Matt Shumaker electric recumbent bicycle
   E-cumbent
An Electric assist recumbent bike project by Matt Shumaker
Page1       Page 2      Matt's Edrive Kit

5/03/2008
It lives, it lives, it lives!

I finished fabricating the battery packs and the basic wiring. I laid out the components for you to see.

 The battery feeds power into the controller (through the data logger). The controller runs the motor (obviously). You can see the throttle (Servo Tester) running from the controller at the bottom of the picture. The Servo Tester requires its own very small power supply. That is the tiny component running from the battery leads to the Servo Tester. You can also see the data logger between the battery and the controller. Finally there is a digital display running from the data logger that will be mounted on the handle bars giving various bits of information from amperage, MAH used, voltage, motor temp, etc.
Along with that display I will mount this little green component.  In RC terms it is referred to as a "VSM" (Venom Speed Meter). It records and displays everything a typical bike speedometer does, speed, miles ridden, total miles (odometer) and max speed achieved on the last run. This item is lighter than a typical bike speedo and it saves all data even if disconnected from the electrical source. It is also very small and easy to mount to a flat console as I will be making.

I cut a portion of the seat bottom braces out to mount the battery packs. The seat is still plenty strong and now the packs have a nice place to rest.
The completed bike is very compact and light. I have not weighed it yet, but I am guessing it weighs about 50 pounds. 

I have been running it today for the first time. It runs absolutely PHENOMENAL!!! I cannot stress that enough! The bike is currently geared for 30 mph. It accelerates very hard. It pulls me from 0 to 30 mph in less than 5 seconds. It is capable of far faster acceleration. But, the main belt skips under load (motor to jackshaft primary belt).

 I will fix that. But, it may be a good thing. Because I have a feeling the bike would wheelie or brake something if that belt did not give way. This thing really pulls. If I blip the throttle, the bike hits me in the back and just hauls! It pulls 3,000 watts under acceleration. Wow, and that is with very cautious use of the throttle to try to protect the driveline.

It looks like I will have about 20 to 25 mile range on a charge (leaving 20% to 25% capacity in the pack). That is exactly what I was hoping for. Also, each time I run it, the efficiency goes up. I think the cells are "Waking up" with each cycle.
 

Anyway, I have about 10 or 12 miles on it right now (rainy day). But, from what I have experienced, I am totally addicted!

Other than the belt skip, the bike shows no mechanical problems what-so-ever.

For those who want the info, here are the parts I used:

  • AXI 5345/14 motor

  • Max Amps 10AH lithium cells (12 cells in series)

  • Castle Creations HV110 ESC (controller)

  • Astro Flight brand "Servo Tester"

  • Dimention Engineering VHV BEC (servo tester power supply)

  • Eagle Tree "Micro Logger" data logger with "Power Panel" display

  • Venom Speed Meter speedometer

Final drive ratio from motor to rear wheel, 25 to 1 (currently) for 30 mph top speed.

5/07/2008
The main belt (from the motor to the primary jackshaft) skips under load. I have to be careful of the throttle to keep it from skipping. So, I am doing a couple things to remedy this.

#1 I ordered more pulleys today. The gearing is a touch low for my needs. It runs about 28 to 29 MPH. I really want it geared for 40. So, I ordered larger motor pulleys. That will up my speed, while giving me more pulley teeth for the belt to engage. That should cut down on the belt skip.

#2 I am machining a larger idler for that belt. That will give me more wrap around the motor pulley, reducing the skip tendency further.

#3 I have an electronics engineer friend that is coming over this Saturday. We are working on an adjustable current limiter. Basically there will be a potentiometer mounted to the left bar to be operated by my left thumb. That potentiometer will adjust how much amperage is allowed to pass to the motor. This will give me much greater control of the power to the motor and I will be able to increase my efficiency by choosing to dial in an amperage that allow the motor to stay within its efficiency range longer.

The current limiter will be a total throttle unit what will take my throttle input and transfer that into the pulse width the controller needs to run just like the servo tester. However, the pulse width will also be altered by the amp sensing circuit to softly roll off the throttle when the amp limit has been reached (again, the amp limit will be adjustable on-the-fly by my left thumb). This will be good for a soft acceleration by allowing only a certain amount of amperage to get me rolling without shocking the driveline, then I can dial up the amperage for some high speed and/or high load running.

