Jump to content
The mkiv Supra Owners Club

Remapped ECU's


Phil Wall
 Share

Recommended Posts

 

 

In a word... no.

 

All you can do is add-on cludges that fool the stock ECU into thinking it is doing something or other that it's not.

 

Which can often be a very iffy approach.

 

For example, Justin came around today and we were playing with his EVO. One of the things we did was to fit an HKS fuel-cut defencer.

 

Well, it stopped the car hitting fuel-cut alright. But in doing so it had the effect of leaning out the mixture.

 

Stock, we were seeing fuel/air ratios of between 11.8 and 12.2 under full power, wide-open throttle conditions: which is about spot-on for a high-power turbo engine. And is what we expected from the stock motor, that is in very good condition.  

 

However, once the HKS unit was connected and set so as to avoid fuel-cut, we were seeing fuel/air ratios of 15.4 under the same conditions, i.e. full-power wide open throttle. Which is flaming crazy as 15.4 is a lean mixture (centre of the two conditions between lean and rich is 14.7) and you never, ever run ANY engine (never mind a high power turbo motor) lean at full-power wide-open throttle.    

 

The reason why the engine was running lean, was because the HKS FCD divides the air-flow meter output signal by a certain percentage. In other words, it fools the ECU into thinking it is flowing less air than it really is. Which causes the ECU to cut back on fuel thus leaning out the mixture.

 

What is really required is a fully adjustable ECU. But failing that, on fitting the HKS FCD, you must add another cludge that will fool the ECU into giving extra fuel by the same percentage that the HKS unit is dividing the air-flow signal.

 

This will have the effect of defeating fuel-cut, and the mixture ratios will remain within sane levels at WOT.

 

Yours,

J  

 

Link to comment
Share on other sites

I agree with the comment on these kludges and thats why I was wary of doing this on the Supra. I went for the Blitz ECU ( a complete replacement ECU), remapped for optimum performance albeit on 102 RON fuel, no speed or fuel cut. Mapping appears to suit my set up perfectly. Maybe I am wrong but I just see these add on bits and pieces as something else to go wrong and I avoid them if I can

Link to comment
Share on other sites

I agree with the comment on these kludges and thats why I was wary of doing this on the Supra. I went for the Blitz ECU ( a complete replacement ECU), remapped for optimum performance albeit on 102 RON fuel, no speed or fuel cut. Mapping appears to suit my set up perfectly. Maybe I am wrong but I just see these add on bits and pieces as something else to go wrong and I avoid them if I can

Link to comment
Share on other sites

Somebody else told me about the HKS FCD dividing the signal, but we werent too sure if it really caused any sort of problems with the Supra fueling. As usual the guys in the states run the HKS FCD even on the 10s drag cars so it obviously works.

 

Pete Betts took the approach of 'holding' the signal just below fuel cut and then feeding that signal back to the ECU even when going above 14psi which seems like a sensible approach to me.

 

I wonder if PBs fuel cut unit works on an Evo.

 

JB

Link to comment
Share on other sites

From mkiv.com on the subject of HKS fuel-cut vs.Greddy (Greddy is similar to Pete Betts FCD) :

 

"The reason that the MkIV owners recommend the GReddy BCC over the HKS Fuel Cut Defensor (FCD)and the Free Fuel Cut Defensor (FFCD) is simple.  The HKS unit reduces the signal coming from the turbo pressure sensor at a fixed percentage, which means that all the data coming from this sensor is corrupt.  The FFCD fixes the value coming from the turbo pressure sensor at something close to atmospheric by capping the pressure input to the sensor."

 

 

 

Link to comment
Share on other sites

I'm not quite sure what you are thinking of doing, Phil. Depending on what you are attempting to achieve I might have a solution for you in what Justin and I are currently looking into.

 

Generally, I just want to make clear whilst I agree an ECU swap is the very best solution, I believe add-on cludges can work very well up to a point. Though I shake my head in disbelief at circumstances, like yesterday, with the HKS FCD we fitted to JF's EVO in the way it defeated fuel-cut... at the expense of sending the fuel/air ratio well lean at WOT (Wide Open Throttle).

 

Provided these add-on boxes are fitted with due regard paid to any other engine system they may be affecting, and they are installed and programmed *properly*. Then, within limits, they can provide excellent fuelling and/or ignition solutions.

 

One of the ways we are going to try this, on Justin's car, is to fit the Apexi AFC in conjunction with the UK-spec injectors.

 

I'm not recommending this as a valid thing to do, yet, as I haven't actually done this before. It's just that JF is eager to do it, and he has already bought the bits, so I'll give it a go. Naturally, I'll post my findings to the BBS.

