2 May 2009
This weekend is May bank holiday. Usually I go to the Stoneleigh kit car show, one of the biggest of the year but I decided that I need to plough any time I have into finishing of my car! First Job is to wire in the oxygen sensor. I had to cut the fly wire to the connector you can see in the photo below. This was very fiddly as I had to reuse the connector. Removal of the pins which click into place is pretty difficult if wanting to preserve the connector. I did try to source a new connector but it seems a special one. I wired up the supply from the relay to the sensor along with ground. Megasquirt ECU will switch this on along with the fuel pump. Issue with this is that the sensor needs to be calibrated to the lambda sensor controller and this needs to be done with the sensor in free air! So I bridged a permanent live to the controller, prior to this I removed the sensor from the exhaust header and held in a ring spanner as this is heated sensor and I didnt want to burn my fingers. Calibration is achieved by pressing and holding down a button on the lambda sensor controller. A LED blnks rapidly for a number of seconds and hey presto it's done. So left the sensor to cool down and installed in the exhaust headers for the final time.
Here's a photo of the oxygen sensor installed in the exhaust header.
The lamda sensor that I have employed is a wide band one which provides more accurate detection of fuel to air ratio mix. Standard narrow band are limited to being able to dector lean, rich or just right! I have a dash gauge which is quite nifty showing the ratio as a bar graph display as well as digital read-out. The bezel around this gauge like like it might be an issue for the SVA so I shall leave this out for the time being. However I am connecting this sensor controller to the megasquirt ECU which is just a matter of connecting WBin+ to pin 23. WBin+ is the blue wire from the controller as shown in the diagram below. WBin- or the green wire is connected to ground.
Pictured below is the wideband controller on the left. On the right is the megasquirt ECU. The smaller connector on the left side of the ECU is a hook-up to the laptop running megatune software.
Here's the laptop running megatune connected to the Megasquirt ECU showing what's going on. The lambda sensor seems to be sending information to the ECU, reading over 19.4 parts oxygen to 1 part fuel which is the leanest value, engine is not running so probably correct.
3 May 2009
MAP which stands for manifold absolute pressure, provides information to the ECU so that it can adjust the fuelling. Another issue I had to overcome here was that the sensor requires a 3mm vacuum tube but the feed from the manifold is 8mm! Drove around various places to try and find a solution. In the end I found a right angled plastic pipe connector which was about 5mm in diameter. My plan here is to stretch the 3mm tube over one end and jubilee clip the 8mm tube to the other. I think it's air tight, also its -ve air pressure, I shall see if this works consistently when the engine is fired up.
Had to drill another hole in the scuttle for the vacuum hose, its the one going through the tight rubber grommit. Turned the ECU and tested, reading 100Kpa which I am lead to believe is standard atmospheric pressure at sea level. Also noted that the manifold temperature was 18 degrees which giving it is 17 degrees outside the garage is probably correct. Coolant temperature readout is 19 degrees, not so sure about that but it cannot be far out.
4 May 2009
Session 284 - TPS sensor
TPS stands for Throttle Position Sensor. This is simply a three pin potentiometer. One pin will be tied to ground, the other pin will be supplied +5v reference from the ECU and one pin will be the sensor. The middle pin is the sensor. The megasquirt site has instructions on how to work out which of the outer pins are connected to ground and +5v. Trick is to give a rising resistance as the throttle is opened. I found the top pin was ground and the bottom pin the +5v ref.
All these connectors and I don't think I have yet detailed how to install these common water proof connectors, so I am doing this now.
Strip the outer sheaf of the wire to reveal the inner wires, you want approximately 25mm length. Now strip some of the insulator to expose 4mm of the copper core. I have a precision tool for this. Now twist the copper strands together. Next install the grommits oriented as shown below.
Next tin the wires with a soldering iron. You don't have to do this but I find it gives a better mechanical join to the connector when crimped as the solder deformes less readily than copper strands and this provides a more positive if not harder crimp.