I will be at the shop this Friday making an aluminum and carbon fiber dash/console to mount the digital displays.

Oh, lastly, I will (eventually) be mounting the other rear wheel I have. That wheel has a disc brake flange and uses larger hub flanges and more spokes. I plan to mount that wheel sometime this summer. I will go up from my current 36 tooth rear sprocket to a 50 tooth sprocket. Then I will increase my ratio up near the motor to compensate. What that will do is allow me to push more power through the rear wheel. The wheel I have now is probably not up to the task of 4000 watts continuous (over 5000 watts peak) that the power system can put out. I am also afraid of breaking something in my power unit. This will cure that by removing some of the torque from that part of the system and moving it to the rear wheel. If I drop the ratio at the wheel and increase it at the motor, the effective load on the motor and pulleys goes down.

Anyway, that part of the project will wait. For now I will gear it up and work on the current limiter.

5/07/2008
Warren stopped by and took some videos of Matt and his bike. Note that the belt was still skipping in these videos, so the bike is actually a lot faster...


Matt's bike, Video 1


Matt's bike, Video 2


Matt's bike, Video 3

5/10/2008
OK, lots to report today.

I picked up a new motor pulley. It is now geared for 35 mph (confirmed with radar gun). I can tell you, 35 is plenty!

Now, there is good news and bad news;

The good news is;
The motor runs just fine pushing me at 35 mph. The bike handles very well.

The bad news is;
I need to make some mechanical changes to handle the power.
Basically, the primary belt still skips under load (though it takes alot more load to make it skip). Also, the secondary belt can be pushed to the point of skipping as well.

Now, that is the bad news. However, I anticipated this. So, I have a solution for each issue.

#1 I ordered new belts and pulleys that are 5mm HTD style for the primary belt reduction. That will totally cure the skip.
#2 The reason the secondary belt skips is because I am currently running an aluminum secondary shaft. That shaft is flexing under the load, causing the belt to slacken and the pulleys to misalign. This, in turn, causes belt skip. That is easy enough to cure. I already have a steel shaft to replace the aluminum one. Also, to reduce the load on the power unit, I am moving up from the current 36 tooth rear sprocket to a 50 tooth. I have room for chain clearance for a 50 tooth. It should fit fine. What this will do is reduce the load on the power unit by 39%. That will bring the load down to a reasonable level. What that will do, however, is move that load to the rear wheel. So, I will go ahead and mount my 32 spoke, heavy duty rear wheel and mount the sprocket to the disc brake flange. All this work will only take a few hours (maybe 5 or 6) to complete. With those changes, the bike should really accellerate hard. Odds are it will be capable of lifting the front wheel with that setup. In the mean time, I just take it easy on the throttle (not an easy task) and it runs like a champ.

I am surprised it takes the abuse it does. I thought I would have broken something long before this. This is all easy stuff to correct. I could always go to a smaller motor and eliminate any issues. But where's the fun in that?

Also, I met with an electronics engineer today. We are designing a new micro processor controlled throttle with adjustable current limiting, throttle exponential adjustment, and delay circuit. This new throttle will substantially reduce the load on the driveline by making the power hit much softer, while making the bike a lot more user friendly.

I also began making the handlebar console to mount the digital displays.
 
5/12/08
I finished fabricating and installing the handlebar console today.

I machined two split clamps out of aluminum on the CNC and hand cut the flat panel from 3mm thick CF/plywood composite.
The digital displays will show amp draw, battery voltage, watts, milliamp hours used, MPH, miles ridden (per trip and total) and top speed attained on that ride.
Eventually there will be a few LEDs and potentiometers related to the microprocessor throttle system we are building installed on the console as well. 
 Also, I have decided to fabricate a slipper clutch. The clutch will be set very tight. It will only slip if the throttle is snapped hard. That will prevent any driveline abuse should I accidentally hit the throttle too hard.

Anyway, it is running well. I am, however, enjoying the fabrication and R&D end of this.
5/14/08
Ok, I have gotten a lot of questions about the machining aspect of this project. So, to that end, I wanted to go a bit deeper into what is involved in making a part a reality.

I began the slipper clutch today.

