Bonanza "Bulldozer" Dual PhaseRunner AWD
- Thread starter Alan B
- Start date
Alan B
100 GW
Do you know if the cable moves half as far or twice as far with the 2:1 stuff?
Here's the shifter I ordered, as I understand it:
So the question becomes, is the old SRAM derailleur a 1:1, which it may not be, so I might have to get a new derailleur as well.
The original shifters were SRAM 5.0 Twist, 9 speed. The rear derailleur is still the SRAM 5.0 9 speed unit, as is the chain. The present rear freewheel is a 7 speed screw-on unit on the 9C motor.
It will be an adventure, and an experiment. With lots of torque from 2 motors, but no suspension except the Thudbuster, it won't be a speed daemon, but it should go most anywhere you point it. Perhaps the biggest problem will be in extremely steep conditions there may be insufficient weight on the front wheel to use its torque, and the rear motor doesn't have enough torque by itself, so that may determine the limiting conditions.
Here's the shifter I ordered, as I understand it:
So the question becomes, is the old SRAM derailleur a 1:1, which it may not be, so I might have to get a new derailleur as well.
The original shifters were SRAM 5.0 Twist, 9 speed. The rear derailleur is still the SRAM 5.0 9 speed unit, as is the chain. The present rear freewheel is a 7 speed screw-on unit on the 9C motor.
It will be an adventure, and an experiment. With lots of torque from 2 motors, but no suspension except the Thudbuster, it won't be a speed daemon, but it should go most anywhere you point it. Perhaps the biggest problem will be in extremely steep conditions there may be insufficient weight on the front wheel to use its torque, and the rear motor doesn't have enough torque by itself, so that may determine the limiting conditions.
Alan B
100 GW
Looks like an SRAM X4 derailleur will work, is not expensive, and is an upgrade from the old 5.0 that's on this bike.
First off, the 2:1 and 1:1 ratio is marketing talk, not actual cable pull. Originally, Gripshift made twist shifters only. They didn't make derailleurs. Their twist shifters were made to work with Shimano derailleurs. They were originally used as bar-end shifters on roadbike drop bars.
http://stwww.bikemag.com/wp-content/uploads/2013/05/grip.jpg
They were much bigger in diameter. Then MTB's came along, and they made them for that market. They were smaller diameter to match flat bar grip diameter, which made them hard to turn, as the grip diameter is pretty small, and it only got worse as your hands got sweaty. They got bought up by a new company...SRAM, which also bought up an old-line chain maker, and some other component makers. With deeper pockets, and more components, SRAM started making derailleurs and twist shifters with more cable pull, to make the twist shifters easier to use. On any cable operated device, more cable pull is a good thing. Campagnolo derailleurs, from the 1990's onward, had more cable pull than Shimano. About the same time Shimano started making MTB V-brakes with more cable pull than traditional brakes. Anyway, the actual difference in cable pull is not double, but about 25-30% greater, as I recall. Definitely enough to screw up shifting, if the components are not matched.
http://stwww.bikemag.com/wp-content/uploads/2013/05/grip.jpg
They were much bigger in diameter. Then MTB's came along, and they made them for that market. They were smaller diameter to match flat bar grip diameter, which made them hard to turn, as the grip diameter is pretty small, and it only got worse as your hands got sweaty. They got bought up by a new company...SRAM, which also bought up an old-line chain maker, and some other component makers. With deeper pockets, and more components, SRAM started making derailleurs and twist shifters with more cable pull, to make the twist shifters easier to use. On any cable operated device, more cable pull is a good thing. Campagnolo derailleurs, from the 1990's onward, had more cable pull than Shimano. About the same time Shimano started making MTB V-brakes with more cable pull than traditional brakes. Anyway, the actual difference in cable pull is not double, but about 25-30% greater, as I recall. Definitely enough to screw up shifting, if the components are not matched.
