• Announcements

    • WhiteWolf

      Donation Info   11/12/2017

      If we can make it thru to the end of March we will be good till then PayPal thechopperunderground@cox.net Thanks WhiteWolf
mrriggs

Homemade Motorcycle Dyno

74 posts in this topic

Well, the motor on that little drill press finally crapped out. It was making nasty sounds and didn't have the power to turn the drum. The night was still young so I tore into it to see what happened. It appears the rotor got so hot that it expanded enough to lose the interference fit and walk down the shaft.

Nothing to lose at this point so I took the rotor over to the press and pressed the shaft back down where it should be. Slapped the motor back together and fired it up. No more noise and it has the power to spin the drum again. :thumbsup:

I finished up the third go-around on side-B and I think it's balanced as good as I'm going to get it. Overall, it is very smooth. You can still feel a bit of vibration but the frame is not visibly deflecting. Best of all, when I spin it slowly and let it run down, it stops in a different place every time.

I still need to turn some steel weights to weld on in place of the lead weights. Then it's time to "calibrate" it.

Edited by mrriggs

Share this post


Link to post
Share on other sites

The only calibration required for an inertia dyno is to figure out how much inertia the drum has. To measure this, I duct-taped a fishing line to the drum, wrapped it around a few times then ran it through a couple of pulleys, one behind the drum and one on the ceiling.

DynoCal01.jpg

DynoCal02.jpg

First you figure out how much weight is required to just turn the drum. This value is subtracted from the test weight to determine how much weight is actually accelerating the drum. I made a hook for the end of the line and added washers until it would fall completely with just a little nudge. If it moves without a nudge then it's too heavy since static friction is greater than kinetic friction. If it doen't make it all the way to the floor then it's not heavy enough.

I spent a bunch of time messing with this and determined that it takes 2.4oz to spin my drum. Then I realized this number is meaningless since I hadn't added the auxiliary rotation sensor.

The drum will only spin about two times before the weight hits the floor so the built-in 1-signal/rev rotation sensor won't work. To get better resolution I drilled a hole in a hockey puck and shoved it onto the end of a distributor, then fabricated a mount to hold it against the drum. The hockey puck spins 7-3/4 times per revolution of the drum. The distributor puts out eight signals per revolution. Eight times seven and three quarters is 62 signals per revolution of the drum.

DynoCal03.jpg

Unfortunately, the distributor added considerable friction. You may notice there is a lot more than a few washers hanging on the line. It ended up taking 2lb 15oz just to turn the drum.

For the tests I used an old barbell weight.

DynoCal04.jpg

Subtract 2.94 from my 8.5lb test weight, the net weight is 5.56lbs. The radius of the drum is 0.979 feet. Multiply those together. The torque on the drum is 5.44 lb-ft.

All I had to do then is make a dyno pull on the falling weight and adjust the inertia value until the table showed 5.44 lb-ft.

DynoCal05.jpg

Um, yeah... Not what I was expecting either.

Share this post


Link to post
Share on other sites

Clearly, I need to spend more than five minutes making an auxiliary sensor. I think the problem, besides the insane friction, is the mechanical advance. It never really gets spinning fast enough for the weights to take up the slack so I can hear it rattling around in there.

That test was a flop so I decided to try a more conventional test using a stopwatch to time how long it takes the weight to hit the floor. With the test weight, distance to floor, diameter of drum, and time of fall, you can calculate the inertia. After several attempts I concluded that my reflexes are not sharp enough to get a meaningful number.

Knowing the exact inertia of the drum isn’t that important. I can plug in any random number and still be able to measure percent improvement from one run to the next. The numbers may not match the numbers from another dyno, but that’s pretty common. I think it was Hotrod magazine that took their car to five different dynos and came back with different numbers from each.

I’m still going to take another shot at the calibration. Not because it’s that important to the finished product, but because it’s an interesting challenge. That’s really what this project is all about. Trying new things, learning, problem solving, and just plain having fun.

Share this post


Link to post
Share on other sites

I was planning to build a new auxiliary sensor but decided to give the distributor one more shot. I removed the mechanical advance weights and got everything lubed up so it spun a bit easier. It was also mounted with less force between the hockey puck and drum. This time it only took 1.6 lbs to turn the drum which works out to 6.76 lb-ft of torque.

The dyno chart still isn't a straight line but much improved from the last attempt.

DynoCal06.jpg

The up and down is very consistent which lends me to believe it is just the runout of the worn-out distributor and less than perfect hockey-puck-with-a-hole-in-it. All the samples average out to the target 6.76 lb-ft with an inertia value of 7.75 lb-ft-sec2. The stopwatch tests I did last week showed about 7.8 lb-ft-sec2 so I think the new numbers are right on target.

So, now it's balanced and calibrated. Just need to mount the wheel chock and add some tie-down points and it'll be ready to stick a bike on there.

Share this post


Link to post
Share on other sites

Right after hitting the post button it occurred to me that if the error was from sensor runout then it could be eliminated. I modified the CSV so each reading was the average of last eight readings. The results are what I was hoping to see. The only variance left is the runout of the drum itself.

DynoCal07.jpg

This correction technique worked so well, I will be incorporating it into my dyno program.

Share this post


Link to post
Share on other sites

The only part of that I got was "hockey-puck-with-a-hole-in-it", and even that has left me scratching my head.

Share this post


Link to post
Share on other sites

I've been very unmotivated the last couple weeks, haven't stepped foot in the garage. When my cousin told me he got his turbo Sportster running but it needs tuning, that got a fire lit under my ass.

