daleburr

My flywheel dyno is now complete.

A friend at work machined the flywheel, I did the rest at home on my desktop milling machine. The flywheel was balanced using a prop balancer and runs nice and smooth. The pod is mounted on a 1/12th car t-bar which has some flex to absorb the remaining vibrations.

The safety shield is 8mm lexan so is basically bullet-proof. I'm going to be testing ESC boost settings which will get the 220g flywheel up to about 50,000RPM; I didn't fancy doing that without some serious protection in case the flywheel comes off or a rotor snaps.

A graph from one of the first runs is attached. This is for an LRP X12 10.5 turn, with 4 dot (27.5deg) endbell timing and zero ESC boost.

These graphs are very useful for choosing your FDR and 'boost start' setting. This motor has its peak power at ~17000RPM, so I would choose an FDR so that the motor is at 17000RPM in the slowest power-limited points of the lap. This would usually be the first point after a slow corner that the car isn't wheelspinning. To do this you also need a data-logger on the car, which is something I always run.

I would also choose a 'boost start' RPM of just over 17000RPM (maybe 18000RPM) so that the boost kicks in just as the motors unboosted power starts to drop off.

More to follow as I do more tests...

mtaddei

Ciao to all, i'm also starting to "play" with home made dyno with sentry data logger. More or less everythings is working fine, but i recognised a ver frustrating problem with novak's program; once downloaded the file from the sentry and stored in the computer folder, if i'm trying to open the file just saved ithe program open an error windwos "unable to open the selected file".
Do you have any idea in which way i can solve this problem?

Tnks
Max

msalazar29

iTrader: (1)

I have a few questions when checking motors on what I need to be looking for. I race dirt oval in a 13.5 sprint car class so if I was to dyno a motor am i looking for looking for the least amount of resistance or how many rpms or is there something else I need to find we are not linited on battery just needs to be 2s please help

daleburr

You want the most amount of power, over the widest possible rpm range, for the least amount of current draw. Which is like wanting a big house with a nice garden, in the centre of a nice city, and having it for free :-)

It turns out building the physical dyno is the simple bit, this was done in a couple of weeks. One year on and I'm still improving my data analysis tools and how to correlate dyno data with track performance.

msalazar29

iTrader: (1)

ok i understand but now is there a way to adjust that without just buying a crap load of motors and hoping you get a good one

daleburr

Most motors have adjustable endbell timing, and you have FDR adjustment. The dyno will allow you to optimise these to get the best out of the motor you're running.

Oval racing is pretty simple for motor setup, compared to circuit racing. The first thing to do is race with the Sentry on your car, so you can see what rev range you're using on track. Compare that to your dyno power curve and adjust the gearing so that you're in the peak power range for as much of the lap as possible.

msalazar29

iTrader: (1)

thank you very much for the help

Steve S

iTrader: (5)

Your torque curve does not look right. It should be almost linearly decreasing to zero at max RPM. Maybe you have some freaky weird motor, but you might want to check your calculations as it appears to be skewing the power curve as well.

frozenpod

iTrader: (1)

Firstly, great thread.

Agreed the torque curve does look quite right but it does appear correct based on the power curve.

IMO if you have a dynamic timing ESC, I would reduce the motor can timing to increase low rpm power output and retest. You can add more timing with the ESC for greater power in the higher rpms.

Combination of the two should given you more power at all rpms.

daleburr

You're absolutely right; that's a very old graph from one of my first dyno runs.

I had been using a Hobbywing Xerun v2 ESC, which turns out isn't a good ESC for dyno testing. I had to ramp the throttle up slowly at the start of the run, otherwise the ESC would error and go into sensorless mode. This results in the torque and current curves you can see in the graph.

I've recently switched to a v3 Hobbywing and this works much better, and can cope with an almost instant throttle ramp-up. I also use a USB servo controller to get repeatable ramp-up times.

That graph was produced using a spreadsheet I got off the McPappy site, which used a cubic fit on the data. I quickly realised that a cubic fit isn't a good fit for any of the variables, and is massively effected by when you choose to stop the run. I now use a 6-order fit for power and torque which works much better.

