Graphics interface for the 14CUX

Graphics interface for the 14CUX

Author
Discussion

Steve_D

13,798 posts

266 months

Wednesday 1st January 2014
quotequote all
stevesprint said:
....Although I’ve adjusted the throttle to read 97% on full throttle my ecu still very rarely/momentarily hits the bottom row of the main fuel table. Is that the correct behaviour?.....
I stand to be corrected but suspect the first row is there for when tickover 'stumbles' for some reason. On aftermarket ECU maps they often shift both fueling and timing to help the engine recover from its stumble.

Steve

Steve_D

13,798 posts

266 months

Wednesday 19th February 2014
quotequote all
A question for the experts please.

14CUX in a TVR.
When first started car holds a high idle which slowly reduces.
When warmed it hold the 900 idle that RoverGauge says it should.
All correct so far.
After a while with the engine well up to temperature RoverGauge shows the speed as 36mph even though the car is stationary. Sometimes the speed displayed jumps about. If the car is driven the speed remains steady and the speedo works correctly.

On the face of it I have to assume the speed module has gone wrong. I plan to send it off for testing/repair.

Now the bit I don't understand.

If the car is driven the stepper motor pregresively opens until it is full open with the tickover at about 2200rpm.
If the speed unit has failed and giving a reading of 36mph why does the ECU not just maintain a fast tickover of about 1100 as it is supposed to?

Confused.

Steve

Pink_Floyd

900 posts

229 months

Thursday 20th February 2014
quotequote all
Steve_D said:
A question for the experts please.
14CUX in a TVR.
After a while with the engine well up to temperature RoverGauge shows the speed as 36mph even though the car is stationary. Sometimes the speed displayed jumps about. If the car is driven the speed remains steady and the speedo works correctly.

On the face of it I have to assume the speed module has gone wrong. I plan to send it off for testing/repair.
Confused.
Steve
The speed in the ECU and the speedo are not linked. Wen the car is stationary you should see a speed from the ecu of 0, ecu now knows that the car is not moving and adjusts tye stepper motor to keep the tick over at the correct value, once the car is moving you should see a speed of 36 no matter how fast you are actually going, this tells the ecu to close the stepper motor as its not required to maintain the engine speed. Its all there to stop the engine from stalling when you slow down and finally come to a stop.

spitfire4v8

4,017 posts

189 months

Thursday 20th February 2014
quotequote all
Rovergauge changes for mappers ?

Now that the mapping addresses have been available in the wild for some months is anyone else geting to grips with mapping ?

I've done a few cars now and there's a couple of things that would make life a little bit easier during set-up and mapping.

Whilst mapping the idle and light throttle use I find myself constantly changing between the long and short trim bar displays in rovergauge to see how the adjustment trends are changing there. I wonder if it's possible in a future rovergauge update to allow viewing of the both the long and short term trims simultaneously?

Also could there be a numerical raw voltage displayed for the throttle position sensor. The percentage bar graphs are great for watching for a smooth and consistent increase in output as the throttle is applied, but for initial setting of the position sensor I still use a voltmeter .. would be great if that voltage was displayed on the screen also.

Have to say though that the combination of rovergauge and tunerpro / moates ostrich is a delight to use.

danbourassa

246 posts

145 months

Thursday 20th February 2014
quotequote all
spitfire4v8 said:
I wonder if it's possible in a future rovergauge update to allow viewing of the both the long and short term trims simultaneously?
I probably shouldn't speak for Colin on this but I think I know his concerns. RoverGauge is being used on almost every continent (we haven't gotten a email yet from Antarctica) and the majority of users are not re-mapping. There is alway a trade-off between optimum default window size for the typical laptop, keeping the screen uncluttered and just displaying what the average user wants. To display both short and long trim, something else would probably have to go. Maybe the speedometer? I don't know if anyone finds that useful.

I would be more inclined to do a special version of RoverGauge just for re-mappers. That way, changes could be made without concern for the general user. I think he and I need to have a beer over this.

