update. 1. Only need to stall engine one time after key off and the once is enough to clear the check engine light. The light staying off also means there's likely no other faults. 2. I didn't bother fitting the plunger with the new rubber end I'd put on, because I want to limit mechanical interventions to bare minimum. It gets quite dusty sand in the air in the summer (la calima). So there's still no plunger at all fitted inside the pump and I can confirm it doesn't seem to make any difference that I can tell.

Next stage is tweak the injection pump timing advance forward while driving to see if I can get my first 10 bhp of increase. That's my usual order of attack. If it goes well I'll post the circuit. As I said I'm spending as close to zero money on this I can and reuse everything original.

[MinorMatt]

Has the stalling trick got rid of the "kangarooing"?

If you advance/retard the timing more than +- 4 degrees (iirc) it will go in to limp mode. There are some small locking bolts on the front though, which I am fairly sure allow you to physically adjust the timing without altering the "pinning" position.

I havent worked out what the calibration resistor calibrates yet - but I have had reports of a noticeable, but not massive, improvement in power by fitting a number 15 resistor (100 ohm +-5) in place of whats fitted. It would be easy enough to mod the wiring, but the resistor (2 terminals with the white surround) just pulls out and replacement pushes in.

Have you found the LDV lucas EPIC manual? There is some useful info in there - if not search for "Ecovippari EPIC" on google

The light staying off also means there's likely no other faults.

the light probably stays on for 5 seconds before going out? If it goes out immediately the fault code memory is clear - if it doesn't there are either "minor" codes stored (e.g. EGR) and/or historic/resolved codes which haven't been cleared from the memory.

I don't know what diesel is like there, but EPIC pumps arn't very keen on low Sulphur fuel - the advance CAM ring in the pump needs the lubrication the sulphur provides, and then it goes sticky and causes limp. Its recommended to add extra lubrication (normally mineral 2 stroke oil) to the diesel to prevent this.

I'm interested to see where you go with this

[marcrbarker]

Has the stalling trick got rid of the "kangarooing"?

At last I've been able to 'hacer una vueltita' on the open road to put my foot down enough to test it. It's still very gently kangarooing as before, although less noticeable now because the normal power has come back that masks the effect- it being out of limp mode. It's a continual regular oscillation in engine torque, always at 2.5 Hz, so I'm pretty sure now it's an artefact of the engine management system. Suspect it's resonance in the injection quantity servo loop triggered by a weak cylinder. If it is then I expect lots of other Transits will do this too

If you advance/retard the timing more than +- 4 degrees (iirc) it will go in to limp mode.

What I've got planned the ECU won't know or realise I'm changing the timing at all, nor do I want it to. Having said that, while I shift the timing about the ECU will be outputting a different PWM % to the cam ring advance because it will think it's accommodating for wear or tolerance. If a code does get triggered then it'd be for 'unfeasable PWM' after I've accidently dialed up 20 deg of advance.

There are some small locking bolts on the front though, which I am fairly sure allow you to physically adjust the timing without altering the "pinning" position.

The pump drive pulley bolts between them the pump timing pin hole? I might be wrong on this but I wouldn't expect offsetting here would have any direct net effect on timing at all, otherwise people would use it to fiddle with the timing. Instead I expect moving pulley position like this would just alter the PWM operating range the ECU needs to command to get the programmed cam ring placement coming back on the 0-5V analogue signal. Rather like AFC on a FM radio, you move the tuning dial slightly back and forth and nothing much happens. Anyway, my objective is change timing on-the-fly with a turny-knob thing in front of me while climbing a hill. Do it electronically and keep all mechanical interventions to absolute minimum.

I havent worked out what the calibration resistor calibrates yet - but I have had reports of a noticeable, but not massive, improvement in power by fitting a number 15 resistor (100 ohm +-5) in place of whats fitted. .

I strongly suspect the calibration resistor is just a sticker they put on afterwards to represent how much residual error is left after manufacture and testing, to tell the ECU how much to trim off or add to static timing. It's that primitive. Bit like shims on bearing. I would expect the intervention caused by changing that square o-ring would technically demand the resistor is redone. I suspect lazy pump rebuild shops wouldn't understand or be bothered with the resistor so improvements changing it for 100 ohm would be sheer pot luck. I would expect the ECU only ever reads the resistor the once, during initialisation after a power-on, very unlikely the ECU would poll it for change while running.

Have you found the LDV lucas EPIC manual? There is some useful info in there - if not search for "Ecovippari EPIC" on google

Yes found the LDV info. Also saw Ford Training etc. that's claiming to be hosted for download in lots of fake discussion forums and fake filesharing sites, obviously it's for harvesting credit card numbers from people searching on google. I didn't bother with those.

on for 5 seconds before going out? If it goes out immediately the fault code memory is clear - if it doesn't there are either "minor" codes stored (e.g. EGR) and/or historic/resolved codes which haven't been cleared from the memory.

The light goes out 5 seconds after start. The EGR was blanked off long time ago and the mechanism still moves ok. I'm pretty sure the ECU can't tell it's been blanked off (by looking for a MAF dip) because there's no MAF I can see. I suspect the stored codes causing the '5 seconds' will clear by themselves after long enough without events, unless I forget to stall the engine and spoil the history. If I keep doing that I'll need to refit the stop plunger.