Alan B
100 GW
Warren said:
Thanks for the history. This kind of marketing and nonstandardization makes it hard to configure bicycles with other than standard components, which is often not possible with ebikes. In this case I needed to move from a 9 speed system to a 7 speed setup. Interestingly enough I've not changed the chain or derailleurs and it has worked fine thus far. I should probably put a small spacer between the freewheel gear cluster and the hubmotor as shifting to first gear has been problematic (as I recall, it has been awhile since I rode this bike).
Just don't put enough of a spacer that the chain can fit between the hubmotor and the first sprocket....done that; sucks for several reasons. 
Alan B
100 GW
amberwolf said:Just don't put enough of a spacer that the chain can fit between the hubmotor and the first sprocket....done that; sucks for several reasons.
Yes, I can imagine that result.
Alan B
100 GW
PhaseRunner #2 has landed. They have made a number of improvements in this model.
I need to look for aluminum, I think I'll put both PhaseRunners in a "U" channel with a box in between to contain the small plugs and connectors. That will provide a major heatsink as well. Perhaps use a pair of "L" extrusions to make a custom "U" with.
I need to look for aluminum, I think I'll put both PhaseRunners in a "U" channel with a box in between to contain the small plugs and connectors. That will provide a major heatsink as well. Perhaps use a pair of "L" extrusions to make a custom "U" with.
Alan B
100 GW
It is always a challenge to fit the controls on a DIY ebike, trying to find the right mix of products to get everything "in".
I removed this picture from this posting due to problems with Google Photos, and a better arrangement is already embedded below. The order is the same and the components are the same but things needed to be rotated a bit. This original setup was done with the bike vertical and it is hard to get the angles right that way.
From left to right we have SRAM X.3 7 speed trigger shifter, Shimano XTR brake lever and Magura throttle. It will probably require some fine tuning when I ride this bike again, but it all fits. The Domino throttle didn't fit quite as well. Everything seems reachable, at least with my hands, but there are limits to how adjustable this is, things are nestled together closely. I might even file a little off the inside edges of the triggers where they touch the throttle body, but it doesn't appear to be necessary.
I've been having a lot of trouble with Google Photos. They often generate links for my photos that require Google login to work, so I can't tell they are bad. It is a real pain to embed these, and at the moment it doesn't seem to work for more than one photo per posting. I have to keep monitoring the postings from a different browser to insure they are visible to others. My apologies for the confusion it creates. A process that works to create a valid embed URL will fail randomly. They make it easy for facebook and twitter but very painful for other forums.
I removed this picture from this posting due to problems with Google Photos, and a better arrangement is already embedded below. The order is the same and the components are the same but things needed to be rotated a bit. This original setup was done with the bike vertical and it is hard to get the angles right that way.
From left to right we have SRAM X.3 7 speed trigger shifter, Shimano XTR brake lever and Magura throttle. It will probably require some fine tuning when I ride this bike again, but it all fits. The Domino throttle didn't fit quite as well. Everything seems reachable, at least with my hands, but there are limits to how adjustable this is, things are nestled together closely. I might even file a little off the inside edges of the triggers where they touch the throttle body, but it doesn't appear to be necessary.
I've been having a lot of trouble with Google Photos. They often generate links for my photos that require Google login to work, so I can't tell they are bad. It is a real pain to embed these, and at the moment it doesn't seem to work for more than one photo per posting. I have to keep monitoring the postings from a different browser to insure they are visible to others. My apologies for the confusion it creates. A process that works to create a valid embed URL will fail randomly. They make it easy for facebook and twitter but very painful for other forums.
Alan B
100 GW
Did some filing on the Surly forks to deepen the fit of the flatted axle today, was able to get the Ebikes.ca stainless "C" washers to seat behind the chrome-moly steel fork's lawyer lips on both sides, and install the torque arm on the non-wire side. To fit the second torque arm will require removing hall sensor connector pins and passing wires through the torque arm hole. I need to find my bag of NordLock washers before I do that.