Steve_D

13,798 posts

266 months

Thursday 20th February 2014
quotequote all
TVR owners do need the speedo.
With this you can tell if the speedo module used by TVR is working properly.

Steve

spitfire4v8

4,017 posts

189 months

Thursday 20th February 2014
quotequote all
danbourassa said:
I probably shouldn't speak for Colin on this but I think I know his concerns. RoverGauge is being used on almost every continent (we haven't gotten a email yet from Antarctica) and the majority of users are not re-mapping. There is alway a trade-off between optimum default window size for the typical laptop, keeping the screen uncluttered and just displaying what the average user wants. To display both short and long trim, something else would probably have to go. Maybe the speedometer? I don't know if anyone finds that useful.

I would be more inclined to do a special version of RoverGauge just for re-mappers. That way, changes could be made without concern for the general user. I think he and I need to have a beer over this.
Could it be done maybe by having two pages ? a front page for the general interface as it is now (and it works brilliantly) .. and page two for map tinkerers ..

davep

1,143 posts

292 months

Thursday 20th February 2014
quotequote all
spitfire4v8 said:
Rovergauge changes for mappers ?

Now that the mapping addresses have been available in the wild for some months is anyone else geting to grips with mapping ?

I've done a few cars now and there's a couple of things that would make life a little bit easier during set-up and mapping.

Whilst mapping the idle and light throttle use I find myself constantly changing between the long and short trim bar displays in rovergauge to see how the adjustment trends are changing there. I wonder if it's possible in a future rovergauge update to allow viewing of the both the long and short term trims simultaneously?

Also could there be a numerical raw voltage displayed for the throttle position sensor. The percentage bar graphs are great for watching for a smooth and consistent increase in output as the throttle is applied, but for initial setting of the position sensor I still use a voltmeter .. would be great if that voltage was displayed on the screen also.

Have to say though that the combination of rovergauge and tunerpro / moates ostrich is a delight to use.
spitfire4v8 have you ever thought about creating a video/webinar of a 14CUX mapping session and putting it up on youtube, along the lines of Tunerpro stuff:

http://www.youtube.com/watch?v=HpUaeiJYk3I

spitfire4v8

4,017 posts

189 months

Friday 21st February 2014
quotequote all
I map on a rolling road so not really applicable to most people maybe but you could do the same thing on a nice stretch of open road with no traffic if you were careful so it's something to think about..

however something that will be of interest to many many people will be stevesprint's data logging package taking wideband lambda sensor readings and cross referencing with the map numbers will allow many people to start refining their fuel map at home, and set-up costs are just 150 pounds for a wideband sensor.

blitzracing

Original Poster:

6,410 posts

228 months

Friday 21st February 2014
quotequote all
Steve_D said:
A question for the experts please.

14CUX in a TVR.
When first started car holds a high idle which slowly reduces.
When warmed it hold the 900 idle that RoverGauge says it should.
All correct so far.
After a while with the engine well up to temperature RoverGauge shows the speed as 36mph even though the car is stationary. Sometimes the speed displayed jumps about. If the car is driven the speed remains steady and the speedo works correctly.

On the face of it I have to assume the speed module has gone wrong. I plan to send it off for testing/repair.

Now the bit I don't understand.

If the car is driven the stepper motor pregresively opens until it is full open with the tickover at about 2200rpm.
If the speed unit has failed and giving a reading of 36mph why does the ECU not just maintain a fast tickover of about 1100 as it is supposed to?

Confused.

Steve
The speedo calibration unit is not well designed- the input circuitry is prone to noise- the problem is there is a comparator chip that senses when there is a pulse from the diff sensor, using the signal as one input against a ground reference. The problem is with no signal from the diff, both inputs can sit near 0 volts, and it only takes a small amount of electrical noise on the earth input to produce a false trigger signal. This can be as simple as bad solder joints on the PCB to make the earth on the PCB float up a volt or two.