 

You never know it *may* turn out to be a good way of cludging the stock ECU that can be recommended.  

 

The theory of the whole thing seems quite sound. Basically, the UK injectors flow about 25% more fuel for the same pressure. So by fitting the UK injectors to the J-spec car means you'll run 25% richer with the ECU running in open loop mode, i.e. when driving at anything other than steady-state cruise, basically.

 

Now this is where the AFC comes in: obviously, you don't want to be running 25% richer all the time. The AFC allows you to correct this by varying the voltage output from the MAP (Manifold Absolute Pressure) sensor in programmable steps. What happens is, instead of the MAP output voltage being fed directly into the engine ECU, it is first fed to the AFC. The AFC is programmed to reduce this voltage slightly, and then it is fed to the engine ECU. Now, because the engine ECU receives a lower voltage, it cuts back on fuel.  

 

Therefore, if you were to program the AFC to subtract 25% off the MAP sensor signal, across the whole rev-range, you'd be right back where you started fuelling-wise (for the sake of clarity I'm ignoring closed-loop mode).

 

Okay, we are right back where we started, but the main difference is that now we have created a 25% extra fuelling headroom.

 

You see, the MKIV is fitted with a 2 bar MAP (Manifold Absolute Pressure) sensor. What this sensor does is to produce an output voltage between zero and +5 volts which is proportional to the pressure in the intake manifold.    

 

Big problem is, anything over 1 bar boost pressure cannot, therefore, be detected by the ECU.

 

That is not to say if you turn the boost up greater than 1 bar, your motor will run lean. With a stock motor it will tend to run richer. This is due to the fact that the turbos are taken way out of their maximum efficiency island; which has the effect of heating the intake charge to a much higher degree, thus making the air - proportionately - much less dense. Which is a bit of a saviour really.

 

Problems start to arise when slightly bigger turbos are fitted coupled with a large front-mount intercooler. Here, the density of the intake charge will be much greater at a boost pressure of 20 psig (say) than if you were running that same boost-pressure with the stock setup. In other words, you lose the "saviour" of having the turbos taken way outside their maximum efficiency island.

 

With a more efficient setup, what would happen is: as boost pressure rose above 1 Bar, the ECU would continue to provide the same amount of fuel, as it was providing, when the boost pressure hit 1 Bar. Simply because 1 Bar is the maximum range of the MAP sensor. So the engine would begin to run dangerously lean from 1 Bar upwards.

 

In using the AFC setup, we create a 25% fuelling headroom. This allows the engine to be mapped accurately at boost pressures that would ordinarily be beyond the range of the stock MAP sensor.

 

The primary effect of the AFC is to divide the MAP voltage signal that is normally sent directly to the ECU. At 1 Bar boost, this pressure-signal will be sitting at its maximum output of 5 volts. But with the AFC setup, we are making a 25% correction. In which case, the voltage the ECU actually sees at 1 Bar will be 5 volts less 25%. Which is 3.75 volts. So, as far as the ECU is concerned, the engine is fuelling for a boost pressure of (say) 10 psig with another 5 psig to go. Normally, that would mean the engine would run mega-lean, but the larger injectors compensate and the fuelling runs normally.  

 

Well, that's the theory.

 

My instinct tells me the fuel pump will probably also need replacing with the UK-spec pump. And there's a few other bits and bobs that make me go... hmm. But, all in all, it could turn out to be a valid way of going about things. Especially in cases where not a large amount of extra fuelling is needed.

 

Yours,

J

 

PS

 

Please note: all figures are theoretical and are given to demonstrate the validity of the process. Actual figures will be determined by road-testing, under various conditions, using a portable and highly accurate UEGO sensor to monitor the fuel-air mixture; together with portable high-speed data-logging of the UEGO sensor and the logging of other factors such as boost-pressure; RPM; acceleration, etc.

 

 

 

 

(Edited by Ash at 1:24 pm on July 1, 2001)

Link to comment
Share on other sites

In the long term I am interested improving the sub 100 mph performance.

 

I am not too concerned with  top end mph as it is very difficult to find anywhere camera free to use it!  But ultra fast  0-70 comes in handy all the time.

 

I have read your last post  three times and I think it makes  more sense each time.

 

I can understand the excitement of designing  and planning these kind of modifications, if you have the technical expertise to do so.  I am OK with spanners and tuning NA engines. I'll leave this stuff to you!

 

I think many of us will be waiting to hear the outcome of the tests on Justins car

Link to comment
Share on other sites

 

 

The HKS FCD situation in relation to the MKIV is something of a moot point. So I'm treading carefully here, and will comment only on the facts of the matter.

 

Personally, I removed the HKS FCD from my MKIV as it was causing my engine, under certain conditions, to run with a fuel/air ratio that, by any natural and normal engineering yardstick, would be judged too lean for a high-power turbo motor.