One thing I really like about these torque arms is the resultant force is straight up into the dropout. In this left side view the wheel rotates counterclockwise and the resulting torque is clockwise. They are also very thick stainless steel to avoid cutting the soft axle steel, and very nicely made without being too bulky and heavy. From Ebikes.ca. At 1300W the BMC might not need torque arms with a Chrome Moly Steel fork, but I'm going to use one on each side, plus the "C" washers and NordLock wedge locking washers. Front wheel failures are ugly.
I need to take a better photo, but the "C" washers sit behind the so called "lawyer lips" on the fork dropouts insuring that the axle cannot come out of the dropouts until the nuts are loose and backed out some distance. Combined with the torque arms it will be very unlikely to ever fall out, and the lockwashers guarantee the nuts will stay tight.
The NordLock wedge lock washers create a condition where, to loosen the nut the tension goes up first, due to the ramped wedges between the two washers. So vibration can't loosen the nuts. To work properly the teeth on the outside surfaces of the washers must bite into a nonrotating surface (the torque arm) on one side and the nut on the other side. No other washers can be adjacent to the lockwashers or the locking effect could be nullified. Installing the washers properly with the teeth and wedge surfaces in the correct orientation is critical.
When you go to remove a wedge lock washer you can feel the tension increase until a ramp drops into the next one, after that the tension drops off. So you can feel the locking effect. The washers are harder than the nut or torque arm so the teeth bite in and don't deform, so the washer is re-usable unlike most locking devices. In the photo above a conventional split ring lockwasher is still present, those are not very effective in retaining tension if there is any vibration or motion present. The rocking motion of acceleration / regen is particularly difficult, here there will be no regen on this geared hubmotor but on the Borg's CroMotor there is both lots of acceleration and large regen forces, and after 2 or 3 tightening cycles they don't move. This is very important on a regen setup hubmotor.
One thing I really like about these torque arms is the resultant force is straight up into the dropout. In this left side view the wheel rotates counterclockwise and the resulting torque is clockwise. They are also very thick stainless steel to avoid cutting the soft axle steel, and very nicely made without being too bulky and heavy. From Ebikes.ca. At 1300W the BMC might not need torque arms with a Chrome Moly Steel fork, but I'm going to use one on each side, plus the "C" washers and NordLock wedge locking washers. Front wheel failures are ugly.
I need to take a better photo, but the "C" washers sit behind the so called "lawyer lips" on the fork dropouts insuring that the axle cannot come out of the dropouts until the nuts are loose and backed out some distance. Combined with the torque arms it will be very unlikely to ever fall out, and the lockwashers guarantee the nuts will stay tight.
The NordLock wedge lock washers create a condition where, to loosen the nut the tension goes up first, due to the ramped wedges between the two washers. So vibration can't loosen the nuts. To work properly the teeth on the outside surfaces of the washers must bite into a nonrotating surface (the torque arm) on one side and the nut on the other side. No other washers can be adjacent to the lockwashers or the locking effect could be nullified. Installing the washers properly with the teeth and wedge surfaces in the correct orientation is critical.
When you go to remove a wedge lock washer you can feel the tension increase until a ramp drops into the next one, after that the tension drops off. So you can feel the locking effect. The washers are harder than the nut or torque arm so the teeth bite in and don't deform, so the washer is re-usable unlike most locking devices. In the photo above a conventional split ring lockwasher is still present, those are not very effective in retaining tension if there is any vibration or motion present. The rocking motion of acceleration / regen is particularly difficult, here there will be no regen on this geared hubmotor but on the Borg's CroMotor there is both lots of acceleration and large regen forces, and after 2 or 3 tightening cycles they don't move. This is very important on a regen setup hubmotor.
Alan B
100 GW
Readjusted the right hand grip equipment today and found a good compromise for trigger shifters, brake and Magura throttle:
The gear cluster is a 14-34 Shimano Megarange with six gears, so the new derailleur was adjusted down to six speeds. It shifts smoothly, though there is a bit of friction on the back side of the downshift trigger so I will probably sand off a bit of plastic there. We'll road test before we decide on that.