Tick over wise, you need both the throttle pot to be at its correct shut throttle voltage, and no speed signal to bring it into idle mode. RoverGauge has a nice green light on it for when this state is reached, and it should then idle at around 800 rpm (dependent on chipping). Off the top of my head I dont know what the car moving idle should be, but I know MA chips lifts it a bit to reduce rear wheel lock up between gears, but 2.2k sounds a bit high.

Steve_D

13,798 posts

266 months

Friday 21st February 2014
quotequote all
blitzracing said:
The speedo calibration unit is not well designed- the input circuitry is prone to noise- the problem is there is a comparator chip that senses when there is a pulse from the diff sensor, using the signal as one input against a ground reference. The problem is with no signal from the diff, both inputs can sit near 0 volts, and it only takes a small amount of electrical noise on the earth input to produce a false trigger signal. This can be as simple as bad solder joints on the PCB to make the earth on the PCB float up a volt or two.

Tick over wise, you need both the throttle pot to be at its correct shut throttle voltage, and no speed signal to bring it into idle mode. RoverGauge has a nice green light on it for when this state is reached, and it should then idle at around 800 rpm (dependent on chipping). Off the top of my head I dont know what the car moving idle should be, but I know MA chips lifts it a bit to reduce rear wheel lock up between gears, but 2.2k sounds a bit high.
Mark
Thanks for the reply. You mentioned this earthing issue a while ago but for the life of me I could not find it. At that time you offered a cure which I believe was a resister fitted somewhere.
Could you explain again what and how?

I've sent the module away to be tested but I suspect it will come back 'no fault found'. Whichever way we can eliminate that as the problem.

The symptoms we have been seeing are more likely to be an ECU problem as I can't see how the speed module could cause the stepper motor to progressively wind itself to the end stops. The speed module should give a speed of either 0 or 36mph (say). I can't see how anything else it could produce would cause the stepper to open.

Steve


blitzracing

Original Poster:

6,410 posts

228 months

Saturday 22nd February 2014
quotequote all
This is from Mike Hayes who originally worked out the issue.

A link to the data sheet for the device used in the ECU speed converter :

http://www.ti.com/lit/ds/symlink/lm2907-n.pdf

The attachment is one of the pages from the spec and from memory the circuit which shows a magnetic speed sensor is exactly the front end of the circuit in the box under the dash.

The circuit needs in built tolerance to supply and sensor noise when used on a vehicle with lots of ignition and earth noise. The following may help understanding.

Basics - The sensor is pulsing into a comparator. A comparator as it suggests is comparing the voltage level of one input to another, and when the voltage on the sensor input is higher than the reference input the output of the comparator will be low. Conversely when the voltage on the sensor input is lower than the reference input the output of the comparator will be high.

The flaw in this circuit is that when the sensor is in an idle state the input is zero volts and as it happens the reference is at zero volts, so the comparator is in no mans land where it's unstable. I'm guessing that there is a small bias at the reference in this particular device that makes this device a few tens of millivolts high, in order to work at all. However it only takes a few millivolts of supply or sensor noise to start triggering the comparator.

This is what I did to resolve it.

I cut the earth connection to pin 8. I then soldered a two resistors to pin 8 one to earth and one to Vcc (pin 6) or any other convenient supply line. Not sure what values I used but if I was doing it now I would go for 100 ohms to earth and 1Kohms to Vcc. The aim is to raise the voltage at pin 8 to about 1.5V. If you have room to add a 0.1uF capacitor across the 100ohm resistor ie Pin 8 to earth that will also help.

The wires on the SPEEDO are Green = +12V, Green/Yellow = Signal from speed sensor, Black = 0V (Grd) - may be helpful to identifying the wires on the converter

My bit:

I have a simpler mod, just involving an additional diode to the earth line- so if you can send me your "faulty" one, Ill send you this one Ive modified as a swap so you can see if it works- I dont have a car to test it on.

Steve_D

13,798 posts

266 months

Saturday 22nd February 2014
quotequote all
Thanks again Mark.
I've sent the module away to be tested so will have to wait until it comes back. When it does I will get the customers car back in and see if the problem is still there. If it is I will first try your resistors unless you believe the other fix is better. Is it something I could do myself. I have a little experience soldering boards (built two Megasquirts) and I have a Metcal soldering iron.

