First, some other quick updates:
For the 4-pole motor, I had a square wheel mount for it's axle, which was just too large at it's corners to allow small chainrings to have clearance at those corners. So I rounded it off by first grinding off the corners, then lathing it round. Then I tackwelded a 21-tooth to it, as that had worked well enough (for the short time it lasted bolted on, back on that May First Friday ride) on the 2pole motor, without causing me to have to windmill my feet to keep up with it at full throttle. If it works out I'll fully weld it on.
I'd wanted to bolt it on, but I need a good drill press for precision holes, and I can't do that with anything I have here. I tried to rig one up using an electric handdrill and a fixture but the handdrills' bearings are all worn too much and wiggle around, on every one I have. Not a big deal if you're hand drilling it anyway, but terrible if you need precision.
Now, before putting the 2pole motor on there again after fixing the controller, I did a quick test fit of the 4pole motor, and it would not even fit in there, because with the derailer hanger where it is, the motor has to be so far over to the right that the front of it (which is wider and longer) is in the way of the crank arm of the right pedal!
That made me sad, but was just one more thing goading me on to fixing the drivetrain problems the rest of this post is about. :-)
So today I decided to do a major revamping of the drivetrain, since I wanted to re-gear it to get my low gears back for hill climbing (even steep driveways are difficult the way it is now, much less actual hills, if I'm not already going fast enough to let inertia help). I also wanted to eliminate that derailer-tensioner on the bottom of the pedal chain, since it's REALLY noisy, inefficient, and in the way.
First, I decided to add the chain-tension-throttle arm so I can test out that control once I have the electronics built. In theory the hall-effect-based sensor is simple, but in reality it requires offsetting it's output voltage (it doesn't start at zero, it starts at half it's supply voltage and goes up or down from there depending on which direction the magnetic field is pointing, and how strong it is). It also requires scaling that voltage, because the 2QD needs about a 4-volt-range input, from zero to full throttle.
Above is the left-side view of the throttle arm/chain tensioner, idea borrowed from AussieJester at the FreakBikeNation forum:
Mine simply uses the actual plastic derailer wheel and roller sleeve, with the ex-brake-cable-holding bolt clamping that to the brake arm. Of course, all he was using it for was to tension his motor chain so it wouldn't jump off, and I'm using it for my pedal chain *and* my throttle control, assuming that my idea works out. ;-)
The 2QD does have a way to scale the voltage built-in, but I want to make a little throttle adapter box that can convert voltage from any hall throttle to a resistance for controllers that check for open potentiometers and thus can't be used with hall-throttles without internal modifications. That way, anyone that wants to can build this adapter and not have to do controller modifications that would probably void their warranty.
Above you can see it viewed from above/behind.
Now, it's only tacked in position and clamped in place with a radiator hose clamp, so I can change it's position if I need to, based on how well the throttle application works. Oh, and I cheated, so that I could clamp it temporarily, I just cut the stud along with a 3" piece of tube off the donor BMX bike, instead of trying to tack the stud on by itself.
I'll put up the circuit diagrams once I have it working, but it basically only takes a quad-op-amp chip (like the LM324), two MAX5160 (or similar) digital pots, and some resistors and a couple of small potentiometers for adjusting the offset and gain of the op-amp circuits to suit the throttle/controller application. It uses the voltage-to-resistance idea in Maxim's AN3284.
The tensioner viewed from directly above, front of bike to the left. Most likely the magnet will be bolted on an extension from the slot where the brake pad used to go, so that it sticks up vertically past the bike's frame. The magnet could be mounted so that it just has one pole within reach of the sensor to keep it compact, and then the sensor only reads either from midpoint to zero or midpoint to max. If it's mounted so that north is above the sensor and south below, such that middle throttle is at zero field in the middle of the magnet, it'll get much more useful range out of the sensor. Either way, the sensor will be mounted on the frame in line with the magnet's arc.
Alternately, the magnet will be mounted north-up/south-down at the back of the lever, just at the right edge of the pic, and the sensor on a bracket from the frame so it sticks out behind the arm. Then as the magnet goes past in a small arc, the sensor picks up both north *and* south plus the changing strength, to get the most range out of the sensor. Depending on the magnet I end up using, I will probably go with this method as it requires the least work to do and has the least stuff sticking out to get bent/etc.
I managed to redo the pedal chain using the chainrings I had on the right side moved to the left side, using the big 52 tooth ring for the pedals' receiver ring (leaving the 40 tooth open and unused), which somehow corrected the chain length so I didn't have the half-link of extra length, which is what forced me to use that derailer-tensioner in the first place.