 

I replaced it, therefore, with a unit that allowed the pressure-signal to pass to the ECU unaltered, which then clamped the signal just before the point where fuel-cut would take place.

 

I know about this because, some time ago, I bought a portable and highly-accurate UEGO sensor setup.

 

The sensor attaches to the rear tail-pipe and there's a wire that feeds through the window to a hand-held digital display. Power is sourced from the fag-lighter socket. You can switch the display to read Lambda, or fuel/air percentage for both petrol and alchohol fuels.

 

In other words, it is not one of these 50-quid jobs you can buy out of Max Power, connect to the stock oxygen sensor, and read off from a line of LED's running from rich to lean. It is a very fast acting and highly accurate unit that cost almost two thousand US dollars. And mine was by no means the most expensive model. A top of the range UEGO setup can cost around five thousand US dollars. Not that these units are any more accurate than mine. It's just that the more expensive units have features such as built-in data-logging, knock sensor readers, and so forth. Something which I don't need as I have a separate data-logging setup which has the ability to monitor knock-sensors, temperature sensors, etc.  

 

My UEGO sensor has a data-logger output that I connect to a virtual instrument. This instrument communicates through the RS232 port, to a portable computer that runs data-logger software. In addition to the UEGO sensor, I can data-log quite a number of other parameters such as boost-pressure, RPM, throttle-position, acceleration, etc.  

 

So what I'm trying to say is, if my UEGO sensor says that an engine's fuelling is running lean, at wide open throttle, as a result of fitting an HKS FCD, as happened yesterday with Justin's EVO... then that engine *is* running lean at wide open throttle. No mistake about it. in fact, seeing that 15.4 A/R was dead scary to the point where I literally shouted for Justin to stop.  

 

BTW, has anyone else been with Justin in his EVO? When I got out of the car the final time I felt sick. I'm sure glad I hadn't had anything to eat beforehand. The guy is the definition of manic.  :biggrin:

 

Note: The PB fuel-cut cannot be fitted to the EVO as there is an RPM wire to take account of.

 

Yours,

J

 

Link to comment
Share on other sites

 

Er, yeah Eric Kasir had a ride in JF's Evo at a Box Hill meet a few months ago.

I think he had a similar experience to you !!

 

Actually i may get myself an EVO at some point in the future.   they do sound like a hell of a laugh !

 

Link to comment
Share on other sites

Hi...

 

I would inclined to ask Racelogic about getting the Blitz ECU re-mapped. It would make sense that it could be set up for the charataristics of the car.

 

If you go to the Racelogic web site they also sell the re-mapping kits for various applications.

 

I read on the US mailing list that someone in the US has written an Java program to get access into the 96-98 ECU....although I suspect this is of no use to you though.

 

Gavin

Link to comment
Share on other sites

 

Thanks, I just looked at their website and sent them a couple of questions. Their ROM emulator stuff looked interesting but I bet it's not compatible with the MKIV. If it is, I'm gonna buy one!

 

Then we can do away with all this fuel-cut and top-speed limiter nonsense.

 

Yours,

J

 

 

 

(Edited by Ash at 7:01 pm on July 1, 2001)

Link to comment
Share on other sites

 

 

Yep, I'll do that.

 

You know, a fantasy of mine is that someone, somewhere, will crack the MKIV ECU. Because, ultimately, it's a nonsense that we all fit different cams, air filters, exhausts, and all that kind of stuff: yet we cannot tune those components, to the max, by altering the ECU parameters.

 

In fact, I know of cases where people have fitted complete new turbo kits yet have retained the stock engine ECU. Okay, the fuelling has been supplemented by some add-on cludge further up the rev-range. But what about lower down?

 

I reckon that a significant amount of the additional lag, caused by the fitting of larger turbos, could be tuned out of the system simply by running more ignition advance, for example.

 

Ultimately, what I would *so* much like to do, is not only fit HKS 256 cams (say) to someone's MKIV... but to actually have a modified fuelling and ignition map that would maximise that modification to the full. And for a reasonable cost, say, something in the region of £150... as opposed to the £1,500 cost of the Motec replacement ECU.

 

Oh well, let's see what Racelogic come up with.  

 

Yours,

J

Link to comment
Share on other sites

Anyone got a stock ECU that's been pulled out and sitting on a shelf at the back of the garage?

 

All these issues and Ash's last comment have prompted me to have a look at the internal architecture.

 

Obviously, I don't want to rip mine out, just to look at the component list.

 

If it's using stand-alone memory devices of whatever type, be they Flash EEPROMs or simple ROMs, it's possibe to stick in a ROMulator (mine's kicking around in a box somewhere). It's then a matter of determining the location of the data tables.