Steve

blitzracing

Original Poster:

6,410 posts

228 months

Sunday 23rd February 2014
quotequote all
My Theory is to remove the black earth wire from the PCB, and solder a small signal diode between the PCB earth and the black wire, in the direction of current flow (ie + to the black wire). The idea is to put the forward voltage drop of the diode onto the earth plane of the PCB, so it sits at about .7 of a volt above the chassis ground. This means that any stray trigger voltages will have to be .7 of a volt or more before they have an affect. Its just an idea as I cant test or measure the voltages in its original situation.

karlspena

24 posts

134 months

Tuesday 11th March 2014
quotequote all
chris_chim said:
Yes - the Raspberry Pi runs various forms of Linux but the standard one is Raspbian (based on Debian) which is what I used. It has an ARM processor so rovergauge needs recompiling from the source code but this was pretty straightforward. For more info on the Pi see http://www.raspberrypi.org/ in short it is ideal for DIY electronics projects and affordable.
I have a RPi laying around. How do I install RG in it?

Steve_D

13,798 posts

266 months

Friday 14th March 2014
quotequote all
blitzracing said:
My Theory is to remove the black earth wire from the PCB, and solder a small signal diode between the PCB earth and the black wire, in the direction of current flow (ie + to the black wire). The idea is to put the forward voltage drop of the diode onto the earth plane of the PCB, so it sits at about .7 of a volt above the chassis ground. This means that any stray trigger voltages will have to be .7 of a volt or more before they have an affect. Its just an idea as I cant test or measure the voltages in its original situation.
Hi Mark
The module has come back 'no fault found' so the next step is to try your diode.
Can you give me a suitable part number for the diode please?

Thanks
Steve

blitzracing

Original Poster:

6,410 posts

228 months

Saturday 15th March 2014
quotequote all
Any diode will do- you just want the .7 voltage drop

http://www.ebay.co.uk/itm/1N4007-General-Purpose-D...

Steve_D

13,798 posts

266 months

Saturday 15th March 2014
quotequote all
Just found some 1N4001 left over from a project. Will they do?

Steve
ETA We have at least 2 customer cars with unstable speed readings when at standstill.
If those are OK I will do the mod and report back.
This could be a good way forward for lots of owners.

Edited by Steve_D on Saturday 15th March 20:33

blitzracing

Original Poster:

6,410 posts

228 months

Sunday 16th March 2014
quotequote all
Thats fine.

chris_chim

207 posts

245 months

Wednesday 2nd April 2014
quotequote all
karlspena said:
chris_chim said:
Yes - the Raspberry Pi runs various forms of Linux but the standard one is Raspbian (based on Debian) which is what I used. It has an ARM processor so rovergauge needs recompiling from the source code but this was pretty straightforward. For more info on the Pi see http://www.raspberrypi.org/ in short it is ideal for DIY electronics projects and affordable.
I have a RPi laying around. How do I install RG in it?
Hi Karl
I have just recompiled version 0.7.1 using QT Creator 4.8 on the RPi and it seems to work. For this version I needed to muck about with some of the functions that don't seem to be supported in Qt4.8 but I have now built it and it seems to work. You also need to build libcomm14cux which requires Cmake to be installed on the RPi and install the RPi drivers for the ftdi device from their website.
To install Cmake use:
sudo apt-get update && sudo apt-get upgrade
sudo apt-get install cmake
Then just follow the instructions in the libcomm14cux source code readme file to build it.
I think I used these instructions to install QT creator:
http://qt-project.org/wiki/apt-get_Qt4_on_the_Rasp...
I then modified an earlier project file to include the current .cpp and header files from the rovergauge source and built it in Qt creator on the RPi.
If you want my compiled executable or modified source code let me know.
Cheers Chris

Edited by chris_chim on Wednesday 9th April 20:23