Now it is just a clean loop of chain, with just a little play in it. Not even enough to be able to pop it off the chainrings, but enough to allow my tension-based throttle to operate. Much quieter and more efficient; easier to pedal, too!
Since I had to change the chainrings around to get my lower gears back, I now had no place to run the motor chain to, unless I put a huge chainring on the motor (which there isnt' really room for).
This shows an idea to put a 24-tooth ring on the *outside* of the 2-ring set that's now on the left side. There's a 28-tooth on the motor right now; the largest I can fit on there without other major modifications that would leave the motor mounting plate too weak.
As it is, I'd already had to cut a notch in the plate to clear the chain and ring for this one:
Only have about another 3/4" or so before I would be hitting the frame, so I can't use a much larger chainring than that. I wouldn't even have done this much, except I needed that 28-tooth ring to let me experiment with using the motor with a significantly smaller ring than the drivetrain input side (at 34 teeth) to see what effect it would have on startup torque vs speed.
The effect was pretty bad, as I didn't really get any more torque out of it, and had far less speed than I wanted--barely more than I could just pedal to. Of course, pedalling to that speed was *much* easier, but I couldn't sustain more than maybe 12MPH. It's hard to even keep it balanced below 9MPH with the batteries on there, so I'd rather have it around 15MPH, and had been able to do 16-17MPH easy for long stretches when I had it as 1:1 (two 24-tooth rings).
So that's why I want to put the 24-tooth ring on the outside of the 2-ring set. There's no room to put it on the inside, since it is too close to the frame for a chain to fit there if I did. I will find 5 identical spacers I can weld to the 2-ring set to ensure clearance of the chain to the outer ring and the bolt-heads that hold the other two together, then weld this ring to the spacers.
I'd love to bolt it on, but I can't drill precisely-enough-placed holes to guarantee it'd stay lined up under tension, or that it wouldn't just shear thru the bolts eventually. So I'll weld it to be safe. I'm not entirely sure if it will clear what remains of the crank arm, but I don't want to cut the rest of that off until I must. I might someday have to use that crank arm and rings for a different bike (although I'd need to drill and tap the hole for the pedal, since the end with that part was damaged so I had removed it), and I have only a single pair of square-taper cranks with rings, both of which are on this bike right now.
I keep hoping to get more from donor bikes, but I have not been able to get any for several months now--alleys and trash piles are being swept clean by salvagers long before I ever get to them, so most of the useful junk I used to use is no longer available to me. I haven't found any bikes on Freecycle either. I tried several times to bid for some of the police auction bikes, but they *start* at $10, which is about my limit for what I could afford to pay for one (right now that is a few days' worth of food!). So I lost all the auctions I bid on, even for the bent-up bikes.
Here you can see how close the two chains are, but they don't touch.
Ignore the extra links--it's just because I am using different size chainrings than I was when I made that chain up day before yesterday. It'll change by a few more links once I put the 24-tooth sprocket back on that 2-pole motor or use the one on the 4-pole motor.
This shows all three together, from the top, front of bike to the right.
"top" is the motor chain, "bottom" is the pedal chain, with the tensioner to the farthest right.
This is the whole drivetrain as it exists right now:
Just in front of the seat you can see the new controller in it's over-sized weatherproof box, heatsinks on top. It's not mounted there, just out of the way for pics. It goes in the triangle with the tensioner/etc. The battery for the lights that was there is now in the rear cargo pod.
I *still* haven't been able to make hangers for the battery boxes that will support the weight, not bend, and not jounce around so much that it makes the bike wobble when I hit bumps. The hangers have to be made of (thick) steel wire, to fit the mounting points on the battery boxes. All I have that will work is old retail-store peghook hangers--those steel rods that stick out from the pegboard that hold product for display. They're good steel, but too springy for this application.
I am probably going to have to make panels I can weld (preferably bolt) to the bike's cargopod rails, then bolt the boxes to those, with a metal panel inside the boxes taking the clamping strain. It just is going to be harder on the plastic boxes that way, since they were designed to support the weight from the top edges, hanging from something, instead of from one of the side walls. It will also be a lot harder to get the batteries in and out of them. If I had the wire hangers done right, I could just lift the whole box off the hangers and unplug the power leads, and then I'd have an instantly lighter bike for those times I want to do that, or when I need access to what's behind them (anything under the seat).
If I bolt them to panels, then I have to unbolt them to get the batteries out (or take off the seat), and then unbolt the panels to get to stuff right behind them if necessary. Lot of work I'd rather not have to do. Also, the wire hangers would be lighter.
Friday, August 14, 2009
First, some other quick updates: