Human Electric Trike Thesis

Design of an electrically assisted human powered trike

Installed drain brain

Posted by Bob Dold on Sunday, October 15, 2006 9:10 PM

Received the Drain Brain on Thursday and spent today installing it on the X1000 scooter. Installation requires splicing the shunt between the batteries and controller and hooking up a wheel sensor. The picture below shows the unit installed on the handlebars:

 

The drain brain displayes the speed, watts, battery voltage, and distance traveled. It also stores the maximum current drawn and maximum wattage. After using it for a couple charge cycles it can calculate the miles left on a charge. After a brief test, it confirmed the maximum current was limited to about 30 amps, this occured when accelerating up a slight grade.

Posted in Electronics | Leave a Comment »

10-15 Dry fit

Posted by Bob Dold on Sunday, October 15, 2006 9:03 PM

  Took some of the bonded parts I put together yesterday and did a dry fit with the parts I have completed. The batteries are sitting in their position between the frame rails.The shock is shown in it’s installed postion connected to the bellcrank.

 

Posted in Components | Leave a Comment »

10-13 Received parts

Posted by Bob Dold on Friday, October 13, 2006 8:57 PM

 Received all waterjet parts from BigBlueSaw yesterday and picked up majority of machined parts from Pete at the WNEC shop. The picture below shows most of these parts laid out:

 

The plan is to bond together what I can over the weekend to be ready for the unfinished parts when they are completed.

Posted in Components | 27 Comments »

Fabrication progress

Posted by Bob Dold on Friday, October 13, 2006 9:19 AM

Stopped in to see Pete this morning to check on progress, he has completed the majority of the parts, the chart below shwos the parts left to complete:

Remaining parts should be complete in the next two weeks.

Posted in Components | Leave a Comment »

CFD Analysis

Posted by Bob Dold on Monday, October 9, 2006 8:50 PM

To get an idea of the aero drag on the trike with a rider I built a FloWorks model of it to study the drag loads and flowline around the trike. My first model is a crude model withhalf of the trike and rider modeled to take advantage of symmetry to allow the model to run faster. The picture below shows the simplified trike model used in the analysis.

For this analysis I used an external air flow speed of 25 MPH, ground effects and spinning effects of the wheels and pedals were ignored. From this run, the solution converged after 44 iterations at about 2.4# for half the trike, or 4.8# total. The plot below shows the pressure on the rider and trike at 25 MPH:

 

To compute the Cd of the rider and trike, the drag is divided by 1/2 the density times the frontal area times the velocity squared.

Cd = Drag / (.5 * density * A * V^2)

Published values for the Cd of a traditional bike and rider vary from .70 to 1.1, a fully enclosed trike with an aero shell similar to the GoOne would be around .150. A recumbent trike should be somewhere inbetween these values, probably on the order of .50 to .70. From the model the frontal area is 5.58ft^2, filling in the rest of the equation yields a Cd value of .53 – right about where it should be. To study further reducing the drag, a fairing may be added to the simulation to see how much it helps reduce the trike’s drag.

 

Flow trajectories at trike centerline:

Posted in Theory | Leave a Comment »

Front Suspension Analysis

Posted by Bob Dold on Friday, October 6, 2006 7:05 PM

Today I used CosmosWorks to check the front suspension design, I used the same loading as on the rear swing arm; 3G up, 2G aft, and a 1G lateral load. G is equal to the loading on each front wheel, which comes to 125#. A solid tetrahedral mesh was used on the a-arms, spindle and pushrod with a total node count of 79,560 and 43,415 elements. The stress plot below shows fairly low stresses except for the lower rod end which is around 50ksi. The rod ends are steel so this is an acceptable stress for them.

 

The displacement plot below shows a deflection at the spindle of about .040in, which is acceptable. However, it also shows a large deflection of the pushrod of up to .080″ which indicates it may have buckling problems.

To check the buckling I modeled the push rod by itself and applied the axial load it is seeing of 530# at the 3G, 2G, 1G case. The results of this analysis show a buckling factor of below 1 (.53), indicating there is a problem. I plan on rerunning the analysis using a steel pushrod to see if can withstand the load without bucking, my other alternative is to go with a larger diameter aluminum pushrod.

Posted in Uncategorized | Leave a Comment »

10/6 Update

Posted by Bob Dold on Friday, October 6, 2006 5:55 PM

Received Garmin Edge 305 yesterday, charged it up and used it in my car on my trip to work this morning. The speed seemed to be accurate to my car’s speedometer within a MPH, I logged the entire ride and was able to plot out the time vs. speed, grade, and elevation. The default sample rate was about once every 3 or 4 seconds, it can be set to sample every second which I will set it for when doing the test runs. See below:

The steepest grade on my way to work was about 15%, right at the start of the ride. I also found a program (G7ToWin) that lets me export the raw data to Excel so I can combine it with the Omega voltage logger data when I receive it.

The 305 also let you export your path to a Google map, as seen below:

Also went ahead and ordered 3 Odyssey PC925 motorcycle batteries from BatteryWeb for $98.88 each, these will be wired together in series to form a 36V, 27 amp-hour pack weighing 78 pounds.

Big Blue Saw emailed me and said they would be delivering the waterjet parts next week, so I should be able to start assembly next weekend.

Posted in Updates | 1 Comment »

Completed design

Posted by Bob Dold on Thursday, October 5, 2006 11:35 PM

Below are images of the final design and a link to an eDrawing of the design:

Link to eDrawing: zeept.htm (may require eDrawings download)

Posted in Design, Updates | Leave a Comment »

USB Current Logger

Posted by Bob Dold on Monday, October 2, 2006 10:58 PM

In order to track motor current vs. time I plan on using one of these $70 Omega units http://www.omega.com/ppt/pptsc.asp?ref=OM-EL-USB-1…  and then sync the data with the trip data from the Garvin Edge 205. I can measure the current using a shunt resistor using these directions provided by Dr. KJ:

 

DC Current Measurement

Place a shunt resistor in the current path and measure the voltage over the resistor.

Note: This technique can only be applied to DC currents, not AC 

The value of the shunt resistor must be chosen carefully as follows:

Step 1. Find the maximum current

You may know this value (in Amps) already, but if not it can be found from the power rating of the motor (or whatever the load is):

(Amps) = Power (in Watts) / Voltage

In your case: I = 30 A

Step 2. Find a suitable value for the resistance

Make sure that the resistance is not too high

Too much voltage will be dropped over the resistor, taking power away from its actual purpose (in this case driving the motor). As a guide, try to avoid dropping more than 1% of the available voltage over the shunt. In the example above 1% of 24 V is 0.24 V

If the resistance is too low

The voltage seen by your A/D will be too low, so the current measurement will be less accurate and noisy or stepped. Avoid this by ensuring that at least 0.1 V appears over the shunt at maximum current. In the example above the voltage over the shunt should ideally be between 0.1 V and 5 V, when the maximum current is flowing. Experimenting with resistance values R in the formula:

Voltage(V) = I R

leads to the choice of 0.01 Ohms, which will develop 0.3 V at 30 A.

Step 3. Calculate the Wattage required for the resistor

We don’t want the resistor to burn out, so fit a suitably high wattage one using the formula:

Power(Watts)= I2 R

In your example, Power = 30 x 30 x 0.01 = 9 W, so you’ll fit a 9 W resistor. Once this is connected up you can configure the acquired voltage to be scaled directly in amps instead of in voltage.

 

OM-EL-USB-3
VOLTAGE DATA LOGGER
Range: 0 to 30 Vdc
Resolution: 100 mV
Accuracy: ±1%
GENERAL
Memory: 32,000 voltage readings
Logging Interval: 1 seconds to 12 hours
Operating Temperature Range: -25 to 80°C (-13 to 176°F)
Alarm Thresholds: High/low alarm thresholds selectable in software
Start Date/Time: Selectable in software
Status Indicators (LEDs): Red and green
Power: 12 AA 3.6 V lithium battery (included)
Battery Life: 1 year typical (depends on sample rate, ambient temperature and use of alarm LEDs)
Weight: 57 g (2 oz)
Dimensions: See dimensional drawing

OM-EL-USB-4
CURRENT DATA LOGGER
Range: 4 to 20 mA
Resolution: 0.1 mA
Accuracy: ±1%
GENERAL
Memory: 32,000 current readings
Logging Interval: 1 seconds to 12 hours
Operating Temperature Range: -35 to 80°C (-31 to 176°F)
Alarm Thresholds: High/low alarm thresholds selectable in software
Start Date/Time: Selectable in software
Status Indicators (LEDs): Red and green
Power: 12 AA 3.6 V lithium battery (included)
Battery Life: 1 year typical (depends on sample rate, ambient temperature and use of alarm LEDs)
Weight: 57 g (2 oz)
Dimensions: See dimensional drawing

Posted in Electronics | 2 Comments »

Swing Arm Stress Analysis – Updated

Posted by Bob Dold on Monday, October 2, 2006 10:57 PM

After checking my intitial analysis I found some errors that caused the arm to see 4X the loadin it should have, here are the corrected results:

 

I analyzed the swing arm with a 3G vertical load, 2G aft load, and 1G lateral load, with G being equal to 140# which is the loaded weight on the rear wheel. I used a mixed shell and solid mesh in CosmosWorks 2006 as shown below:

 Von-Mises stress plot and displacement are shown below:

 

Max stress at this extreme condition is 17ksi, which is well below the 42ksi ultimate and 35ksi yield for 6061-T6 aluminum tube (Mil Hndbk 5)

Displacement is accptable at about an 1/8 of an inch at the dropouts.

 

– Updated 10-8-06

Posted in Design | Leave a Comment »