The way any part begins is with a design. Here I drew the part on graph paper. Once it is drawn at 100% scale, I can easily program the CNC off that picture using the scale to the left and bottom of the picture as a reference.  

 

Once the part is drawn, I need to cut a piece of material to fit the mill. Here is a piece of 5/8" thick 6061 aluminum being cut on the band saw. It took about 15 minutes to cut that piece.

 Next I begin programming (actually, I was programming while the aluminum was being cut in the band saw).

I do all programming in line code, called G-code (General machine code). I do not use CAD software yet. Actually, this is easier than CAD for programming simple parts.

You can see the basic text code used.
Once the programming is done, I load that program into the CNC milling software (easy enough). At that point, I run a "Plot Preview". This shows how the part will look once it is cut. Here is the screen view for the plot preview. 
Next I clamp the material onto the CNC table. For aluminum plate, simple tension clamps work fine.

Here you can see the CNC cutting out the part. The white plastic item sticking down by the end mill is a nozzle for the shop vac. This pulls most of the chips away keeping the mill much cleaner.
Once the part is completely cut out, I chucked it into the 5 axis manual mill. This allows me to cut tapers and radiuses on parts. My 5 axis mill is the most versatile machine I own.

In this view, you can see the angle of the table allowing me to cut the 30 degree angle required for this part.
 Here you can see the part next to my original sketch. Other than the slot length, it is exactly as drawn. The taper removes material, while retaining the strength of the part. This part is one of two pressure plates for the slipper clutch. This one is the rear plate.
 So far, so good!
Here are both clutch pressure plates finished. I have 4 hours in these so far.

These parts are the pressure plates for the slipper clutch. The clutch will be comprised of these plates, the pulley that will also be the center friction plate, and a strong spring for tension. The plates I made today will sandwich a large pulley between them. If the throttle is banged too hard, the clutch will slip a small amount protecting the driveline. This is common in RC cars to protect the drive train. A slipper allows much more power to be used without ruining the drive train components.
I will bond simple friction pads onto the back side of these plates (probably very dense cardboard or similar fibrous material). That friction material will allow a small amount of slip without damaging the metal they are in contact with.

This slipper will not slow the bike down at all. It will only slip under the harshest of use, and only for a very short time to remove any unwanted jarring of the driveline.

My pulleys arrived today. I will be back in the shop tomorrow doing more work on the clutch.
 

5/15/08
I put in some more hours at the shop today.

I made the pulley (center friction plate). This was made from a pulley blank. That is a generic pulley that can be machined to whatever is needed.
This pulley is not the standard XL pitch I was using before. It is HTD 5mm pitch. This belt and pulley arrangement will not skip. However, without any belt skip, I will need this clutch.
 
 Without a clutch, or some other part that will skip or otherwise slip if the throttle is hit too hard from a stop, something in the drivetrain will break. This much power does have its drawbacks! But, with a clutch in the system, drivetrain breakage will not be a concern. The only issue is to make sure the clutch only slips when the system is overloaded, not during normal riding. That will be easy to adjust.

 
 Oh, my data logger is capable of reading two separate RPM inputs. So, I could put a magnetic pickup on the pulley itself and another on the clutch hub and my data logger would read the difference in RPM and log it in real time.
That way I could monitor clutch slippage under acceleration and set the clutch according to what I am looking for (preferably very little slip only when the throttle is hit too hard). I love technology! 
5/19/08
I got the clutch together and running. It works exactly as designed.
 
I decided (due to a number of emails I received regarding aluminum against fiber pads) to machine carbon fiber friction plates that are bonded to the friction portion of the pulley. The fiber pads now run against the carbon fiber plates. The clutch runs very well, indeed! It completely removes the touchy startup. No longer do I have to be super careful as I turn the throttle.
 
The tension spring for the clutch is a rubber elastomer. I got one that is far too stiff. That makes proper tensioning of the clutch difficult. So, I will order a few softer elastomers for experimentation.
 
The bike is proving to be very reliable (no matter how hard I ride it, I have never been stranded with it) and relatively efficient, especially considering how hard I ride the bike.
 
I found out the laws in Illinois are very relaxed for E-bikes. The law regarding "motor powered pedal cycles" is 2 horsepower (1500 watts) and 30 mph. Though my bike is well over 2hp, it is easy enough to stay under 30 mph. Plus, my motor looks so small, I doubt anyone (officer) would question if it is over powered.
 
I am adding outboard support bearings to the ends of the center jackshaft. I am seeing shaft deflection under load. The deflection is responsible for the belt skip I have been experiencing.
 
The bike ran well right after building it.  At this point, I am increasing its ride-ability and acceleration. So, I am just pushing for greater refinement and performance at this point.
 
More pictures and updates coming
5/20/08
I have been pushing the bike harder and harder (accelleration). I have gotten pretty good at tuning the clutch aas well as getting the bike to really launch. Anyway, as you know, I have been working primarily on eliminating belt skip. The clutch helped. But, I still want to get more accelleration out of the bike. So, after alot of R&D and pushing the limit of my belts I delaminated the 15mm wide belt on the secondary reduction. I was pretty bummed about that, not because I had to repair it (no big deal) but because I know I have alot more power available if I can get the belts to survive. So, I called a few engineers who recommended moving up to GT2 belts rather than the HTD belts I have been using. The GT2 belts are twice the load carrying capacity of the HTD belts I have now. So, that is the direction I am moving.

Tomorrow I am placing an order for the new belts and pulleys I need.
 5/23/08
Refinement of the system is rolling right along!

Here is the new rear sprocket. I tried for a 50 tooth. But, I settled on this 43 tooth sprocket (up from the previous 36 tooth). If you look closely, you will notice it is an old style Redline Flight sprocket machined for my wheel. this new sprocket reduces the load on the power unit and transfers that load to the rear hub instead. I also geared the power unit up to compensate.
 

The clutch is finished and working wonderfully. I have a simple butterfly screw to adjust tension. I can coast with the motor off and adjust it on the run if need be. Very simple setup. I went ahead and CNC cut carbon fiber plates for the fiber pads to drive against. This eliminates and possibility of the aluminum galling or wearing down. Again, the clutch works great! 

As I mentioned previously, I have been battling belt skip. Both the primary and secondary belts have been skipping. The primary belt is not a big deal. I have longer belt coming that will increase the amount of wrap around that pulley. That should (hopefully) cure it. If not, moving to a GT2 Power Grip belt and pulleys will definately cure the problem. Those belts handle twice the load before they skip. The seconday belt is another story, however. That belt was skipping and there is no easy solution for it. In fact, that belt (15mm wide) delaminated from the torque (shredded).

I considered going to the GT2 Power Grip belt for that drive. But, in the end, I settled on a 1/4 inch pitch chain. The chain makes a tiny bit of noise at high speed, but very little and it is bullet-proof. The chain drive runs great! I cannot run a chain for the primary reduction, however, because of the high RPM that drive runs. The chain would make a huge amount of noise up at 7,000 to 10,000 RPM the motor runs at.
 

At this point, the bike accelerates very hard. I still have some primary belt skip (still haven't received the longer belt yet). But, it only skips at a specific RPM. I think there is a harmonic going on. When I start off, there is no skip. Once up to about 6 or 7 MPH, the belt skips easily. Move past that speed up to about 9 MPH and the belt will not skip anymore, no matter how much load it sees. It is really strange. So, as long as I am careful with the throttle as I transition through that speed range, I can hammer the throttle from about 10 mph to 35 mph top speed and it just pulls HARD! I am thrilled. The longer belt should eliminate the last bit of skip.

Next I will be working on wiring the dashboard displays.


5/26/08

I spent another 3 hours at the shop today.

I had 3 things on the agenda;
#1 Replace the shredded drive belt (I did some serious acceleration testing yesterday).
#2 Install new clutch springs.
#3 Fabricate a caliper bracket for my new front brake disc.
You can see how the belt delaminated. I was doing some hard launches to see how far I could push the bike. I gradually increased clutch tension doing hard launches until something objected. Well, you can see, the primary belt objected. I will move to a 15mm wide belt next (the belt in the picture is 9mm wide). 
Once the belt was replaced I was on to the disc replacement. The reason for this new disc is to cure an issue I have been having with the stock disc brake. The original 160mm disc was designed for 20mph. I am pushing the bike up at 35mph (and over) with an extra 18 pounds of E-bike equipment on it. Well, the stock disc required a good hard 2 finger pull on the lever to stop the bike and it never really felt like it was happy about the abuse I was putting it through. I was also seeing brake pad wear from the hard lever pull and brake fade. So, I went on e-Bay looking for a larger disc. This is an Avid 203mm disc designed for downhill mountain bikes. It was new in the package for $21.99. Perfect! 
I machined the caliper mount (2 hours on the milling machine). The mount is light and very stout. 

