Monday, December 16, 2013

Hall-effect throttle replacement

As I understand it, there are 9 of us Gizmo owners who have the Hall-effect throttles which go along with the Sevcon PP745 SepEx controllers. Hall-effect throttles are essentially a non-wear item. The only thing that should ware on them is the pivot point of the trigger and the return spring. The proximity of a magnet to the sensor changes the output voltage of the signal line which the controller then interprets as a torque value. A really great setup, actually.

Mine has been working flawlessly until about a month ago. The throttle started cutting out at various times. Sometimes this was when I was attempting to accelerate from a stop, sometimes it was when my variable regen cutout (I posted about this in March/April 2009), and sometimes it was while I was driving along at a constant throttle. Once in a while it would quit when I hit a bump or slapped the side of the handle. I took the handle apart and took the Hall-effect sensor off and couldn't see anything immediately wrong with it except that the back of the case was dented where the screw mounts touched it.
 If you look closely at the edge of the sensor you will see what looks like glue. I noticed that when the trigger is pulled fully it doesn't stop until it impacts the sensor and the side of the sensor pushes in just a little. I reassembled the handle and then started looking for a solution. I contacted Ron Anderson of Black Sheep Technology, the builder of the replacement "NORM" circuit. (BTW, Carl Watkins told me that the name of the one who built the original circuit was named Norm so that is what the circuit was called.) Ron told me that the replacement circuit was a pass-through circuit meaning that it only took a measurement of the throttle voltage to do its thing but that the signal was a direct line to the controller so it was unlikely that the replacement Gizmo Interface was the problem. In all of Ron's wisdom he built in a serial display output to the interface. As it turns out, I had run extra wires from around the tub to the dash and I had used some of these wires to hook up the Gizmo Interface display. This turned out very useful because I could then watch the throttle signal on the dash as I was driving to try to figure out what was going on and when the throttle cut out.

Prior to removing and reinstalling the Hall-effect sensor the cutout issues were somewhat random but not quite. Afterword, however, it seemed that the throttle would cut out when the trigger was pulled fully on. This seemed strange but I figured I must have wiggled something inside the unit in the removal/installation process. I clearly needed a replacement.

Carl said that they purchased the handles as complete units so I went to my list of bookmarks and the original parts supply spreadsheet and found a website for Sure Grip Controls, Inc. I emailed the sales department and told them I need the hall-effect part. They replied that they only sell through dealers and that Pape' Machinery right here in Kelso, WA was a dealer. I stopped by the next day and naturally the Gizmo was a topic of conversation for a bit. The sales/order rep was very helpful and said that he had an order for Sure Grip going out the next day so I could get the part without added shipping. Awesome! Found that the part only comes as a kit and that the last item he ordered with it in it was about $125. Ouch. I don't have a choice and knew that even if it was over $100 that it would be cheaper than having someone build one from scratch so I had him order me the kit. It is part number PC-V2-T-KIT. The Whole kit is shown on the left and a close-up of the individual parts is on the right.
Note that the Hall-effect sensor is fully potted with a hard resin. This might have been done for cost savings but it looks like there won't be any more flex in the packaging like on the original part. I was going to use the original trigger since the replacement is a gray plastic but usually these sensors are tuned to the magnet in the trigger so I decided I would use the new trigger. It turns out that I needed to use the new spring too since the design of the trigger has changed slightly.
Next was to hook up the new sensor. The replacement comes with pig tails rather than solder posts. There isn't much extra room in the handle so I decided to just solder the new wires to the old ones and heat shrink the joint. The problem is that it was more than a two handed job so I decided to cut a high quality copper but connector in half and push the wires in from both sides and crimp them. This is actually better in an automotive environment any way because there isn't any solder to wick up the wire and cause a crack point in the wire. I was delighted to find that Carl and crew used the basically the same color wires as came with the replacement sensor. I got everything mounted up and put the handle back together.

I checked the readouts on the Gizmo Interface and other than the "zero" point being slightly higher it worked as needed. The Gizmo Interface already had a buffer programmed in so that "zero" was a little above the low voltage of the original hall-effect and it was enough to accommodate the slight difference.

Now that I had a working throttle I was a little more willing to take apart the old sensor. I carefully pried the FRP cover off and then slowly pried the contents out of the gray plastic box. Then I removed the glue (which is like silicon rubber) off the back and then the front. What I discovered was that the actual hall-effect sensor it self had a broken solder joint on one of the pins.This may have been a weak point to begin with or it could have just broken after the sensor moved slightly every time full throttle was requested. One of the mounting screws is also right beside the connection so when I reinstalled the unit after my initial inspection I probably tightened the screw a different amount and with the compressible potting material the joint might have made better contact, at least to the point that only when full throttle was requested did the connection break. I re-bent the pin and re-soldered it. I also did my best to reattach a very tiny SMD resistor I accidentally cut off when prying the edge of the potting compound off. I hooked up the circuit board to my power supply, gave it 5V and measured the output. It puts out the original 0.82V to something over 2-3V. I didn't try too hard to get the trigger magnet in the right place since I don't expect to need this any more but I'm satisfied that it works somewhat.
How does the new throttle work? Great! The edges of the trigger are more rounded so shouldn't dig into my finger as much. The spring tension feels just slightly softer and response is definitely smoother. Oh, and the kit ended up costing $84.00 and I picked up a replacement momentary switch for one of my blinker switches since it sometimes doesn't activate. The switch is part number SW-00 and was $5.00. In the process I found out that Sure Grip Controls handles are used in many places around the world and on most equipment so parts shouldn't be a problem. The only thing is that the handles are gray instead of black and the button faces are larger in diameter than they used to be.