ECU - replace/upgrade?

[adam quantrill]

Simon - these days people want convenience. Why not stick a bluetooth chip in the ECU too? Then with the plethora of Android phones out there, it would be relatively easy to write an app to either transfer map files to the ECU, or change parameters on the fly (without having to lug a laptop around). Then comes the interesting bit - the phone app can access the accelerometer in the phone, and you have a moving rolling road. You can match the acceleration with the ECU performance in real time and play around with the map to improve the performance, or optimise for economy if needed.

[simon_h]

adam quantrill said:

Simon - these days people want convenience. Why not stick a bluetooth chip in the ECU too? Then with the plethora of Android phones out there, it would be relatively easy to write an app to either transfer map files to the ECU, or change parameters on the fly (without having to lug a laptop around). Then comes the interesting bit - the phone app can access the accelerometer in the phone, and you have a moving rolling road. You can match the acceleration with the ECU performance in real time and play around with the map to improve the performance, or optimise for economy if needed.

Nah, just install a Megasquirt and it's already done for you apart from the accelerometer as the ones in a typical phone are not accurate enough. Granted it is quite cool to drive down the road and have a realtime display on your Android phone - http://tunerstudio.com/index.php/shadowloggerms

[RCK974X]

Mark, back to original - from the comment a while ago, if your ECU is constantly overfuelling, perhaps you could temporarily fiddle it weaker the same way, by putting a resistor in parallel with the temp sender to make ECU think the engine is hotter than it really is....probably somthing like 100-200 ohms ? perhaps a bit higher...

I have spent quite a bit of time on old Ford EEC-IV programs (as read out of the EEC boxes), the really old ones (like 1985) are fairly simple, but still require a lot of knowledge to decipher, the later ones with adaptive learning, Oxygen sensors, coil packs etc are really, really tough to understand.

The megasquirt looks good to me, from the websites and forums, but yes, it probably will take a while to get a good base setup...

[mark387mw]

I don't have the facilities nowadays to do my own repairs so just looking for a direct replacement.
If a megasquirt will do the job then that'll do, it's just finding one!

[simon_h]

mark387mw said:

I don't have the facilities nowadays to do my own repairs so just looking for a direct replacement.
If a megasquirt will do the job then that'll do, it's just finding one!

Hi Mark.
this company http://www.extraefi.co.uk/RV8_Kits.html#pnp used to produce a MS which was pretty much a drop in replacement for the 4CU but they don't advertise it any more. They've moved onto the later 14CUX now but it's worth contacting them just in case they can still supply you with something which would plug straight in with the minimum of fuss.

This thread http://www.v8forum.co.uk/forum/viewtopic.php?p=758... discusses the very issue.

Simon.

[honestjohntoo]

RCK974X said:

if your ECU is constantly overfuelling, perhaps you could temporarily fiddle it weaker the same way, by putting a resistor in parallel with the temp sender to make ECU think the engine is hotter than it really is....probably somthing like 100-200 ohms ? perhaps a bit higher...

True-ish! Inserting a 200 ohm resistor is a 'get you home' trick for a failed temp sensor, but destroys the Efi system's ability to start when cold.

10,000 ohms is a tweak introduced by Rover for SD1 post production idle speed hunting (surging) issues. Technical reasons of doing this are explained in my essay, here:

http://www.vintagemodelairplane.com/pages/Snippets...

As an 'outside the box' fix, experimenting with lesser values may do as suggested for a cure to overfuelling across the range by a rogue ECU.

However, 'inside the box' it seems to me that what owners would like nowadays, not withstanding any minor AFM variations, is a simple table of internal resistor changes to allow an ECU from one car to be used in a different car.

With apologies to those still collecting 'cheese scone', to achieve this would require a degree of altruistic collaboration between knowledgeable hobbyist in the interest of preserving the marque. Easy for me to say, being retired for more than 20 years.

[adam quantrill]

Getting back to the other sub-thread... I wonder if we know much about the original production process? We know that ECU's are matched to flapper AFM's.

So I wonder how they were set up at the time - were the AFM's put on a rig and had the carbon track adjusted first? Then were the resistors chosen for the ECU? How were the pairs then kept together before assembly in the car?

Then what happened to the batches destined for TVR - presumably these were taken off the production line as matched pairs - how were these kept together?

Any anecdotes from the factory would be interesting to hear....