Below, you can see the difference in disc size. Pretty substantial.
So, how does it perform? Fantastic! The brake is now a 1 finger moderate pull affair rather than a 2 finger (pull with all your might) ordeal. Also, the brake fade I was experiencing with the 160mm disc is gone. Not only that, the lever feel is better and braking is 100% consistent. I am thrilled!


Old disk


New disk
 Lastly I installed my new clutch springs. I originally installed a rubber elastomer. However, I used one that is far too stiff. I ran across these Belleville springs. They are essentially flat washers that are dish shaped. I am using 5 of of them mounted top to bottom in alternating succession in a stack. These springs have pretty much perfect tension for this application. The clutch is very consistent and easy to adjust now.
 

The bike is pretty far refined at this point. The last item is still the first issue I had, primary belt skip at a specific RPM. I think I will just go with GT2 belts and pulleys in 15mm width and be done with it. I have been resisting that because of a loss of efficiency. But, it is a small loss. I think I can live with it to cure the belt issue. Beyond that, developing the current limiter is the last detail.

Moving right along!
 

6/04/08
Belt skip cured!  Alright! I finally cured the belt skip and belt shredding problem.

Here's what I did; The belt I was using was 9mm wide 5mm pitch. I decided to go with a 15mm wide belt. You can see the difference in belt width between the first setup, and the new setup.

That did two things for me;
#1 - No more skip.

#2 - I was able to eliminate the belt wrap idler. You will notice in the pictures, the idler is still used. However, I removed it and still have no belt skip, yet the motor runs cooler with better efficiency (much better efficiency). However, I did not want to pull the side plate off to take another picture without the idler. But, I am thrilled to say, no idler is needed anymore and it runs great! 
One issue I was having as well (related to the belt skip) is motor shaft flex. So, I bored a 1/2 inch hole in the motor pulley and pressed a 1/4 inch by 1/2 inch bearing into that hole. Then I fabricated a support plate form 3/16 thick carbon fiber plate I got surplus from Boeing a few years back into a shaft brace. This CF is tough as nails! I have been saving it for just the right application. 
The CF plate is adjustable with the motor for belt tension adjustment. You can see the steel standoff in the end of the CF plate. That standoff goes into the bearing at the end of the motor pulley. This CF plate and bearing setup eliminates motor shaft flex and prevents the motor from moving back and loosening the belt under heavy load.

There is one final issue that I cannot easily cure, though. The AXI motor I am using has a strange trait.  It tends to confuse the controller (the back EMF going to the controller) under certain load at certain RPM.

It only happens once in a while under just the right circumstances. But, when it does, the motor makes sort of a singing sound and does not want to speed up any further until I reduce throttle a touch and get back on it. This is part of what was aggravating my belt skip. I spoke with the manufacturer of the controller. He explained it to me. It is not the controller, but this particular motor. He has seen it before in RC planes. It is no big deal. But, eventually, I will go to a different brand motor to cure this. I almost hate to bring it up, because it is so minor. But, the bike is running so perfectly, this one issue really stands out. 
 

Tomorrow I will be back a the shop modifying the clutch. I am removing the fiber friction pads and installing tough, dried leather friction pads. The pads I am running now are wearing down. the leather pads have roughly the same surface friction properties without the tendency of wearing down to dust like the fiber pads. I will post pics of the clutch update soon.

6/06/08
Here is a picture of the bottom of the power unit. It gives you an idea how compact everything is under the seat. I thought it was an interesting perspective shot and gives a much better idea of the complexity of this system.

I did some refining work on the bike today.

First I reduced belt tension a bit. That quieted the drive down to its lowest point ever. It runs really quiet now! I also gained back much of the efficiency I lost going to the wider belt.

Next I mounted the speedometer magnet and pickup. You can see I mounted it to the chain idler.

