Established 22 years
I've had three new rear speed sensor brackets 3D printed (Thanks Andy Laurence). I've fitted them on the gearbox and the sensors are now much more rigid. I was adjusting the sensor gap on the offside, when I found a dead spot on the reluctance ring. There is a tiny difference in the height of the teeth around the ring. I should have had them machined when they were fly-pressed on to the outputs cups. So a few more minutes spent reducing the gap to the teeth and I think the sensor is now working properly, detecting all 44 teeth as I rotate the rear wheels. No wonder I was losing the frequency from the X10 and the rear speeds were failing.
I've fitted another pair of thermocouples to the EGT-CAN box. One is measuring the temperature of the fuel in the large 10l Gemzoe tank, and the other is attached to the entry to the intercooler. I've set the ECU to log both channels, and I'll have some useful data I hope after the Combe event. I've also relocated the starter battery from inside the car, to the rear of the intercooler side-pod. The battery is now on yellow Anderson connectors, so I can isolate it from the chassis electronics to prevent the parasitic load from the Cartek isolator from draining it between events.
I am car number 8 at the Dick Mayo sprint, seeded number 1, so no pressure then LOL. I've added a gurney to the near side barge board so it matches the gurney thats built-in on the opposite side. This should again help the board deflect more air.
The barge boards are now modified and fitted back on the car. I've ordered the battery posts and ring terminals to allow me to relocate the Lithium Battery to the intercooler side pod. The parts should be with me next week so I can crack on with that piece of work. The battery will be connected to the loom using a small Anderson connector to allow it to be rapidly removed.
The professional pictures from Blyton are below.
Spent the afternoon cutting out a set of new firewall panels to fit both the sides of the drivers seat. These should isolate me from the temperature buildup that I was experiencing on Tuesday. At speed my right side felt like it was burning from the heat travelling forward from the turbo in to the cockpit, which was most distracting. I've also removed both the Dallara barge boards and I've cut out some carbon and stuck it in to the gaps on the ends of the boards, to force all of the air outboard of the floor, which should help seal it even more effectively.
I've got some more K Type thermocouples on order, so I can monitor the gearbox oil temp, fuel temp, and the air temp before it travels through the intercooler. I'll get these wired up and the ECU configured to read them in the next few days.
Two weeks to go to the Dick Mayo sprint, and I'm practically ready.
Testing went well at Blyton on Tuesday. I had a very high speed spin in the morning, at 106MPH losing the rear end around Port Froid, which was caused by too much toe-out with the larger front wheels fitted on the car. Later I adjusted the steering arms three times, taking a 1/2 turn off between each drive in the afternoon, to reduce toe-out and it definitely helped remove the dead spot from the steering. But I still wasnt comfortable with pushing the car near the limit, so for Castle Combe I've reset the ride height etc, so I can use the 2019 tyres. I think with more testing the larger tyres can be made to work, but I dont want to be making changes at Combe, and I at least have a starting point when I go back to the larger tyres. If I get to drive at Aintree in September, I'll probably try them again. I reached 2.5G in the corners, and on the spin I reached 2.8G so there is definitely more grip from the chassis this year. The floor and barge boards are all contributing. The Yaw angle sensor also worked, and I now have lots of data to pour through.
Mechanically the car was perfect. No issues with the gearbox, engine, or the temperatures. The steering wheel electronics worked perfectly, with no missed shifts, and the traction control (when it worked) worked brilliantly. The front and rear wheel speed sensors were throwing up errors which compounded the TC testing. The printed brackets that hold the rear sensors were allowing the sensors to move, which caused them to fail, so a new pair of brackets with additional support for the sensor are being printed. The front left sensor continued to fail, and this is a problem I've lived with for some time in 2019, so I've replaced it with a new one, and wired it in directly to the loom, having removed the rubber sureseal connectors which I loathe. I'll see if its solved the problem on the 18th July.
The front left sensor deteriorates to the point where it stops altogether. Front right is working 100%. The two rear sensors exhibit issues under acceleration/deceleration.
Andrew O'Malley has done a fantastic job of the switch box case that he's made for me. He's basically made it from layers of polycarbonate, with a carbon fibre lid, which is attached using magnets to the case body. It looks brilliant and houses the CAN switch board, keeping it out of harms way on the rear of the dash4pro display.