 

As a Japanese unit, it's most likely to be based around a 68K if it's CISC and the Hitachi SH family if it's RISC but that's exactly what I want to determine. If it's any of the obvious ones, reading the ROM and reverse engineering it should then be a chore rather than a challenge.

 

I promise not to destroy any loaned device although I'm going to have to 'go in'.

 

Link to comment
Share on other sites

Pete Betts and Adrian Thake have both done a lot of work on the ECUs, I know that Adrian has tracked back what each and every wire is supposed to do but neither of them have really hacked about with the ECU itself too much. The problem is the ECU controls so much stuff. Fitting something like a Motec is still a cludge on an auto as you need to leave the ECU in place to handle specific functions like the auto box. If you can get a Motec fully programmed with every single feature then it will work but as said before, the ECU controls so much in the car that you may miss a feature.

 

JB

Link to comment
Share on other sites

 

I understand exactly what the ECU controls on the car and understand the auto stuff as well. What I'm having difficulty doing is cracking open the actual chip.

 

The last computer I designed used a Motorola 6800 series uP and ran at about 1 MHz.

 

Are you saying that Adrian *can* modify the program within the chip?

 

Yours,

J

Link to comment
Share on other sites

Im not sure how far Adrian CAN go, but he certainly has exceptional electronics skills and might well have the right equipment to do it. He's a bit busy running his computer company at the moment so probably wont be able to come onto this list, but I will ask him just how far he can go with the ECU.

 

JB

Link to comment
Share on other sites

 

 

Okay, it seems Paul sounds like he knows his way around a microchip, and Adrian has done quite a lot of work and I know about the engine and gearbox stuff. Plus I used to be an electronics design engineer in the early 80's. But I've not got a lot of idea as regards modern-day processors.

 

So between us we may be able to make some progress on this.

 

One thing I was going to do a while ago is to get an engine and gearbox ECU up and running by using a computer to drive simulation devices.

 

My idea was to have a program that would simulate fuel-cut and so on. In the end, though, I didn't bother because the situation is I have the knowledge to set all that up and measure things, but I don't yet have the knowledge to change stuff inside the chip. It may well be that the info is blown into the chip and *cannot* be changed.

 

I've heard that other ECUs have the info stored in a separate ROM. In which case it's a simple case of using a ROM emulator. But the MKIV ECU doesn't appear to have a separate ROM. My instinct tells me this is probably the big stumbling block.

 

Yours,

J

 

 

Link to comment
Share on other sites

My 15 minute claim to fame was the speech in the Lagonda and the dashboard on the Jaguar XJ220.

 

Anecdote for you on that subject:

I profiled the instruments slated by Jaguar to go into the Jag (the mechanisms were off a MKII F*rd

Granada, if you care) and was able to characterise them down to half a percent. When I reported this to the Jaguar design team, they specified that they wanted them to read exactly 6% fast right across the range (6% is for legal reasons). Top Gear performed their review with one of those bike wheel odometers on the back and stated (on TV) that they were the most accurate instruments they had ever tested. What does that tell you about every other car?

Finally, a technical note: because the F*rd

instruments were calibrated to 140 MPH, I had to insert a 'soft' gearbox between the sensors and the panel so when it was doing 180 the clock thought it was actually doing about 120 MPH

 

Anyway, I'd love to play nice.

 

If it's based around an SH1 MCU, I can possibly get an ICE from my nice friends at Hitachi (Japan - so it wouldn't be anytime soon). The down side is that they tend to put the boot vectors in a small mask ROM in the MCU and it will be likely the tables are also masked in.

However, if they've just placed the boot vectors in mask, so long as we keep the same entry points in ROM, we could frig it.

 

If it's of 68K origin that tends to be less sophisticated and should be more hack-able.

 

Anything else is on a case by case basis. Motorola were pushing a PowerPC based MCU into a number of US ECUs in the early 90's (shudder) but hopefully they never got any footing in Japan. If it is, I'd redesign it for a better chip first.

If it's MIPS based (some Japanese companies got committed to MIPS in their embedded solutions), that might be irritating but I might be able to get support out of Philips (big MIPS supplier) if any of my old colleagues are still there (long time now).

 

If the device in question is a seriously integrated part then that might create issues re 'getting' at things without the specific device 'header' for the ICE and that might well be proprietary to the ECU manufacturer (e.g. Fujitsu).

If they've gone for a less integrated solution with external peripherals (extremely highly unlikely), we could drop an ICE in easily.

 

These are just some of the reasons I'd like to get an ECU's knickers off and look.

 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. You might also be interested in our Guidelines, Privacy Policy and Terms of Use.