This speedometer is designed for RC cars and, as such, it is adjustable to a maximum of 20 inches of roll out per magnet revolution. That is not long enough for normal bicycle use (my rear wheel is 78 inches circumference, not 20 inches). It also achieves a neat added benefit; With the magnet on my idler, I was able to achieve the 20 inches roll out (maximum) the speedometer is looking for, and that also gave me much better resolution.

 A typical bike speedometer relies on one read per wheel revolution. It looks at the last few revolutions an averages them to give a reading. So, a typical speedometer refreshes the display slowly and looks jumpy as it skips from one speed to another every few seconds. This arrangement gives me 3 to 4 magnet readings per wheel revolution (depending if it is used on a 20 inch or 26 inch wheeled bike). Now I get 3 to 4 times the resolution and the speed changes on the display in a very smooth linear manner with greater accuracy in speed measurement. It also means at slow speed, the system is more accurate than a typical bike speedometer.
 I also fabricated a mount for my old trusty Burley trailer. I machined an aluminum mount and threaded two holes into the rear dropout. Now I can pull my trailer.

Anyway, that is it for now. I will update again when I do more work on the bike.

7/04/08
I have put about 200 miles on the E-Cumbent since building it and I am having a blast with it! I have, however, encountered a few problems. A couple of them are simple issues, not worth discussing here. However, one problem has been nagging from the beginning;

The AXI motor has an odd issue. The Back EMF (magnetic pulse the controller uses to "Read" the motor armature position) is not consistent. Basically the motor can resonates at one specific RPM which confuses the controller and make the motor eliminate a loud screeching sound. When that happens, I get off the throttle, let the controller read the motor RPM properly, then get back on the throttle. I did a lot of research into this. It is not a controller problem, it is a specific issue unique to AXI motors. It is rarely discussed because the motor is designed for airplane use. As such, it does not affect its operation in an RC plane. However, in my application, it is an issue.

So, I began looking into replacement motors. None that I found is a direct replacement. One, however, looked fantastic, but would not fit my bike without serious modifications. That is the Plettenberg Terminator 30-8 heli motor. I have run Plettenberg motors in RC helicopters for years. They are the best outrunners money can buy. Well, I e-Bayed a LARGE amount of RC stuff to buy this motor and it was worth every cent!

The issue mounting it is related to the diameter (among other things). It is 14mm wider than the AXI. The AXI was already a tight fit. This thing just barely fits and only after a large amount of machining.

I removed the power unit and the AXI motor. I then removed the motor front plate. That was done to drill and tap two 4mm holes to fit the 1.75 inch AXI mount hole locations. This was done to eliminate the need to machine a new plate on the power unit. This step was far quicker.

Next, I completely disassembled the power unit. Every screw and aluminum piece was removed, machined (if needed), cleaned, drilled, tapped, etc, etc, etc. Very few parts were not modified for this conversion.

Once I began disassembling the power unit, I removed a center divider wall in the power unit to provide clearance. Next I machined the heck out of the power unit sides. I removed every bit of aluminum I could from the power unit to make the motor fit. It is such a tight fit that the sticker on the can was too thick and got rubbed off in one spot! The fit is so tight, in fact, that I had to make a new chain idler bracket for clearance, and I only have 3mm fore/aft movement of the motor for belt adjustment, and that 3mm is only possible by moving the entire power unit back 3mm.

I spent a total of 7 hours on it. Now, how does it run? Wonderful! Here's the scoop on its performance:

The back EMF screech is totally gone. No problem at all! It accelerates perfectly smooth all the way from 0 RPM to full throttle. This motor has very little sound compared to the AXI. It does not growl and whine like the AXI. It makes a wonderful muffled whistle that is pleasing to the ear. It also is more tractable at very low speeds (5mph pulling my kids in the Burly trailer). This motor is also more powerful. The AXI is rated at 4,000 watts.
 

This Plettenberg is rated at 6,000 watts! However, I only pull 3,000 watts from it most often (hard accelerating). I have the slipper clutch set to protect the driveline. But, it is obvious the Plettenberg is less stressed when pushed hard than the AXI was. The AXI never really seemed to like being pushed really hard. It would run well, but always felt like it was unhappy with the use. Not so with the Plettenberg! 