I don't know what diesel is like there, but EPIC pumps arn't very keen on low Sulphur fuel - the advance CAM ring in the pump needs the lubrication the sulphur provides, and then it goes sticky and causes limp. Its recommended to add extra lubrication (normally mineral 2 stroke oil) to the diesel to prevent this.

I'm interested to see where you go with this

If the cam ring sticks at full advance it'll be easy to hear as the engine will clack loudly at low RPM and the red light comes on. I can also look out for any sticking cam ring easily by looking for step changes on the 0-5 V analogue feedback signal.

I think it's the exact same "gasoil" (diesel) that's on mainland Spain. It's very unlikely anyone here would had habitually added twostroke oil, the lifestyle here is free and easy. There's an epidemic here of rusted out core plugs and bypassed failed cabin heaters because many people never heard of antifreeze, though having said that they are marketing it more nowdays as 'refrigerante para verano' = summer coolant. I'll ask around and see how well known it is.

Next instalment will be altering timing on-the-fly from the driving seat while hill climbing, turning a control back and forth .

[marcrbarker]

Now I've figured out how to post images, here's the circuits I'm using and what they do.

The first thing of all is a diagnostic aid because I don't have a compaitible scantool, it's little piece of cardboard with two LEDs attached to pump wires with safety pins and taped over.

D1.png

This has been so handy I've left it attached to the bulkhead behind the pump. It shows two lights PLU and POS. The engine stop plunger signal, which was discussed earlier, and the other LED the rotor inside the pump going past the 'speed sensor' under the square o-ring.

With the POS you can easily see the rotor's "3 long & 1 short" pulses as you turn the crank around by hand. I've marked them on the pump pulley with marker pen.

manchas.jpg

Don't take what you see here as gospel because here they're in the wrong place by 90 degrees clockwise. I've mistakenly put the pump back together 90 degrees out internally, only realised after using up an entire can of brake cleaner down the air intake puzzled why wouldn't start. It's an easy fix for time being, I just relocated the pump pulley anticlockwise by 6 teeth and the engine started right away.

Here's a functional block view of the injection pump.

C.png

The ESOS ("PLU" LED) is on pins 2 & 3 highlighted numbers. The bottom right corner is the rotor ("POS" LED).

The timing hacking takes place with the bit on the bottom left corner. Pin 19 highlighted. Details to follow

[marcrbarker]

Here's the "Timing-on-the-Fly" modification. It's a bit of cardboard, on front has a knob to tweak while driving.
There's a fairly long pair of wires to reach round between the back of the pump and the cab. Either the driver or passenger operates it.
It's some components and battery holder tacked onto the back of the card with hotmelt glue. At the injection pump connector a wire needs to be cut, the white/blue one on pin 19, and joined with a screw connector block. .

D2.png

What it does is adds a small DC offset to the analogue feedback signal that represents the Cam Ring absolute position. The ECU will automatically command a modified PWM signal to compensate for the offset, thinking the mechanism is wearing out. Result is a user-controlled variation in timing. It's a bit like mechanically adjusting the cam sensor, but on-the-fly.

trim.png

mod1.png

[MinorMatt]

All very interesting stuff!

It's a continual regular oscillation in engine torque, always at 2.5 Hz, so I'm pretty sure now it's an artefact of the engine management system. Suspect it's resonance in the injection quantity servo loop triggered by a weak cylinder. If it is then I expect lots of other Transits will do this too

Its not something I have ever noticed, or had reported to me...




MinorMatt wrote:
If you advance/retard the timing more than +- 4 degrees (iirc) it will go in to limp mode.

What I've got planned the ECU won't know or realise I'm changing the timing at all, nor do I want it to. Having said that, while I shift the timing about the ECU will be outputting a different PWM % to the cam ring advance because it will think it's accommodating for wear or tolerance. If a code does get triggered then it'd be for 'unfeasable PWM' after I've accidently dialed up 20 deg of advance.

The pump drive pulley bolts between them the pump timing pin hole? I might be wrong on this but I wouldn't expect offsetting here would have any direct net effect on timing at all, otherwise people would use it to fiddle with the timing. Instead I expect moving pulley position like this would just alter the PWM operating range the ECU needs to command to get the programmed cam ring placement coming back on the 0-5V analogue signal. Rather like AFC on a FM radio, you move the tuning dial slightly back and forth and nothing much happens. Anyway, my objective is change timing on-the-fly with a turny-knob thing in front of me while climbing a hill. Do it electronically and keep all mechanical interventions to absolute minimum.

No, this will alter where the pump pins... its 3 or 4 very small (m4 maybe?) cap headed screws - but as you say irrelevant to what you are trying to do.

I strongly suspect the calibration resistor is just a sticker they put on afterwards to represent how much residual error is left after manufacture and testing, to tell the ECU how much to trim off or add to static timing. It's that primitive.