Once the solder has cooled down push up the grommits so the end is level with the end of the wire insulator as shown below.
Now to crimp on these connectors you will need the special tool. Others have used pliers etc but there really is no substitue for the tool you can see pictured. Place the connector on the witre wo tha tthe copper is in the channel and the rubber grommited located as shown below. Sometimes this is easier to do with the connector held in the crimp, it depends on the style of connector.
Photo below showing the crimps complete. As you can see the rubber grommit is held in place which also holds the wire by the insulator and the copper core is crimped very well and the mechanical joint is very strong. I have tried pulling one of these off before. The wire breaks before the joint.
Next is to push the pins into the connector, double check that pins are in the correct positions before pushing in because ones these pins are in, they click into place and are not comming out again, ever! there is a special tool to remove which I do not have but I went to an automotive electrian once with a connector I had cut off whch I wanted to save and he could not remove the pins even with the tool. Looking at the photo below you will see that the pins once located in the housing are sealed by the grommits. Once the pins are home a guard is clipped on the back which further seals and also locks the pins into place not they will ever come out anyway. You will need long nose pliers to push the pins in at the base of the wire, quite a firm fit.
Rear guard clipped into place
View of front of connector showing pins located
Finally the connector installed onto the TPS.
Next I powered up the ECU again. Calibrated the TPS which involves taking a reading of the throttle closed and then a reading a throttle fully open. I then moved the throttle from closed to fully open steadily whilst looking at the throttle gauge in megatune, yep that's working.
5 May 2009
Session 285 - Crank position Sensor
Left the biggest, most important sensor to last. You may remember the crank trigger wheel and VR sensor I installed a while ago, well know it's time to hook it up. The VR sensor will be wired into the EDIS ignition module and the EDIS will output a signal to the Megasquirt ECU and in turn the Megasquirt will send an ignition advance signal back to the EDIS module. I have discussed this at greater length previously, anyhow you can go to the megasquirt website or simply google for EDIS8.
Here's a diagram I have shown before.
So today I am wiring in pin 4, 5, 7, 1 and 3. When I installed the connector on the back of the VR sensor it had two wires, red and black.I hadn't checked if the sensor had a certain polarity which as you can see from the diagram it does. It transpires that my -ve is red and +ve is black!
With exception of the tach signal I need to wire in which I am hoping pin 2 - IDM will suit that concludes all the wiring.
8 May 2009
Session 286 - Fuel pressure test
My special order relays came in today, these are Double Pole, Single Throw I used these for neater wiring. Anyway I can now start performing some testing prior to attempting to start my engine for the first time tommrow. Inserted 9 litres of fuel into my tank. Paused and checked all round the tank and line to fuel pump looking for leaks. No leaks so processed to turning ignition on, ECU sent a several second signal to the fuel pump relay to prime the pump, instant 3 bar pressure (42psi). fuel pump off and pressure dropped steadily over 2 minutes. I checked for leaks from pressure rail all the way back to the fuel pump, again none. I bringed the relay with a piece of wire to turn the pump on for a minute. Checked for leaks again. Whilst I did this I removed the fuses for coils, injectors etc, if I had a leak I do not want to chance a source of ignition, daunting stuff, I had my fire extinguiser right next to me.
Downloaded ECU configuration from a fellow builder (combatsapph), reviewed it and accepted it as it was and burnt into my ECU via PC.
Replaced the fuses for the coils and injectors. checked that power supply to these on the connectors, no problems there. This concludes all the pre-start checks. I am ready to start the engine but I am doing this tommrow morning. I am in for a restless night I think.
9 May 2009
Session 287 - Engine first start attempt day
It was 03:00am this morning when I finally managed to get to sleep. Very anxious about today, it would either be very good or bad! Over three years in the making consuming over 500 man hours; D-Day has arrived.