The only downside is a slight reduction in efficiency. This motor is larger with a lower efficiency rating. I have not calculated it exactly (been pulling the kids in the trailer). But, it is taking a touch more amp hours to recharge after our typical ride. With the AXI, I averaged roughly 16 watt-hours per mile. If I had to guess, I would say this motor is more like 17 or 17.5 watt-hours per mile.

Not a bad drop, but it does seem to be a touch lower than the AXI. I will calculate it exactly and get back to you with the numbers. What I find odd about the drop in efficiency is the motor heat. I would expect an increase in motor heat with a drop in efficiency. However, that is not the case. This motor runs the same temp as the AXI.

I just got back from a ride. My watt-hours per mile on that ride was 14.49 at an average speed of 20mph!

8/28/08
I have put a lot of miles on the bike this summer and have learned a great deal. One weak link on my bike became apparent when I installed the Plettenberg motor. That new motor draws a lot more current (and pushes the bike a lot harder). However, with that extra power, I have found the controller cannot deal with it for long. After 10 miles of hard riding I blew the HV110 controller. I installed a new controller and that one lasted 100 miles as well. Hmm, frustrating. I spoke with Castle Creations about this, at length a number of times and I decided to go with another of their controllers designed for RC boat use. It is the Hydr HV240. That ESC is designed for 240 amps if liquid cooled or (according to Castle Creations) 160 amps without liquid cooling. Perfect!
 So, I ordered one. Once the new controller arrived I read the instructions. This new controller uses different programming than the HV110 airplane controller. There are too many differences to discuss here. But, suffice it to say, this new Hydra ESC is not going to be ideal for this application. Hmm, what to do? After staring at the Hydra for a while, I came up with an idea.........

 I noticed the Hydra has the same Logic board as the HV110 and uses the same FET boards. However, the HV110 uses only 3 FET boards, while the Hydra uses 4 FET boards. I called Castle Creations and asked them if the HV110 logic board would snap onto the Hydra FET bank. They said absolutely! So, I merely unplugged the FET board from one of my blown HV110s and plugged it onto the Hydra. Perfect!
Next I removed the liquid cooling tubes leaving the bare heat sinks. Last I removed two of the eight 180mf caps and added two 1,000mf caps in their place effectively increasing the capacitance from roughly 1,000mf to 3,000mf.
I have done a lot of research into blowing my HV110 controllers. I was told it may be due in part to voltage ripple. That ripple causes the FETs in the controller to see very short term low voltage situations (millisecond duration) dropping the FET voltage to 0 volts. This low voltage, but high amperage situation causes FETs to blow. The only way to eliminate that is to add more capacitance to the controller to feed those transients caused by the ripple.

Kind of technical, and a touch over my head. But, I had 3 electronics engineers tell me the same thing. 
 
You can see the two controllers and the differences between the two. I effectively built an HV160 controller.

I haven't run it yet. So, I will have to get back to you with the details on its performance. I will let you know how it runs when I get it back together!

I need to make it abundantly clear that this ESC issue is related to my use of the huge Plettenberg motor under hard acceleration. I have never run into anyone else blowing HV110s in bike use. So, please do not assume the HV110 is inadequate! I have merely been pulling 7,500 watts through a 5,500 watt controller. So, this is my solution. Also, Castle Creations is willing to custom make controllers with more FET boards for custom applications. They are costly and take 8 weeks to get, though. So, I went this route. We shall see how it performs!

11/22/08
Long time since an update! Sorry for the wait. I have received many emails asking for an update on the project. So, here we go!

First off, one major reason for no new updates is related to build success. The bike just works. I have over 600 miles on it with no problems other than a couple controller issues I hvae since resolved. Other than that, the bike just runs and runs and runs. I pull my kids around town, I run errands, and just plain have fun. I am hooked!
 

I was at the shop for 6 hours with "Nothing to do". So, I grabbed some 3mm carbon fiber I had lying around and some aluminum plate and decided to make myself a set of pedals any BMX rider would lust after. I am tired of slipping pedals.
 

These are the most aggressive pedals I have ever owned. They took alot of time to make. But, I am thrilled with the results.

The bike is currently at the shop (winter here). I am doing some upgrades over the winter. No hints just yet. You will see pictures as the work progresses.
 

   
   
   
   
If you have any questions, feel free to email me at shumaker@owc.net

-Matt

Back to the recumbent projects page