[honestjohntoo]

adam quantrill said:

So I wonder how they were set up at the time - were the AFM's put on a rig and had the carbon track adjusted first? Then were the resistors chosen for the ECU? How were the pairs then kept together before assembly in the car?

It's just my opinion, but I don't hold too much much credence for the ongoing importance of the AFM/ECU matching situation, especially on post production maintenance.

I doubt, too much, when an ECU croaked early in its life, the SD1 service agent rematched the replacement ECU from his stores location, to the existing AFM - AND THEN RESEALED THE SECURITY NUTS.

In the hobbyist era, for example, I had three normally working SD1 Efi's at the same time, one twin plenum and two single plenum's, and interchanged their ECU freely without any noticeable difference in performance or characteristics. Likewise to a lesser degree I interchanged AFM's and - again - no detectable diff.

Undoubtedly, the purist will attest that any differences would be seen on a dyno or rolling road, but in the heyday of the Efi SD1 when thousands of these machines were being sold over a four year period mainly as company cars for executive types and to police and other security establishments, the general experience of latterday SD1 hobbyist/owners is that all these SD1's had the same ECU's and AFM's and are/were essentially interchangeable for normal road use - Then and now!!!

Not so for the ECU's inside range rover, land rover and other lesser marques like the TVR, MG, Morgan (?) and what have you. Even for different continents, USA, Europe, Antipodes, the resistors and other innards were changed according to local law.

And today, it would be rare to find an Efi system with all its original components, but there are many of us who have managed to find the "RIGHT" replacements (TP,CSI,ECU,AFM,FPR,Etc) and the car continues to function as intended.

For TVR's during production, with different specifications and engine capacities, no doubt matching the ECU to the "car" was de-rigour and in today's hobbyist after-market an absolute mix 'n match nightmare - hence the current bout of discussion.

[Wedg1e]

adam quantrill said:

Getting back to the other sub-thread... I wonder if we know much about the original production process? We know that ECU's are matched to flapper AFM's.

So I wonder how they were set up at the time - were the AFM's put on a rig and had the carbon track adjusted first? Then were the resistors chosen for the ECU? How were the pairs then kept together before assembly in the car?

Then what happened to the batches destined for TVR - presumably these were taken off the production line as matched pairs - how were these kept together?

Any anecdotes from the factory would be interesting to hear....

If you pull any AFM apart (including the larger-bore Jaguar unit) you'll find they used the same ceramic resistor network. Effectively it's a ladder of resistors with the carbon track paralleled across it. On no two networks that I've measured have the values of each 'step' on the ladder been the same. It looks as though each resistor started off as a certain value and was then 'cut' (and I suspect mechanically rather than by laser at it would probably be now) to create a new value - you can see the cuts in the resistive pads.
What this will do is create a number of fixed points (equating to currents or voltages input to the ECU) as the wiper sweeps, modified by the value of the carbon track as well.

[Wedg1e]

honestjohntoo said:

I doubt, too much, when an ECU croaked early in its life, the SD1 service agent rematched the replacement ECU from his stores location, to the existing AFM - AND THEN RESEALED THE SECURITY NUTS.

Why not? It was still new technology then, they wouldn't want it tampering with. The 4CU only uses two little aluminium caps, it's the work of a couple of seconds to replace them.

I work in a calibration facility, an awful lot of the instruments and devices we see are similarly anti-tamper sealed, we keep stocks of the necessary bits to maintain that integrity.

It is of course entirely feasible that for a 'cooking' 3.5L, the issue of ECU swaps is less of a problem than for 'hotter' versions as you say, and it's also likely that a large percentage of the ECUs that people buy from the likes of Ebay are not working properly anyway so swapping doesn't solve the original problem.

[Wedg1e]

honestjohntoo said:

...the general experience of latterday SD1 hobbyist/owners is that all these SD1's had the same ECU's and AFM's and are/were essentially interchangeable for normal road use - Then and now!!!

Maybe we should ask for a 'hands-up' of all those who've bought ex-SD1 ECUs and/or AFMs off Ebay and had absolutely no issues with them on their TVR 350i.

Then a show from all those who have had grief...

[simon_h]

Wedg1e said:

honestjohntoo said:

...the general experience of latterday SD1 hobbyist/owners is that all these SD1's had the same ECU's and AFM's and are/were essentially interchangeable for normal road use - Then and now!!!