The left shifter was damaged so a replacement will be here in a few days. After that is mounted some road testing will sort out the details.
The gear cluster is a 14-34 Shimano Megarange with six gears, so the new derailleur was adjusted down to six speeds. It shifts smoothly, though there is a bit of friction on the back side of the downshift trigger so I will probably sand off a bit of plastic there. We'll road test before we decide on that.
The left shifter was damaged so a replacement will be here in a few days. After that is mounted some road testing will sort out the details.
Alan B
100 GW
Here we have installed the wire side hardware over the cable, having removed and reinstalled the pins to get it past the connector.
From the inside out (left to right):
Steel Axle Shoulder (BMC)
Chrome Moly Steel Fork dropout (Surly, black)
Stainless Steel "C" washer (Ebikes.ca)
Thick Stainless Steel Torque Arm (Ebikes.ca)
NordLock Wedge LockWashers (McMaster-Carr)
Nut (standard BMC)
Alan B
100 GW
Another view, Google photos finally started working.
The NordLock Wedge LockWashers are clearly visible between the torque arm and the nut. The ramps in between the two washers cause the tension to increase when undoing the nut, small teeth dig into the nut and the torque arm. You feel it when undoing the nut, the tension rises and then lowers. So vibration and small motions from torque and regen won't loosen the nut. This has worked very well on the CroBorg which has a lot more torque and high regen, so it will work easily here with no regen and modest torque.
Alan B
100 GW
There are two pieces of software that interact with the PhaseRunner motor controllers. One is from the manufacturer, called Bacdoor; and the other is a simplified version from Ebikes.ca called Phase Runner Suite.
These programs communicate with the controllers using TTL serial, in fact the Satiator's serial adapter is well suited for this as it is protected against the voltage transients that can occur with high voltage devices like the Satiator and perhaps these controllers as well. They use the prolific chipset so one has the usual problems getting the drivers loaded and working. After preloading the drivers my interface worked fine, and I used Bacdoor v1.4 to load new firmware (really xml parameters) into my older controller in preparation for testing and configuration. During this operation I powered the units with a small convenient Luna 52V mighty mini pack and an XT90 to PowerPole adapter.
The two units I will be using on this bike are both BAC800 type PhaseRunners as pictured earlier, embedded in waterproof rigid plastic. The older one is from the first shipment, I plan to use that for the front BMC gearmotor. The newer one has an embedded programming jack and on/off pushbutton and I plan to use that for the rear 9C 6x10 motor.
I set them both up for 72V nominal, I'll start with 52V and work up to 66 or 72V nominal later on. I'll record some of the settings here for later reference. These values are selected for 18S 66V Lipo testing/operation:
72-75V max regen
60-63V min LVC (note this will have to be reduced for testing on 52V batteries)
Pole Pairs, 9C 26; BMC 80
Initial Kv 6
max front BMC battery current 20A, regen 0A (40A motor)
max rear 9C battery current 30A, regen 15A (50A motor)
sensored start, sensorless run
torque throttle
torque ramping increase 200mS on 9C, 400mS on BMC
remainder of settings are generally default values or will be discovered in configuration
I made adapters for the phase wires to connect between the PowerPoles that come on the PhaseRunners and the 4mm tubular blue shielded connectors that EbikesSF uses on the BMC. The 9C already has PowerPole phase connectors.
I changed the BMC hall plug to be compatible with the PhaseRunner 5 pin JST-SM connector.
The next step is to setup a throttle with the appropriate plug, and physically restrain the BMC forks to allow the wheel to rotate safely and not fall over with torque reaction, and set the system up for initial testing, calibration and configuration.
Note that the serial communication protocol for the PhaseRunner is via Modbus over TTL serial. This is a common protocol and fairly simple, so we will be able to interface to these controllers from our own microprocessors, such as an Arduino. With this kind of interface there are many possible uses, far more than what the usual three speed switch can do. More on that later.