Here's a more recent graph. You can still see the current ramp-up at the start, but this happens quickly and at an RPM we'll never see on track. This graph shows an endbell a timing comparison (+2 and +3 on the Team Wave timing marks). Note there's no Torque line; since Power is a function of Torque and RPM, I don't find any need to have both displayed.

daleburr

Here's a nice recent graph I posted in the Trinity D4 thread to show an example of 17.5 power curves.

Howard Cano posted the following analysis:

I did measure the voltage during the run. It drops to around 7.9v at the start of the run (under highest current draw) and recovers linearly to around 8.25v by the end of the run as the current reduces.

This particular pack has slightly higher IR than my race packs (which is why it's consigned to dyno work), but I have done a comparison to the best packs available (new SMC 7200 70C) and the data/curves are pretty similar. I think this is to be expected of any battery source; the voltage will always drop under load, giving an asymetrical power curve. Since this will also happen on-track I'm okay with that.

When I started out with my dyno project all it could do was produce graphs like the one attached. These are useful for checking a motor is still good, making sure a new motor is in the ball-park power wise, and determining maximum sensible boost settings. However they don't really tell you which setup will be best on track, especially once different gearings are considered. It's even harder with boosted as the power/current curves are all sorts of crazy shapes.

So I continued to develop the dyno software to calculate lap times and motor temperature for a given dyno run and track model. Fortunately I've been running a Novak Sentry for years so have plenty of track data available to create the track models I'm using.

The situation at the moment is that it does a pretty good job of calculating the 'power-limited' laptime and current at the start of the run when the car matches the dyno run (cold motor and full lipo). It can do a sweep of gear ratios and show which ratio will be fastest over a lap.

The next step (and ultimate goal) is to calculate how much dropoff each setup will have (due to motor heat and lipo discharge), and therefore calculate the performance over a full race.

howardcano

iTrader: (37)

It sounds like you have been quite busy! Thanks for posting this info.

Two items catch my eye: Your test shows a particularly non-linear current vs RPM, and the power output in your test does not peak at 1/2 of the free-running RPM. Both are different than I would have expected, and also differ from most published curves I've seen. Here is one example from Mr. Stranahan in this thread:
http://www.rctech.net/forum/5684698-post24.html
And here are some typical examples from the web:
http://lancet.mit.edu/motors/motors3.html#tscurve
http://www.namiki.net/product/dcmotor/curve.html

Do you have any ideas on why this is occurring?

P.S. Your dyno shown in post #331 is absolutely gorgeous!

daleburr

In truth I'd never looked into what the curves should look like, so maybe it's time I did. The current line looks pretty similar to the one JohnS published, the dip is just more noticeable as I used a larger graph scale. I can only guess this is due to Lipo behavior. I could try a couple of packs in parallel to see if that reduces the dip.

It's possible I'm sampling data for too long, John talks about cutting off the data when the voltage stabilises as it does get noisy at the end when the acceleration is quite small. I'm on holiday but will play around when I get back.

Thanks for the feedback, it's pretty quiet in here these days!

old_dude

iTrader: (1)

I use a Fantom dyno and with what I know about electric motors is verified on it. When a flywheel load is attached to the shaft and a motor is given a plug start it will do the following.
1. Attempt to accelerate the flywheel with its maximum capability.
2. Draw all the amps it can handle with that decreasing as the motor approaches its peak rpm and and acceleration decreases.
3. Peak wattage will not be at time zero but peak torque will be. With wattage being the product of torque and rpm that point will move around based on the motor timing and the motor configuration.

The Fantom's curves are very much like those Howard's post.

With that being stated you should see max amp draw and torque very near time zero. The issue I suspect is the power source. That is why the Fantom wants a fat one. I use a garden tractor battery which through the Fantom is regulated to 5 volts to the motor controller.

kufman

iTrader: (6)

Reviving an old thread here. Anybody else have this problem with the optical rpm sensor on their eagle tree system? (see picture) Am I exceeding a d-rpm/dt limit or something?