The steering wheel is finished, and I have configured the traction control to use the additional TC Add switch that I've interfaced. I'll see how it performs when I go testing on Tuesday.
My G logo routed in to the carbon lid
The inside of the box, showing the magnets on the four corners that secure the lid in place.
Once I carefully drilled two holes in the side to allow the wires to pass in and out, the box is then stuck to the carbon back plate using sikaflex
The wires are carefully pushed down inside the case to allow the lid to fit properly
The steering wheel shows the new rotary switches, with some hint sheets stuck to the chassis for the TC Add and TC % settings.
The ECUMaster USB-CAN box arrived, that allowed me to program the CAN switch board (cswb). It has a pair of flying leads that you just solder across the CAN Hi and Lo wires on the cswb, and it is powered from the USB port on the laptop. With the box running, I could see the cswb on CAN, and that allowed me to run the Light Client program and setup the number of switch positions for each rotary switch, so they returned the correct value (1 to 12) to the ECU.
The first thing I tackled was the paddle switches. We know that the analogue voltage returned on the A frames was a 16 bit format (0 to 5000), and therefore the ECU couldnt use that data. So the implementation I ended up with was to wire the paddles to analogue 6 and analogue 8, and use B frame 3 and 4, to read back the rotary switch values that the Switch box calculates according to the number of resistors chosen in the Light Client. With analogue input 5 unused, the leading 4 bits of the value returned are 0000, with the trailing 4 bits representing analogue input 6. The Switch board needed configuring to use 2 resistors with an offset of -1. So instead of returning 1 / 2 for the two positions of the paddle switch, the -1 offset meant it returned values of 0 / 1. With these values returned, I was able to assign Gear Shift Down Paddle Switch to use X CAN RECEIVE B #03, and Gear Shift Up Paddle Switch to use X CAN RECEIVE B #04. Testing in the garage shows that these both worked, and the pneumatic solenoids both fired when I pull the paddles with the clutch depressed switch closed. Result.
Basically the same approach had to be taken for two of the four rotary switches. I've used analogue 2 for the TC % switch, configured it for 12 resistors with an offset of 0 to return 1 to 12, and again analogue 4 for the TC Adder switch, configured the same way. And these both work. The ECU expects a value from 1 to 12 over CAN as an alternative to seeing an analogue voltage. I'm using X CAN RECEIVE B #01 and #02 for the two rotary switch inputs.
So at the moment, I've only used the even analogue channels, 2-4-6-8, and the two remaining rotary switches are disconnected, and I've left the original CAL and Launch RPM switches connected to the X10 enabled, so I can still select the power map and launch RPM using those switches.
The way forward is to find a different CAN Switch board, one that presents the data in a friendlier way. Add a second ECUMaster CAN Switch board, to allow four more analogue channels to be connected. Buy a CAN Dashboard, that can read the values from the Switch board, and convert them in to voltages. And potentially a few other solutions are available I'm sure.
A big thankyou to James Middleton from Relentless Performance for loaning the USB-CAN box, and Richard O'Donovan from Zen Performance for the tech support.
ECUMaster Switch board configuration
Paddle switch configuration using the B frames 3 and 4 from the Switch board
With the steering wheel complete, I am a step closer to getting the switch board working today. And a few more challenges to overcome. Because the board isn't slaved to the ECU, the data that it presents are unitless. So 5000mV is presented as the number 5000. Not mVolts, just 5000. So the ECU sees the number, and thinks that it represents the value of 5000 Volts. And we cant change the format of the data. So getting the ECU to work with the analogue voltages presented by the board isnt looking good. Ordinarily, if we were just capturing and logging the value, we could use the customizing options in the ECU to change the format. But if I assign a received sensor to a can frame, and change the format in to say Volts, I cant then assign the same frame to TC Cal. It just doesnt work like that.
So Plan B (or is it C) is to use B Frames (8 bit bytes) to read the third frame from the switch box, which returns the actual rotary switch positions as a value from 1 to 12. But, the switch board returns a pair of rotary switch positions per byte. So byte 1 holds switch R1 and R2 eg the first 4 bits are for R1, and the second are for R2. That throws up another challenge. I dont think the ECU can be assigned to read a frame with a mask to ignore the first or last 4 bits.