These programs communicate with the controllers using TTL serial, in fact the Satiator's serial adapter is well suited for this as it is protected against the voltage transients that can occur with high voltage devices like the Satiator and perhaps these controllers as well. They use the prolific chipset so one has the usual problems getting the drivers loaded and working. After preloading the drivers my interface worked fine, and I used Bacdoor v1.4 to load new firmware (really xml parameters) into my older controller in preparation for testing and configuration. During this operation I powered the units with a small convenient Luna 52V mighty mini pack and an XT90 to PowerPole adapter.
The two units I will be using on this bike are both BAC800 type PhaseRunners as pictured earlier, embedded in waterproof rigid plastic. The older one is from the first shipment, I plan to use that for the front BMC gearmotor. The newer one has an embedded programming jack and on/off pushbutton and I plan to use that for the rear 9C 6x10 motor.
I set them both up for 72V nominal, I'll start with 52V and work up to 66 or 72V nominal later on. I'll record some of the settings here for later reference. These values are selected for 18S 66V Lipo testing/operation:
72-75V max regen
60-63V min LVC (note this will have to be reduced for testing on 52V batteries)
Pole Pairs, 9C 26; BMC 80
Initial Kv 6
max front BMC battery current 20A, regen 0A (40A motor)
max rear 9C battery current 30A, regen 15A (50A motor)
sensored start, sensorless run
torque throttle
torque ramping increase 200mS on 9C, 400mS on BMC
remainder of settings are generally default values or will be discovered in configuration
I made adapters for the phase wires to connect between the PowerPoles that come on the PhaseRunners and the 4mm tubular blue shielded connectors that EbikesSF uses on the BMC. The 9C already has PowerPole phase connectors.
I changed the BMC hall plug to be compatible with the PhaseRunner 5 pin JST-SM connector.
The next step is to setup a throttle with the appropriate plug, and physically restrain the BMC forks to allow the wheel to rotate safely and not fall over with torque reaction, and set the system up for initial testing, calibration and configuration.
Note that the serial communication protocol for the PhaseRunner is via Modbus over TTL serial. This is a common protocol and fairly simple, so we will be able to interface to these controllers from our own microprocessors, such as an Arduino. With this kind of interface there are many possible uses, far more than what the usual three speed switch can do. More on that later.
Alan B
100 GW
PhaseRunner Testing has Started
BMC on test stand
The motor runs (but has a puzzling clicking noise), there are some controller configuration problems I'm working on.
BMC on test stand
The motor runs (but has a puzzling clicking noise), there are some controller configuration problems I'm working on.
Alan B
100 GW
I dug out my tach and measured RPM and voltage and calculated Kv at 4.8, instead of the estimated 6, so I entered that into the PhaseRunner Suite 0.9.4 (the about menu dialog says 0.9.2 but the filename says 0.9.4) software and saved it to the controller. It still errored out on Autotune.
I fired up the Bacdoor 1.4 software again and looked around, didn't see anything that looked wrong or that I was willing to try changing, so I went back to the PhaseRunner Suite and tried Autotune many times. It tripped out in different places each time, complaining of no motor, disconnected phase wires, etc. But occasionally it would go further, and finally it went all the way through. I changed nothing, other than leaving the hall cable plugged in despite it throwing errors.
BMC 4TT motor in 26" wheel
80 pole pairs (16 pp times 5:1 gearing)
Rs 68
Ls 194
Kv 5.53
Halls OK, 16.996 degrees
Unloaded max speed 300 RPM battery current 1A at 56.4V
I let it set everything else back to defaults, which is mostly what I was using anyway. It chose sensored start, sensorless run which is good.
The motor runs smoothly unloaded, aside from a clicking noise that appears to be motor related. This is a new BMC that I bought a couple years ago for this project and never used. It is mounted in a new steel fork which has not yet been mated to the frame.
