One thing leads to another

Lots of grease

During a recent commute, a thumping sound was audible between about 10 mph and 20 mph.  The sounds was speed dependent and seem to be correlated to wheel rotation.

Looking underneath, two things were seen - A torn outer CV boot (driver side) and a torn steering rack boot (driver side).  Also, the oil drip from the main shaft seal seemed to be getting worse, so it was time for a major enhancement session (3 weeks on the jack-stands).  Here is a summary of the latest updates.

• New Rear Main Seal (LUF10005).  Previously, the bolt holding the primary gear to the shaft adapter for the motor would not easily slide through the seal and onto the motor shaft.  Some light grinding and thinning of the bolt head resulted in it not tearing the seal when attaching it to the shaft.  The old seal had an obvious gash.

• Idler gear float - confirmed the idler gear float is 0.004 inches (right in spec).  Verified the torque on the bolts holding the flywheel housing/transfer gear cover.  The gears all turn very easily by hand when in neutral.  This was part of an effort to reduce gear noise.
• New gasket for gear-box cover.  Removed the RTV "make-a-gasket" material and installed a custom gasket, making it easier to service in the future.
• CV Boot replacement (GDG0233) - This required disassembling the driver side wheel hub - 2 ball joints.  Then the drive shaft could be removed, repacked both CV joints and reassembled.  Replaced ball-joint dust covers (BTA0377) since they were destroyed in the disassembly process.  Needed to apply some crowbar pressure to get the stud to seat into the knuckle so that the nut could be tightened.  The old boot had a plastic zip-tie on the large end, sitting inside the wheel hub, and it may have snagged on the hug causing the boot to tear.    Used the metal band that came with the new boot.   It has a much smaller head.
• Steering Boot replacement (BHM7113) - Removed the tie-rod end ball joint.  The ball-joint seal was not installed with the upper clip so the joint was frozen.  This is a non-greaseable joint.  After some freeze-all and lots of wiggling of the stud, the joint freed up. Then a good application of silicon grease and standard grease had the joint twisting very smoothly.  The cap was installed with zip-ties on both the top and bottom to prevent water incursion.    Greased up the steering rack as best possible and installed the new boot.  It is hard to tell whether the steering is easier now but this certainly did not hurt.
• Brake cylinder rebuild (BHM7068) - one of the brake cylinders had a significant leak so a quick replacement of the inner and outer seals resolved it.  Also, replaced the broken water ring for the inner front hub seal (GHS0173).

Electrical Updates

As an experiment, I inserted a spare CA60FI LiFePO4 battery in series with the existing 26 cells.  This raised the resting pack voltage to approximately 87V.  The Kelly controller shuts off at 90V, so there is not much margin but the pack should never exceed this during driving.  Charging voltage goes higher, but the Kelly controller is not operating during charging.  Regenerative braking does not produce enough boost to raise the pack voltage above about 88V.

This resolved:

• The DC-DC converter would shut down if the Pack voltage dropped below about 72V.  During heavy acceleration (300A-400A), the pack could drop to this level, especially, if it was below 50% charge, resulting in the 12V system dropping down and relying on the small starter battery.  Lights would dim but nothing else noticeable, just annoying.  The extra 3.2 volts made the difference and now the DC-DC converter never drops out.

Now however, the 27 cells means two things - a new BMS since the current (Green Bike Kit GBK) unit is customized to 26 cells, and a new mechanical arrangement.  The 13 x 2 rack will have to be redone to a 9 x 3 setup, if that will fit in the boot.  This is a good reason to re-engineer the battery setup anyway.  Just need to order a new BMS.

Second benefit - About a 4% increase in battery capacity - not a lot, but every bit helps.

New Digital Dashboard

New Dashboard

It was time to upgrade the Digital Dashboard.  Here is a quick summary and a few fixes

1) New DUE Arduino controller

2) New Serial CANBUS from Longan Labs  http://docs.longan-labs.cc/can_bus/ 

3) Sainsmart 7 inch TFT LCD with CTE Shield (uses SSD1963 controller)

4) Using the UTFT and URTouch library for the display

Setup

1) UTFT myGLCD(20, 25, 26, 27, 28);  //Setup for display  This is CTE70=20

2) URTouch  myTouch( 6, 5, 32, 3, 2);  // the 32 was a change from the default 4

     - Touch Calibration worked with #define TOUCH_ORIENTATION  LANDSCAPE

3) Serial CANBUS connected to Serial 3 on the DUE  (tx-rx rx-tx)

4) 12V DC supply to the DUE (see below for 3.3v fix)

Fixes

• TFT display would not consistently initialize after being powered off for longer than a few seconds

1.  Added a 9K resistor as shown in http://www.dimdim.gr/2016/03/fixing-startup-issues-with-arduino-dues/
2.  Used the latest UTFT library and un-commented out all of the unused configs in the memorysaver.h - 
3. Removed the 3.3v supply from the DUE to the TFT Shield by bending header pin outward.  This solved the cold-start startup problem.  It seems like the 3.3v suppliers (one on the TFT shield and the one on the DUE may have been fighting each-other.

Now the system consistently boots in less than 1 second.

•  Serial Can Bus Library - Since Serial Can Arduino Master relied on SoftSerial, a global replace of "canSerial"-> to "Serial3."

•  Add some buffer clearing since 1939 CAN Data is on the bus starting at power-up.  This floods the serial3 port on the DUE making it difficult to perform the setup.  Sending "+++" to the Serial CANBUS board, sets it to Command Mode, but if the a baud rate is unknown, then the "+++" must be sent at each of the 5 possible baud rates.   After sending the +++; flush the receive buffer then return to Data mode.

• Change the Serial Can Bus factorySetting() routine to set the serial baud rate to 115200 and to start searching for the current baud rate at 115200 since that is most likely where the board is set to at power up.  The Serial CanBus board stores all settings in nonvolatile memory.   Also set all 6 filters to EXT 0 and the two masks to EXT 0.  EXT is for extended frames used in J1939 CAN Bus 2.0b broadcast protocol.  The 0x00 filter and masks allow all frames to pass through.  When set right the Receive LED on the Serial Can Bus board will flicker dimly with traffic.  If set wrong, it will either be bright or off.  Also, the Green power LED flickers with serial traffic to the DUE.