Another thing we want to try is to enable the pull-up resistor on the two analogue channels that the paddle switches are connected to, and see if the switch board then outputs 0 and 1, rather than the analogue voltage on the switch. So I need to obtain a USB-CAN box from ECUMaster, to reprogram the switch board, and thats wednesdays job.
The four rotary switches are now connected to the CAN Switch board, and the ECU is reading the voltages from them. I've renamed the CAN channels to use friendly names, and the next job is to rewire the paddles to use the next two analogue channels on the Switch board, which I've now done, and I'll test the paddles in the morning.
Rotary switches returning 5000mV (Position 1 on all switches)
Rotary switches returning 430mV (Position 12). Switch #2 rotation is locked by a pair of pins, and can only select from position 1 to 4 since it selects the calibration maps 1-4
I then removed the assignment from the Beacon input on AN#03, since its not used by the ECU, and added wheelspin to the logged channels using LifeCFG, so I can gauge how well the TC is working.
Here is the finished switch plate, with the four rotary switches mounted and wired together. Soldering the PCB's to the rear of the switches was very easy, they certainly take the effort out of assembling everything, and look far tidier.
It looks like the ECU only had four frequency inputs. I tried to connect the exhaust VVT sensor to AN#16, but according to the Life manual for the F88R, only channels 9 to 12 are capable of measuring frequency. These are all used already. So to connect the exhaust VVT to the ECU, I'll have to move the front wheel speed sensors to the X10.
Another thing I noticed is that the ECU is still configured for the Beacon input, on pin AN#03. As I'm not using that input, I'll unassign it.
The switch PCB's turned up from @Trackformula and I quickly set about soldering the rotary switches to them. They make the construction of the switches so much easier. Instead of trying to solder 11 0.25W resistors to the rear, I just place the switch on the PCB and solder 13 pins. Simply brilliant.
Yesterday I renamed all of the 8 received channels from the EGT-to-CAN box, to friendly names, in the Customising Options menu. I also changed the default values to 700C where the sensor fails (or isnt connected).
I then loaded a tweaked linearisation file for the Yaw Rate sensor, after tweaking the data I generated in Excel.
I also updated the wiring diagram schematic on my Life F88R page to show the Yaw sensor connection and the addition of the CAN-Switch steering wheel circuit board to CAN1.
Now I'm focusing on loading some slip targets for traction control to work with Yaw Rate rather than Lat G. If you have a known working set of spin targets for Yaw, please get in touch.
I have fitted the Bosch Yaw rate sensor and it works really well. The sensor outputs 2.5V when there are no rotational forces working on it, and as the sensor is rotated left and right the voltage rises and drops in proportion to the angle of rotation, and it then returns to 2.5V again when the rotational force stops. I had to create a new linearisation table for it, which I've saved in my Sensor DB folder. This ensures that the output matches the characteristics described by the manufacturer. I've setup TC to use Yaw instead of Lateral G and I'll now work on creating some new spin targets. I'll have four maps to select from, with the brace of new switches on the steering wheel. The ECU can read the Yaw rate at 100Hz, which should make the traction control very responsive to changes in direction.
Finally I reconnected the steering wheel angle sensor, and reassigned the SWA input in the ECU.
I made a cable to connect the Bosch sensor to the Chassis Loom
The sensor is fitted to the chassis just ahead of the fuel tank.
Traction Control is now set to use Yaw
I tried to connect the exhaust camshaft sensor directly to the ECU by re-assigning AN#16 which is the steering angle sensor. I modified the spare inputs loom, and took 5V, GND and the white wire (Pin 5) from the 6 pin connector, over to the sensor, using shielded cable, and everything checked out ok. Until I tried to program the ECU. It displays an error as below:
This error indicates that the cam sensor input needs a specific type of hardware to support the connection to a cam sensor. With the same pin assigned to the steering wheel angle sensor, it just programs with no errors. But try and use AN#16 for VVT and computer says no. I think is a hardware limitation in the ECU, since they were built with a reduced IO. So I'll reconnect AN#16 signal to the steering wheel sensor, and have a look at using one of the X10's spare inputs, which I'll have when I've moved the calibration switches to the Steering Wheel CAN Switch.
Lesson learnt: Before trying to repurpose any input, make sure the ECU will allow you to reassign it first.