Similar experience here. 3.5 VDP EFI, Vitesse Single, vitesse twin plenum, standard 1986 350i. I have interchanged AFM's and ECU between them no problem - They were all manual gearboxes though.

Simon

Maybe we should ask for a 'hands-up' of all those who've bought ex-SD1 ECUs and/or AFMs off Ebay and had absolutely no issues with them on their TVR 350i.

Then a show from all those who have had grief...

I suspect that a lot of that grief has been caused by faulty ECU's. I accept that the RR engines have a different compression ratio, cam etc, and the ECU has been tweaked slightly to satisfy those requirements.

If you wanted an AFM in 1986 for an SD1 you just ordered one and stuck it on. There was no additional selection process and Rover were happy enough to supply you with one. The engine in my 350i was just a bogo Vitesse unit and kicked out the same power on the rollers as my single plenum Vitesse so why would TVR have felt it necessary to tweak anything for no or little gain?


I tested an RR ECU i got hold of in the stock 350i before I tore it apart (the ECU that is) and it ran fine. Granted I didn't drive it for long so there may have been some slight differences in characteristics that went undetected.

Simon.

[Wedg1e]

simon_h said:

I suspect that a lot of that grief has been caused by faulty ECU's. I accept that the RR engines have a different compression ratio, cam etc, and the ECU has been tweaked slightly to satisfy those requirements.

If you wanted an AFM in 1986 for an SD1 you just ordered one and stuck it on. There was no additional selection process and Rover were happy enough to supply you with one. The engine in my 350i was just a bogo Vitesse unit and kicked out the same power on the rollers as my single plenum Vitesse so why would TVR have felt it necessary to tweak anything for no or little gain?


I tested an RR ECU i got hold of in the stock 350i before I tore it apart (the ECU that is) and it ran fine. Granted I didn't drive it for long so there may have been some slight differences in characteristics that went undetected.

Simon.

TVR didn't tweak anything at any time. The 350i was box-stock SD1 and the hotter engines were outsourced. Don't forget that although the different versions of the 4CU carry different part numbers, there's no guarantee that Lucas didn't alter the calibration for each model over the years of production which itself could explain why different AFMs and ECUs don't work well together. I'm not disputing that you and others have swapped parts around merrily with no apparent ill-efects, what I am telling you is that from my research (and that of others) there appear to be no two AFMs, and no two ECUs with PRECISELY the same set of resistance values. Something was being set up to do something with something else. The only wild-card suggestion that anyne has come up with is that the Ferranti chipset was so variable that they had to mess about with the discrete components to make them work properly.
If it was so simple to crack this, then why hasn't it been done already or if it has, why isn't there a widely-known source of 4CUs pre-tuned to the skies? The internet is full of Rover V8-powered vehicle forums, someone would know somebody. Still, give it your best shot, it could be you

[simon_h]

Wedg1e said:

The only wild-card suggestion that anyne has come up with is that the Ferranti chipset was so variable that they had to mess about with the discrete components to make them work properly.

Funny that! I was pondering this on the way home from work and was going to put forward the same hypothesis. That is that the chipset manufacturing process may not have been very consistent back then so not only were the calibration resistors selected to suit a particular application, they were also trimming out the variance between individual chipsets. If this is the case then there is no way to simply have a resistor lookup chart mapped to a particular application.
I guess the only way to prove this would be to run an ECU up in a simulated environment and capture the response, swap out the chipset, repeat the test, and see if there are any major deltas. If not repeat again until this theory is proved or disproved. A lot of effort!
The AFM is still bothering me as well. Between standard 3.5 v8's I can see no reason for them to vary. However, if you were designing the system you'd want the AFM to be at nearly full scale when the engine was making full power so that you had the best resolution. If you suddenly bolted the same AFM onto a 3.9 or 4.0 it would go FSD way before full power and not give a true representation of air flow to the ECU. That's where tightening the spring would come in isn't it? In that case an AFM off one car may not present the same resistance compared to another at the same given airflow. I struggle to believe that they actually trimmed the AFM resistor matrix on a per application basis as all you are trying to do is measure the volume of air entering the engine, not distort the fueling curve as that's the job of the ECU? BUT, to compensate for that change you'd have to adjust the ECU accordingly because at airflow X the AFM would be outputting a lower reading however the engine would still need the same fueling. You'd then have to fiddle the ECU to match the AFM which is what you have been saying all along.