I fired up the Bacdoor 1.4 software again and looked around, didn't see anything that looked wrong or that I was willing to try changing, so I went back to the PhaseRunner Suite and tried Autotune many times. It tripped out in different places each time, complaining of no motor, disconnected phase wires, etc. But occasionally it would go further, and finally it went all the way through. I changed nothing, other than leaving the hall cable plugged in despite it throwing errors.
BMC 4TT motor in 26" wheel
80 pole pairs (16 pp times 5:1 gearing)
Rs 68
Ls 194
Kv 5.53
Halls OK, 16.996 degrees
Unloaded max speed 300 RPM battery current 1A at 56.4V
I let it set everything else back to defaults, which is mostly what I was using anyway. It chose sensored start, sensorless run which is good.
The motor runs smoothly unloaded, aside from a clicking noise that appears to be motor related. This is a new BMC that I bought a couple years ago for this project and never used. It is mounted in a new steel fork which has not yet been mated to the frame.
Alan B
100 GW
I took the Bonanza RWD out for a test ride today. It has been a long time since I rode it, the main change since last time is the trigger shifters.
The new shifters are great. I can shift without moving my hands away from the handlebars, dropping the throttle, etc. The indicators show what gear is engaged. There is slight rubbing between the top trigger and the throttle / grip flange but it isn't bad and more clearance can be made. Edit - filed some away on the inside edge of both top triggers, reducing the rubbing.
This bike rides nicely, between the Big Apples at 40 psi, the front suspension and the Thudbuster. It steers quickly yet doesn't feel unstable.
It looks like something dangerous with all the wires and Lipo bricks, we will have to fix that. Edit - 72V triangle pack on order.
The rear brake squeals like a banshee. I hope I can fix that.
My son and I made a lunch and light grocery run. Here it is with groceries onboard. The solid rear rack will carry a heavy load.
The new shifters are great. I can shift without moving my hands away from the handlebars, dropping the throttle, etc. The indicators show what gear is engaged. There is slight rubbing between the top trigger and the throttle / grip flange but it isn't bad and more clearance can be made. Edit - filed some away on the inside edge of both top triggers, reducing the rubbing.
This bike rides nicely, between the Big Apples at 40 psi, the front suspension and the Thudbuster. It steers quickly yet doesn't feel unstable.
It looks like something dangerous with all the wires and Lipo bricks, we will have to fix that. Edit - 72V triangle pack on order.
The rear brake squeals like a banshee. I hope I can fix that.
My son and I made a lunch and light grocery run. Here it is with groceries onboard. The solid rear rack will carry a heavy load.
Alan B
100 GW
Here's a short video of the BMC runninng on the PhaseRunner:
https://photos.google.com/share/AF1QipNe_ak1Dk4lHUuSfwujSKazg5YoB8O78VISyl0EYbK7GiTPsLrLNWwCJvR5eKCYWQ/photo/AF1QipPIjUvin0vArFJm37Q4wybiVSNofUwaiNeuPuIC?key=YnQ4S0FrVnZrLWRVTHJ0d3RSRWpWMmF1dFZiTkdR
It doesn't sound quite right, there is some clicking that seems to be related to the motor RPM (not the wheel RPM).
I started a separate thread on the BMC noises, here:
https://endless-sphere.com/forums/viewtopic.php?f=3&t=82809
I received some great responses both here and from the vendor, apparently something in the motor is rubbing, and it is not critical, might go away, and not hard to repair (after getting the motor apart). So I'll have to decide whether to get it fixed right away or let it go awhile.
https://photos.google.com/share/AF1QipNe_ak1Dk4lHUuSfwujSKazg5YoB8O78VISyl0EYbK7GiTPsLrLNWwCJvR5eKCYWQ/photo/AF1QipPIjUvin0vArFJm37Q4wybiVSNofUwaiNeuPuIC?key=YnQ4S0FrVnZrLWRVTHJ0d3RSRWpWMmF1dFZiTkdR
It doesn't sound quite right, there is some clicking that seems to be related to the motor RPM (not the wheel RPM).