Tomorrows job is to wire in the Bosch Yaw rate sensor.
I soldered my CAN split connector on to the end of the CAN bus having first removed the redundant 12-pin Deutsch connector from the end, and I then connected the Life GPS board to the CAN using the new loom. The 120 Ohm terminator resistor is built in to the GPS board, so it still needs to go on the end of the CAN-BUS to terminate any reflections. I powered the ignition on, and I was still able to read GPS and accel data from the board.
So I then connected the ECUMaster steering wheel CAN Switch board to my CAN split connector, and powered the ignition on again, and the tiny green LED on the board started blinking rapidly, which shows that it was also working. Next job was to program the Datastream in the ECU to enable frames A13-A20, and make sure the address of A13-A20 were set to 640-641-642h. I then went to the IO Config and under Pin Assignments assigned X:CAN RECEIVE A #09 to Inputs R09, programmed the ECU, and then I added a gauge to display R09, and the voltage from the rotary switch connected to Analogue Input #1 on the switch board appeared. Bingo.
Now I can add the remaining three rotary switches to the Switch board, and perform the test again to make sure the ECU can read all four.
Whilst I had the CAL open, I lowered the engine oil temperature trip temperature from 200°C to 150°C which is still high, but can you imagine the engine oil temperature at 200°C ?!?!
The Binder connectors arrived so I made haste and fitted a green male connector to the CAN Switch board, a blue male to the Life GPS board, and then I made a loom split connector, to allow the two devices to be attached to the CAN connection on the loom. Tomorrow I'll remove the 12 pin Deutsch connector from the loom, and solder on my split connector. The Binder 720 connectors have five pins, and I've designated pin 1 = 0V (Black wire), pin 2 = 12V (Red wire), pin 4 CAN Hi (Blue wire), pin 5 CAN Lo (Green wire). I've buzzed all the wires through and they all test out OK.
By removing the GPS board from the plastic case, I've saved almost 180grammes in weight.
The ECU Master CAN switch board and wiring (below) will live on the steering wheel. The four rotary switches and the two paddle switches will all be wired to the switch board, and thats the benefit of using the CAN bus.
Frustrating delays whilst I wait for parts to arrive. I had ordered 3mm aluminium dome head rivets but the supplier sent countersunk ones. When the right ones arrive I can finish off the rear diffuser. I'm waiting for four Binder 720 5-pin connectors to arrive, so I can wire in the CAN Switch board to the CAN-BUS. I've removed the Life GPS board from its large plastic housing, and I'm also changing the GPS connector from the 12 pin Deutsch type to another 5-pin Binder 720, which will again save further weight. The plastic casing and connector weighs around 200 grammes in total.
The Grayhill 10 and 12 position rotary switches arrived from RSWWW and CPC, and I've built the 10 position switch to allow me to fit it on the CAN board so I can test the operation of the board out. Once its proven, I'll build the remaining switches with resistor ladders and wire them to the CAN switch board, test them, and then I can test the paddle switch operation. Once its all working, I'll fit everything to the steering wheel and print the stickers.
With the tufnol side skirts now fitted, I turned my attention to the steering wheel. I'm moving three calibration switches to the wheel, and fitting an ecumaster CAN Switch board, which should allow the steering wheel to attach to the ECU with a simple 4 pin connector. The wheel already has the paddle switches on it, which are connected to the ECU using a connector. These switches will be wired to the CAN Switch board, and three rotary switches will also be wired to the boards analogue inputs.
All I need to provide the board with is 0V, 12V, and CAN Hi and Lo. Once the ECU is programmed to receive data from the board, I can re-assign the pins used to provide the switch positions for CAL, Launch RPM, and Traction Control.
Once thats done, I can then use the spare analogue inputs on the X10 expander unit, which were connected to the rotary switches, as suspension pot inputs, so the suspension movement can be logged by the ECU.
First job was to replace the metal plate on the wheel (80 grammes) with a carbon plate (25 grammes) which gives me an area to mount the switches. I like working with carbon, cutting it can cause a lot of dust, but if you spray water on it whilst cutting, or drilling, you can greatly reduce the mess.
Now the plate is made, I looked at making some artwork for the switches. I've designed several different styles, all using Powerpoint, and I can print these on to the adhesive backed A4 label paper using my colour printer, and then carefully cut them out and stick them on to the carbon.
I've put a video together explaining how to make the ECU log data whilst the engine is stopped. I'd managed to set this mode last month when I was testing the wheel speed sensors and logging the data from them, but when I tried to make it work at the weekend, the ECU was logging a new file every 1/10th of a second. The video explains the sequence ;)
Almost inevitably, Borough 19 came to the decision to cancel their Snetterton Sprint and Lydden Sprint in July. Which is a shame, but, the welfare of those involved is the most important aspect at the moment, and the club couldn't see how they could manage with the burden of guidelines placed upon them by MUK. And then yesterday the Midland Speed Championship also announced the cancellation of their championship, with registration fees being rolled over to 2021.
An email came out yesterday from the organisers, which doesnt really say a lot. As we've lost Snetterton and Lydden in July, they've suggested that the Snetterton round is rescheduled to another date. I know Borough 19 are meeting tonight to discuss whats going to happen to the event following the MUK Email on Tuesday saying there would be no championship rounds in July. They could continue to run it as a club meeting, rather than a national, so it could still go ahead, but there'd be no points for BSC drivers. And without a maximum entry, they're going to lose money, since Snetterton isnt cheap to hire for a weekend. The BSC is now hedging its bets on N.Ireland running at the start of August. But N.Ireland might not happen since there are travel restrictions in Scotland, and we still dont know whats happening regarding overnight camping. Besides I cant afford to go to N.Ireland. I'm without any income at the moment, since my contract finished on April 3rd, so I simply cannot justify the costs of travel and competition for a few points. If Knockhill is moved to September, and N.Ireland is cancelled, we end up with 6 rounds (Knockhill x2, Aintree, 3 Sisters, Anglesey x2) with the best 5 rounds counting.
The governing body for motorsport in the UK sent an email out yesterday, stating that there would be no National or British Championship rounds taking place in July. That now means that the BSC rounds at Snetterton and Lydden in July will no longer be taking place. The Borough 19 Motor Club were still planning on running both events when I spoke to them on Tuesday evening, however, since they are no longer able to have the Sprint Leaders, BSC, or any other invited championship attend, I'm not sure how many people will enter, so the events might get scrubbed altogether. The next rounds of the BSC are in N.Ireland at the start of August, and Knockhill at the end. I doubt restrictions will be lifted in time for the event at Kirkistown, and the rumour is that Knockhill might be moved to September. Still no news on the championship itself. No communications. No emails. No phone calls. So we still dont know what the coordinators intentions are for abandoning or pressing on with the championship this year.
I've sprayed and fitted the Tufnol strips. I used matt black spray, probably should have used gloss.
I've fitted a set of 50mm x 2mm Tufnol strips to the carbon floor, which sit 50mm above off the ground. They will be pop riveted to the floor once I've sprayed them black. I would have sprayed them over the weekend, but when I got to Halfords on Saturday, the queue of customers outside must have been an hour long, so I went home and ordered the paint on-line. It could have been a line of people wanting to buy bikes in the queue, they'd have been in the store for an hour LOL. Anyway the Tufnol is very light, hard wearing and I fully expect to replace sections as they touch the ground and wear.
While researching barge boards, I read a number of white papers on race car aerodynamics. This picture taken from one report, illustrates the wake without and with the boards fitted, and I'm hoping the Dallara boards I've fitted will divert the wake away from the rear wing, increasing downforce. Reference: Honda R&D Technical Review 2009, Aerodynamics Analysis of Formula One Vehicles
I updated the firmware in the Dash4pro display yesterday, and I made a couple of slight changes to the two screens that I use. I've added the lift pump fuel pressure (fp1) to both screens. I also updated the shift lights, so they show green at 6000-6150-6300-6450rpm, red at 6600 and 6750, and flash red at 6800. The warning buzzer lets me know when I've reached 6400 rpm in every gear.
Small tweaks to the layout to display the fuel pressure from the lift pump (fp1) during tickover.
I've removed the Life GPS PCB from inside the plastic casing, and it is now affixed directly to the chassis on top of insulating blocks. The plastic case is too big and bulky to fit in the space behind the front dampers, and despite cable tying it down, it would move and that would affect the offsets for latG and longG. As I discovered last week, when latG is greater than 0.1G the logic that produces the front wheel speeds, assumes that the car is turning, and ignores the inside wheel sensor as it assumes it is unloaded. So at rest, if latG wasnt zero, the speed sensors wouldnt report independantly the true wheel speeds. By removing the PCB from the case, the footprint of the GPS board is much reduced, and lighter, and easier to attach to the chassis. I've saved 140 grammes by removing the case. I've got a Deutsch 12 pin connector on order from RS, and I will remove the remainder of the plastic housing that surrounds the connector left on the PCB which will save more weight.
Removal of the PCB is easy enough, its just held in by a couple of plastic clips and slides out.
The PCB is attached directly to the chassis using cable ties, mounted on some firm insulation blocks to prevent electrical short circuits.
The front wheel trigger disks have been swapped for some heavier gauge ones, again with 20 slots. The ones I found on ebay were paper thin, and the sensors didnt appear to work very well using them. The new disks are from B&Q and were £5 each, but they're better quaility. This means I dont need to change the sensor at £60. I'm now reading the speed every 4th tooth, which averages the reading, and I'll drop this to every 2nd tooth when I get a chance to drive the car. I've now secured the tops of the carbon aero covers with M6 fasteners, to eliminate any chances of the covers from moving. If the covers move, then the gap between the sensor and the trigger disks would increase, which would prevent the sensors from working. I've had a few failures on the front left, which might have been caused by this. The ECU is set to fail a sensor after a set number of spikes, so when the ECU is power cycled or a length of time passes, the sensor would be re-enabled. So its not a permanent failure, but one that the ECU determined as a possible fault, hence the sensor is turned off.
I've fitted the Dallara barge boards that I bought last year. They needed trimming and modifying, and I had to make a pair of aluminium brackets where they join the bodywork. Overall I'm very pleased with their appearance. Lets hope they make a difference to the wake from the front wheels, and remove the turbulance from spoiling the flow over the rear wing.
I've setup the ECU to log data without needing to have the engine running. I've always lacked data for testing, and the ECU was set to start logging when the engine speed was above 300rpm, and stop when it was below 100rpm. This makes diagnosing problems very difficult, since you dont want a noisey engine running when you're looking for problems on the chassis sensors. With the change made, and the ECU powered on, I was able to download speed data from the front wheel speed sensor, which can be seen in the picture below.
With the front right sensor plugged in to the front left connector on the chassis loom, the noise is still present, so I'll swap the front right sensor out for a spare and see if it cures the noise.
The front right sensor plugged in to the front right speed sensor connector, and then the front left speed sensor connector.
This is the noise I was talking about on the frspeed channel. The front right wheel hub is being turned using the electric drill, at around 120mph and the speed jumps rapidly to around 127mph. I'll swap the left and right sensors over on the loom and see if the front right sensor registers the same noise when connected to the front right channel on the ECU.
The sink drainers have arrived, so I fitted them to each hub, and positioned the sensor to allow it to capture movement, and the results arent bad. Spinning the front wheels at 110mph using the Bosch power drill with the wheel nut attachement, the front left sensor works perfectly, and the front right is a little noisey, with 10mph spikes. That will need investigating.
Both the Mygale front hubs have been removed from the car, and given a good clean on the bench. I've refitted them both, and replaced both the front pushrods with the freshly sprayed ones which have flats on the ends to assist with adjustments, and I'm now waiting for the replacement front wheel speed trigger disks to arrive in the post. The trigger disks on the front axle have just 8 holes drilled in them and I am looking for around 20 pulses per revolution. I made a 90mm diameter disk and cut out the slots in it to trigger the sensors, but it didnt work very well. So I had a good look on the internet for something I could maybe repurpose to do the job, when I came across a sink drainer plug. Yep, a plug, made from stainless steel, that is the right shape and size, and has 20 slots cut out in it. Its even dished to go over the nut in the centre of the hub. Perfect I hope. I tried the plug from my kitchen sink, against the inductive sensor, and it triggers the sensor alright. So I've ordered a pair of drainer plugs, for £3.95 inc postage, and when they arrive I'll remove the post from the centre, and screw them to the hubs and test them out.
The front hubs, cleaned and refitted
Sink drainer plugs, which will work as trigger disks for the front wheel speed sensors