Some people have said that the AFM is fully open before full power is reached, and that the ECU uses the basic RPM / Load map in the Ferranti chipset to lookup the fueling for the rest of the curve. I don't know if this is true.?

[RCK974X]

Reading with interest - a couple of comments....

When I said a resistor in parallel, and the response was it may not start - well it's called a SWITCH wired in series with the extra resistor....could even do it from the dash !

Ferranti chips - I did work for the big F in my youth - I don't think their chips would be all that variable, as they had a high standard of manufacture (and did a lot of military stuff too). Don't know if these chips were were outsourced though....

Flapper meters - from doing quite a bit around Ford EEC-IV, there does seem to be a range of calibrations, even for same meter !! (I think this was Ford doing fine tuning though). I understand there used to be an official spec for each meter by part number (air flow vs voltage). Did the Rovers have Bosch units (or Bosch copies ?). I guess there MUST be a spec somewhere. The Bosch units are ADJUSTABLE inside, although sealed at manufacture (you can get them open though), design seems to support a calibration on flow bench as part of manufacture ...

Mark, megasquirt is orderable from web, although there's quite a few options to work through (because they cover a wide range of engines, manufacturers etc). There's even a build it yourself option !

Edited by RCK974X on Thursday 19th July 04:55

[adam quantrill]

Wedg1e said:

honestjohntoo said:

I doubt, too much, when an ECU croaked early in its life, the SD1 service agent rematched the replacement ECU from his stores location, to the existing AFM - AND THEN RESEALED THE SECURITY NUTS.

Why not? It was still new technology then, they wouldn't want it tampering with. The 4CU only uses two little aluminium caps, it's the work of a couple of seconds to replace them.

I work in a calibration facility, an awful lot of the instruments and devices we see are similarly anti-tamper sealed, we keep stocks of the necessary bits to maintain that integrity.

It is of course entirely feasible that for a 'cooking' 3.5L, the issue of ECU swaps is less of a problem than for 'hotter' versions as you say, and it's also likely that a large percentage of the ECUs that people buy from the likes of Ebay are not working properly anyway so swapping doesn't solve the original problem.

I thing you're both right. HJ2 has the experience of inter-series swapping. The "calibration" on a series (in this case SDi) AFM would be to make it as standard as possible. The ECU's wuld then be plug'n'play as the resistors would also be chosen to 'normalise' the response. This makes production line procedure much simpler and also simplifies the parts supply chain. To some extent the SDi Vitesse-derived 4CU's should then be a "straight swap" into a 350i.

HOWEVER there are many different applications out there now. E.g. Range Rover with bigger engines, torque requirements etc. LDV vans. Morgans. Each application could/might have some adjustments away from the SDi curves.

So this is the situation we are in now - lots of different ECU's on fleabay or derived from scrapyards where the donor vehicle has not been noted now mean we have to be more careful - plug'n'play won't work any more.

[Wedg1e]

adam quantrill said:

So this is the situation we are in now - lots of different ECU's on fleabay or derived from scrapyards where the donor vehicle has not been noted now mean we have to be more careful - plug'n'play won't work any more.

I think there are a couple of scenarios at work though: first is the obtaining of a spare ECU to eliminate it as a cause of running issues that apparently can't be otherwise traced using available diagnostics; second is 'converting' say a RangeRover ECU to SD1/350i spec. and thirdly is 'fine-tuning' or otherwise optimising an existing working ECU to better suit the state of tune of the engine it serves.
We know from the type numbers which ECU was Rangey, SD1 etc. so we can save ourselves some work in the case of the first scenario by simply not buying the wrong one (as long as the label hasn't been removed, which they often have).
Scenario two only really comes into play when/if the only ECUs available are the 'wrong' type for scenario one... which leaves the third and most likely (or most potentially rewarding!) - that of potentially doing Mr. Adams et al out of business. In theory this is where the big spreadsheet of resistor values come in... except it doesn't because there's still no definitive description of what the various 'tunable' areas of the circuitry are doing in the fuelling 'mix'.
It would be easy to suggest making R24 1000 Ohms higher or lower to give such-and-such an increase or decrease at whatever part of the rev. range IF we knew what R24 was contributing.
I did get as far as designing a board with numerous adjustable potentiometers on it that could be plugged onto the standoffs in place of the existing fixed resistors and whilst this still has some merit, it overlooks the fact that the standoff resistors are themselves paralleled across other resistors mounted normally on the PCB that ALSO change with each application (so the 'trimmer' values would have to vary for each ECU series at least, if not within each series).
A further issue is the fact that if the final resistance value is so critical it has to be set using 2% tolerance components, you couldn't realistically hope to hit that precise a value by twiddling a pot. at random - probably whilst having your fillings jarred loose in the typical TVR ride.
Also, whilst that degree of adjustability is fine in a piece of electronics kit that you have on the bench with an array of test instruments hanging off it, where you can immediately see the effect of introducing more voltage or decreasing some current, it becomes a different matter when it's in a car and the only feedback you have is the colour of the exhaust smoke and the feel of your arse, as it were. Yes, there's the rolling road, but how many of us have the funds to book one for the undoubted hours it'd take to see perhaps a piddling improvement?