I started a separate thread on the BMC noises, here:
https://endless-sphere.com/forums/viewtopic.php?f=3&t=82809
I received some great responses both here and from the vendor, apparently something in the motor is rubbing, and it is not critical, might go away, and not hard to repair (after getting the motor apart). So I'll have to decide whether to get it fixed right away or let it go awhile.
Mmmm...you could add a big multiband antenna rig, and some nice big old Hammarlund gear with lots of knobs and vacuum tubes...oh, wait...that might be worse.Alan B said:
(and wouldn't help much with handling).
One simple quick fix would be a couple of coroplast (or other material) with slots cut in it's edges, to either ziptie or velcro-strap around the frame tube. Then go for something more permanent later on.
Alan B
100 GW
Coroplast is great stuff. I would definitely do something like that. I should get another sheet or a few, I've used up the one I had.
I have a new triangle battery bag, and the battery is in the mail, so I'll be ripping all the wiring off soon, and then I'll clean it up.
For today, I made an adapter between the Lyen and Phaserunner hall connections and tested the 9C on the PhaseRunner today on a stand. It fired right up and Autotuned the first try. About 200 mOhms and 460 uH, and 5.3 RPM/V which compares to the BMC at 5.5 RPM/V, so a little phase weakening should be able to match the Kv's.
I noticed that I hadn't matched motor phase colors, so I matched them and did another Autotune. It spun backwards, so I swapped a couple phases. Then it would not Autotune anymore. It was passing the first test, but the spinning motor test always ended with communications failure. I noticed that it accelerates really hard, and shakes the bike and stand. I retried a few times. Always failed. Finally went to the starting parameters and reduced current and extended time, and then it passed.
I injected 11 amps of phase weakening and the RPM went up 20%. So that works.
Next I need to haywire the PhaseRunner on the bike for a test run under load, and get the BMC repaired and mount the new forks. The new battery will probably arrive Tuesday and after that we can tear into the wiring and fabricate mounts for the PhaseRunners. I also need to modify the throttle for Justin's mixed voltage brake/accelerator.
On the regen braking, the throttle I'm using goes from 0 to 4V or so. In the 0 to 0.8V it should be doing regen. The software says it is doing braking and regen, but the deceleration of the unloaded wheel is very gradual, not at all like I would expect. So there's work to do on the regen to understand what is going on. 40A of regen should STOP an unloaded wheel almost violently, yet it glides down almost like it is not regenerating at all.
I have a new triangle battery bag, and the battery is in the mail, so I'll be ripping all the wiring off soon, and then I'll clean it up.
For today, I made an adapter between the Lyen and Phaserunner hall connections and tested the 9C on the PhaseRunner today on a stand. It fired right up and Autotuned the first try. About 200 mOhms and 460 uH, and 5.3 RPM/V which compares to the BMC at 5.5 RPM/V, so a little phase weakening should be able to match the Kv's.
I noticed that I hadn't matched motor phase colors, so I matched them and did another Autotune. It spun backwards, so I swapped a couple phases. Then it would not Autotune anymore. It was passing the first test, but the spinning motor test always ended with communications failure. I noticed that it accelerates really hard, and shakes the bike and stand. I retried a few times. Always failed. Finally went to the starting parameters and reduced current and extended time, and then it passed.
I injected 11 amps of phase weakening and the RPM went up 20%. So that works.
Next I need to haywire the PhaseRunner on the bike for a test run under load, and get the BMC repaired and mount the new forks. The new battery will probably arrive Tuesday and after that we can tear into the wiring and fabricate mounts for the PhaseRunners. I also need to modify the throttle for Justin's mixed voltage brake/accelerator.
On the regen braking, the throttle I'm using goes from 0 to 4V or so. In the 0 to 0.8V it should be doing regen. The software says it is doing braking and regen, but the deceleration of the unloaded wheel is very gradual, not at all like I would expect. So there's work to do on the regen to understand what is going on. 40A of regen should STOP an unloaded wheel almost violently, yet it glides down almost like it is not regenerating at all.