The roller still had the lead weights on it when I was calibrating it. Last night I finished turning the steel weights and got them welded on. The goal is to have a bike on this thing by Saturday.

Share this post


Link to post
Share on other sites

What made you choose eight samples? Is that based off the iteration of sample data rate? For the purposes of the tool that data smoothing sounds appropriate.

Share this post


Link to post
Share on other sites

What made you choose eight samples? Is that based off the iteration of sample data rate? For the purposes of the tool that data smoothing sounds appropriate.

I averaged eight samples because the distributor puts out eight signals per revolution. If the distributor only put out a single signal per rev then you would never see the runout because it would automatically average out. I could have just cut seven of the eight teeth of of the reluctor in the distributor and run the procedure again. That would have given the same flat line but was unnecessary since I already had the data necessary to interpolate the same readings.

The data averaging won't do anything on my dyno since the roller only puts out a single signal per revolution. But I'm designing the software with the idea that other guys can download it and use it for their own dynos. Despite how complicated I've made all this look, inertia dynos are extremely simple. The only challenge is making the roller. Once that's done you can have it balanced at an electric motor shop which would bypass all the crap I just went through.

Share this post


Link to post
Share on other sites

I get it. There's no reason to do things if they aren't complicated, difficult, and annoying, right?

Share this post


Link to post
Share on other sites

IT WORKS! IT WORKS!

We ran bikes on the dyno this evening. I blasted it over 90 miles per hour and it is SMOOTH. The software worked well too.

Here is a run from my XS650.

XSdyno01.jpg

My uncle shot some videos but I don't have copies yet.

I am just so excited. This is HUGE! Now that we know it works, I can focus on refining it into a practical tool.

Share this post


Link to post
Share on other sites

Hell yes. That rules man.

Share this post


Link to post
Share on other sites

Excellent! Congrats,

I almost hate to see it work, the build has been so entertaining.

Dusty

Share this post


Link to post
Share on other sites

Excellent! Congrats,

I almost hate to see it work, the build has been so entertaining.

Dusty

It may be working now but it is still far from done.

I picked up an Arduino Uno board from Radioshack on the way home today. It's a programmable microcontroller board. These thing are really trick, much easier to program than the microcontrollers I was playing with ten years ago. I'll use it for capturing the analog signals; thermocouples, pressure sensors, wide-band O2, etc. That part will be easy. The real trick will be getting it synced up to the dyno signal from the soundcard.

Share this post


Link to post
Share on other sites

I fumbled through programming the Arduino this evening. This is the first time I've played with one of these. The syntax took some getting used to but the language isn't that difficult. In one sitting, I was able to learn the code, write the program, and upload it to the device.

The Arduino Uno board has six analog inputs and connects to the PC through a USB serial port. I set it up to read the analog inputs and send them to the PC whenever it receives a signal. I tested it with the Windows Hyper Terminal and it works well.

The next step is to modify the dyno program to read from the Arduino.

Share this post


Link to post
Share on other sites

I figured out how to get Python to read the Arduino at the same time it's reading from the sound card. The command line dyno recorder has been updated to read everything and record it to a csv. I still need to update the Excel workbook to read the csv.

The total cost for this dyno at this point is about $110. It'll probably cost another $500 to get some wideband O2 sensors, thermocouple transmitters, MAP sensors, etc. At that point, it will be able to do anything that a commercial inertia dyno would do.

Share this post


Link to post
Share on other sites

I'm done messing around with the software, for now. The Excel workbook has been updated and I made a lot of changes to the command line recorder so it's easier to use. The GUI program can wait.

If anyone wants to build their own dyno then here is the new and improved software;

DataLog.ino is the "sketch" for the Arduino Uno.

WAVdyno.zip contains WAVdyno.exe and all the files it needs to run. This is the command line program that records the dyno run. You're supposed to be able to run WAVdyno.exe without installing Python, but it only seems to work on my computers that have Python installed.

WAVdyno.py is the source code for the command line recorder. You can run this instead of WAVdyno.exe but it does require Python2.7, as well as pyAudio and pySerial (all free downloads).

DynoCalculator.xls is the Excel workbook that processes the raw data from the command line recorder and creates the table and graph for the run.

Share this post


Link to post
Share on other sites

I dug through the junk boxes and found some terminal strips and a die-cast enclosure for the Arduino board. Now it looks like a proper data logger.

DynoLog01.jpg

DynoLog02.jpg

Share this post


Link to post
Share on other sites

This is pretty inspiring.

Share this post


Link to post
Share on other sites

My cousin was trying to figure out gearing for his Bonneville racer so I added a few more features to the Dyno Calculator.

If you plug in the gear ratios and tire size then it will calculate the force where the rubber meets the road (thrust) and determine the optimum shift points. It will also estimate top speed and tell you which gear and rpm it will be reached.

ShiftPointTable.jpg

It also makes a graph of the thrust per mph as well as the estimated drag per mph. Where the thrust and drag lines intersect is the top speed for that gear.

ShiftPointGraph.jpg

The values shown are from my bike. As you can see, the first gear ratio is poorly matched to the rest. I told it to limit shifts to 7500 rpm. With a higher redline it calculates the 1-2 shift at 7700 rpm. The early model XS650's had a 14/31 first gear which would bring my shift point down to 7100 rpm.

Share this post


Link to post
Share on other sites

the maths! :blowup:

Share this post


Link to post
Share on other sites

Well done. Threads like this really motivate me to get back out there working on my own projects again.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now