[honestjohntoo]

adam quantrill said:

HJ2 has the experience of inter-series swapping. The "calibration" on a series (in this case SDi) AFM would be to make it as standard as possible. The ECU's wuld then be plug'n'play as the resistors would also be chosen to 'normalise' the response. This makes production line procedure much simpler and also simplifies the parts supply chain. To some extent the SDi Vitesse-derived 4CU's should then be a "straight swap" into a 350i.

HOWEVER there are many different applications out there now. E.g. Range Rover with bigger engines, torque requirements etc. LDV vans. Morgans. Each application could/might have some adjustments away from the SDi curves.

Seven years ago I conducted an inter-series swapping experiment with the following purpose, results and conclusions.

Purpose

A question seen quite frequently on SD1 forums is whether a Range Rover ECU will work OK in a regular flapper type SD1 Efi system. In order to try and get a resolution I purchased several Flapper ECU's at quite low cost on Ebay and tested them under near identical road conditions in my TP.

Common status

My car's resident ECU is the Mark Adams Tornado unit (since 1994), so that was my benchmark.

Most tests were conducted in 4th gear with Cruise Control “On” over the exact same test track (Yes! Cruise is retro-fitted on my TP)

The results were taken and recorded from my working trip computer set and reset at the start and end of the test track.

The test track was the newly opened Bedford Bypass, an 8.7 mile section between bridges, making for identical 'test start' and 'test stop' locations, once the car was brought up to speed and the test terminated before clicking out of cruise control.

Because the cruising speed was just below 50 mph I chose 4th gear so that the engine rpm was operating at or near maximum efficiency. Also chosen so as not to be impeded by slower moving traffic.

Results



Comment on Results

I did not run the car flat out (obviously) so it was not possible to establish if the upper speed performance was limited in any way but in all practical sense, I judged the upper speed is pretty much going to be a function of whether the top rpm is affected so I redlined the engine in a lower gear for each of the five ECU's and could not detect any high end limitation or acceleration differences from my resident 'Tornado' ECU.

For an accurate assessment I suppose one would need to do the tests on a rolling road - way outside a retiree's budget?

I did not have any qualms about fitting a "wrong" ECU in the Vitesse as I had already established (by asking around) that they were practically identical in function, with relatively minor changes in tuning and performance.

I did not try to adjust out the idle speed problems during the experimental process as the car was running fine with its resident ECU and the propensity for messing things up was high.

Also - Unless the resident ECU gives up the ghost the experiment remains currently "on hold". Even after seven years.

Conclusions

The result tables show that the car actually cruised OK on all the Range Rover ECU's but had idling problems on two of them. The RR numbers refer to the ECU Model/Part No. The other Vitesse ECU behaved normally.

My provisional conclusions were that the idle speed problems on the Range Rover units might possibly be overcome by adjusting the Efi system variables such as idle speed and mixture control, but probably not the throttle pot settings which generally only come into play for acceleration and full load enrichment.

The other system variable would be the throttle plate residual idle gap of about 2 thousandths of an inch normally seen on Range Rover units with their built-in throttle stop screws, but not normally seen on the single or twin plenum Vitesses.

Not-with-standing the above, the three Range Rover ECU's had a very similar cruise performance at the chosen setting and in itself bodes well for use in an Efi SD1 with the possibilty of corrective external tweaking to get some control over the idle speed.

Out of interest My ECU is rigged for quick change-over as its cradle is removed from the car and the ECU resides (otherwise well protected) in its usual location.