40A? Yeah, that might even try to spin the offground bike in a reverse wheelie (or a stoppie on a front wheel), applied suddenly.Alan B said:
I'm not sure what would be wrong to cause so little braking--is it actually pulling 40A back out of the wheel as it spins down? Or just programmed to do that?
Alan B
100 GW
It is programmed to do that, it says it is braking and regenning, but clearly it is not doing 40A. It is set to accelerate at up to 80A motor current, and that shakes the bike stand significantly with the unloaded wheel. But the decel/regen appears gentle. Must be another setting preventing full regen.
Alan B
100 GW
So today was the first test ride of the PhaseRunner on this Bonanza. To recap the setup:
9C 6x10 rear DD hub
18S 10AH Turnigy pack at 68V
PhaseRunner, latest model with pushbutton on/off
5K Magura throttle
Set for 80A phase, 30A battery
Hall start, sensorless run
CAV2 large screen display
Verry smooth. Like a torque throttle should be.
Silent. Never heard any motor sound, no grumbling, etc.
Accelerates nicely to 26 average
So gentle it feels a bit slow
The CA seems to work, then breaks into nonsense current/power levels. 1,2,3 kw then 15kw. Tried low and high range. It has always had this problem, I tried the bypass cap on it as recommended, it still has the problem. So no real power data.
But we are definitely in good shape with the PhaseRunner and the 9C.
After a couple miles of "testing" hard accelerations and climbing steep pavement the motor was slightly warm, and the controller was.. barely above ambient, and it is mounted on foam now, about as bad as it can get.
Except for regen and ebraking. Nothing there at all. Throttle goes clear to zero, but no braking.
Looking forward to seeing how "2" wheel drive works, and how much change there is with 72V.
9C 6x10 rear DD hub
18S 10AH Turnigy pack at 68V
PhaseRunner, latest model with pushbutton on/off
5K Magura throttle
Set for 80A phase, 30A battery
Hall start, sensorless run
CAV2 large screen display
Verry smooth. Like a torque throttle should be.
Silent. Never heard any motor sound, no grumbling, etc.
Accelerates nicely to 26 average
So gentle it feels a bit slow
The CA seems to work, then breaks into nonsense current/power levels. 1,2,3 kw then 15kw. Tried low and high range. It has always had this problem, I tried the bypass cap on it as recommended, it still has the problem. So no real power data.
But we are definitely in good shape with the PhaseRunner and the 9C.
After a couple miles of "testing" hard accelerations and climbing steep pavement the motor was slightly warm, and the controller was.. barely above ambient, and it is mounted on foam now, about as bad as it can get.
Except for regen and ebraking. Nothing there at all. Throttle goes clear to zero, but no braking.
Looking forward to seeing how "2" wheel drive works, and how much change there is with 72V.
Best I can say is to check the circuit path from the MCU all the way to the shunt wire input PCB pads on the CA (I'll assume you've long since checked the actual wire from that point onAlan B said:
). Every time I have had a problem with impossible power readings it has always been a connection issue to the shunt (in my case at teh JST connector or crimps, but it could happen anywhere in the path).
Alan B
100 GW
Good thoughts.
Here we have a CA plugged into a PhaseRunner. Both Grin products, their cables and connectors, etc, etc. Nothing homemade, crimped or soldered involved. Previously it was CA plugged into Lyen controller. Justin indicated at this voltage (66V) he had seen some break into oscillation, adding capacitance generally solved it. It seems fine with no load, it only goes nuts when the power gets higher, so I can't easily troubleshoot on the bench.
Here we have a CA plugged into a PhaseRunner. Both Grin products, their cables and connectors, etc, etc. Nothing homemade, crimped or soldered involved. Previously it was CA plugged into Lyen controller. Justin indicated at this voltage (66V) he had seen some break into oscillation, adding capacitance generally solved it. It seems fine with no load, it only goes nuts when the power gets higher, so I can't easily troubleshoot on the bench.