Okay first, double checked for leaks, found none. Check oil level, check coolant level. Removed spark plugs and cranked engine over, read 1bar of pressure from oil guage, good, oil pump is primed. Replaced spark plugs but left off HT leads and cranked over, oil pressure reading now just under 2 bar.
Here goes, attached HT leads to spark plugs, ignition and a single press of the start button, instant bust of life, wow, whoah, I was not expecting it to fire first time and so instantaneously. Idle speed needs setting as I had to hold the throttle. little rough at slow engine speed, timing probably out a bit. Let run for 30 seconds whilst watching as many guages, pipes as possible. Oil pressure seems to be around 3 Bar at fast idle.
Removed spark plugs to examine. All seemed to have fired so each getting fuel and spark. Plugs seem to be quite wet with petrol. so a little rich but better than too lean as this can cause major damage.
I noticed a small oil leak from the dirver side front corner of the sump gasket. This I am not surpised about this as I had a bit of trouble pushing the sump up straight as the front cross member limits this.
I cannot believe how lucky I am this worked first time, after all I have built this up from scratch, we not talking ready assembled engine or all standard parts, its a collection of Rover,ARP, Piper, Mustang, Ford, Real steel, parts fabricated by myself and specialist companies not to mention a hand built ECU, wow.
Once I have sorted out my small oil leak and ran in the cam properly, i.e. brought the engine temperature up and discard the running in oil I shall start work on checking the static timing then I shall start tuning at bit more. Well that's the plan which may change.
The Megasquirt ECU is tuned using open source software called megatune. Screen shot below.
These guages are just a small selection of the many available. Settings and tuning parameters there are 9 pages of them. Below are screen shots of the two most important, spark advance and fuelling maps.
Going to have hours of fun with these. Just imagine having different settings, one for normal road use, one for track days, etc. You can even change settings whilst the engine is running! You can log in real time whats going on as well although you need a PC connected via the serial port. There's some more open source software called megaviewer which you can use to analyse the log file and play back showing what's be going on. See screen shot below.
The other bit of hardware that I can connect upto my PC is the digital dash. This sofware allows you to change the sensors, calibrate them, change the dash layout and much more. Screen shot below.
13 May 2009
Session 288 - Week night fettling!
Yesterday I called up Westfield to ask about the RT dash sensor calibration as I was not happy with the accuracy of the readings. It transpires that they have a revised configuration file. I downloaded this file and compared with mine and merged them together.
Fuel sender settings. This gives a reading between 0.1 and 1 I might scale this up a bit later so that it effectively reads in litres:
Oil pressure sender settings. This gives a reading in bar. 1 bar equals aproximately 14.5PSI or 100Kpa:
Water temperature sender:
RPM configuration:
Gear ratios entered so that gear indicator can be shown in dash:
I have compared water temperature with that from the ECU which uses another water temperature sender, seems to be with 2 degrees of each other. Oil pressure reads as expected. Fuel level seems to be correct to what I have jerry canned in. I had to adjust the number of pulses for RPM as it was set to a 4 cylinder car, now RPMs match that read from the ECU.
14 May 2009
Session 289 - Fuel mixture settings
After I started the engine for the first time it was a bit obvious that the fuel mixture was a bit rich, pratically spits petrol out of the exhausts. I had removed the plugs and could see that they were blackened. So tonight I'd thought I would be brave enough to run the engine long enough for the lambda oxygen sensor to heat up enough to take some Air to fuel ratio readings (AFR). The ideal AFR is 14.7 that is 14.7 parts of oxgen to 1 part of fuel. I am told that a at idle that I should aim for 13.2. The ECU has a map for AFR targets as shown below.
So the only value I am changing is the require fuel for one cylinder (ms). I reduced this value down by 1ms at a time and noted the AFR ratio from the megasquirt dial display and continued until I got around 13.2.
Changing the fuelling down as actually changed the engine note. Seems smoother especially at idle. The other thing I noticed, totally unrelated, is that the battery gauge in megasquirt is reading 14.2 which means my alternator is working.
Go to June 2009