It will be to take up a manufacturing tolerance, the rotor position sensor (position of the rotor dictates fuel quantity injected) is shimmed, but I think thats the only sensor that is. The Hall sensor has adjustment by slots for the screws

I would expect the ECU only ever reads the resistor the once, during initialisation after a power-on, very unlikely the ECU would poll it for change while running.

That is a correct assumption

MinorMatt wrote:
Have you found the LDV lucas EPIC manual? There is some useful info in there - if not search for "Ecovippari EPIC" on google

Yes found the LDV info. Also saw Ford Training etc. that's claiming to be hosted for download in lots of fake discussion forums and fake filesharing sites, obviously it's for harvesting credit card numbers from people searching on google. I didn't bother with those.

I have got some Lucas training manuals on the EPIC - I will see if I can find them - there may be some useful info.

The light goes out 5 seconds after start. The EGR was blanked off long time ago and the mechanism still moves ok. I'm pretty sure the ECU can't tell it's been blanked off (by looking for a MAF dip) because there's no MAF I can see. I suspect the stored codes causing the '5 seconds' will clear by themselves after long enough without events, unless I forget to stall the engine and spoil the history. If I keep doing that I'll need to refit the stop plunger.

There is no MAF on these, and you are correct ECU can't tell its blanked.

MinorMatt wrote:
I don't know what diesel is like there, but EPIC pumps arn't very keen on low Sulphur fuel - the advance CAM ring in the pump needs the lubrication the sulphur provides, and then it goes sticky and causes limp. Its recommended to add extra lubrication (normally mineral 2 stroke oil) to the diesel to prevent this.

I think it's the exact same "gasoil" (diesel) that's on mainland Spain. It's very unlikely anyone here would had habitually added twostroke oil, the lifestyle here is free and easy. There's an epidemic here of rusted out core plugs and bypassed failed cabin heaters because many people never heard of antifreeze, though having said that they are marketing it more nowdays as 'refrigerante para verano' = summer coolant. I'll ask around and see how well known it is.

I was more advising to check the diesel you have (its called city diesel here) and add some extra lube if you need too.

What you are looking at is pretty interesting - I agree the system is fairly simple (bear in mind it was introduced in 1991 as the first electronically controlled LCV diesel engine and only had minor tweaks afterwards).

It will be interesting to see if it all works, its taken a few minutes to get my head around what you are proposing - im not entirely convinced it will work, but there is only one way to find out! I have obtained extra power on one of my vans by shimming the rotor position sensor rod - but that increases the revs at idle and if you take it too far it goes in to limp.

The other thing I think would be beneficial is to modify the boost sensor so that it is tricked in to injecting more diesel by reading a higher boost lower down the rev range - putting an offset in a bit like you are doing with the timing.

If you want to mess around inside the ECU - I will happily send you one to experiment with?

[marcrbarker]

MinorMatt wrote:
its 3 or 4 very small (m4 maybe?) cap headed screws - but as you say irrelevant to what you are trying to do.

I'm not sure what this is. I realize now I'm confusing slightly between the cam advance sensor and the rotor speed sensor (that's under the square o ring) . I expect the former is concerned to timing. The latter disregard what I mentioned earlier about this square o ring sensor & calibration resistor. As I understand this latter rotor sensor has to be more than one belt tooth off sync to trigger a DTC and isn't involved.

MinorMatt wrote:
It will be to take up a manufacturing tolerance, the rotor position sensor (position of the rotor dictates fuel quantity injected) is shimmed, but I think thats the only sensor that is. The Hall sensor has adjustment by slots for the screws

I think you're referring to the rotor axial position if it's fuel quantity. I would expect the shims are to do with PWM operating range.
The Hall sensor adjustable by slots I guess is what I've called the square o ring sensor.

MinorMatt wrote:
I have got some Lucas training manuals on the EPIC - I will see if I can find them - there may be some useful info.

I think would be a good idea.

MinorMatt wrote:
Its recommended to add extra lubrication (normally mineral 2 stroke oil) to the diesel<snip>

I was more advising to check the diesel you have (its called city diesel here) and add some extra lube if you need too.

I will investigate more on this and come back

MinorMatt wrote
It will be interesting to see if it all works, its taken a few minutes to get my head around what you are proposing - im not entirely convinced it will work, but there is only one way to find out! I have obtained extra power on one of my vans by shimming the rotor position sensor rod - but that increases the revs at idle and if you take it too far it goes in to limp.

I'll try it tomorrow morning. The timing tweaks. I'm sure it'll work.
What you've done with the shims on the axial position I would expect the ECU would compensate against it the whole time because it would think the mechanism is wearing out, until it runs out of PWM action and you'll get a very unstable fast idle.

MinorMatt wrote
The other thing I think would be beneficial is to modify the boost sensor so that it is tricked in to injecting more diesel by reading a higher boost lower down the rev range - putting an offset in a bit like you are doing with the timing.

If you want to mess around inside the ECU - I will happily send you one to experiment with?

That's a task for later on. The order I'm doing things in is 1. Timing tweaks. 2. Fuel amount increase. 3. Turbo boost. All without the ECU noticing what's going on. The last one could break a con rod and then it's game over.
Thanks for the offer of ECU, I don't plan on touching the ECU itself but that might change.