John Stranahan

iTrader: (27)

Joe I like your setup. Matt may make something very similar except it will have a heavy axle that can be spun up to dyno the chassis.

The only thing that troubles me is that you are using amps measured on the track as the gauge to be repeated. If the motor cannot be geared to run at maximum power due to overheating then this is a reasonable aproach. For example the 17.5 pulls 31 amperes to make 169 Watts maximum power. If by chance you thought gears were optimized at 25 amps your logger would just tell you what your seat of the pants told you. If you had geared considerably taller and lower and got poorer lap times then understandably you have the best data available, track lap time data is better than dyno data.
I see Nick Case starting to use gears the dyno predicts are the best from a max power standpoint. I wonder if others should at least try them for a heat. When you throw in 1:1 gears your amp draw will increase, so will your power, and so will the heat evolved, A temp check at 2 minutes in a practice would be highly desirable. A slow throttle roll on to keep out of the 75 amp range will be desirable from a heat standpoint. Not meaning to meddle in your buisness, but this is the approach I took on the flat oval and it worked very well.

If you tried this gear and laps improved and the motor was not too hot then of course you should try a bigger one. If the motor got too hot then you are forced to gear below max power, but maybe not for a short speed run.

If you mail me a 21.5 I will put it on the dyno and then would be able to suggest a gear based on max power. In addition you would know what amp draw produces the most power. That can be used directly by your rig.



Novak SS 10.5 motor with Replicate tests

This is the motor I run in our World GT (pan) cars on a road course. I have run it quite a bit. I ruined it when the female plug on the Sentry RPM harness lost a piece that prevents plugging it in backwards and so I did. It ruined the sensors. I happened to have another motor with burned windings. I took the sensors out and put them in the 10.5 If you have a good soldering iron with a very hot tip this is not too difficult.

I ran three tests with Matts Heavy Flywheel. I rested the motor 15 minutes between runs the motor averaged 304 W + 1.4 W. That is just outstanding precision from a flywheel dyno and a homebuilt one at that. That + number is the Relative Standard Deviation and is only about .3% of the total value. Note the decimal. The actual values were 301,307,304 W. Attached is a sample graph from the output of the centered value.

I run this motor with a 72/24 gear and 2.4 inch tires. I will measure the runline with a scale drawing of the track and see if I can make any relation to average speed and gear.

SWTour

iTrader: (9)

John,

Some of the limits we face, because I have our series locked into a fairly LOW mAh LIPO battery is RUN Voltage, and watching the voltage drop curve.

With the 3200/3400 we pull the 21.5 motors around 18-20 amps, and race for 5 minutes.

I regularly do a discharge test on my batteries in "RACE" trim temperature condition.

Of course the initial launch off the line sees a higher amp draw, so does my test method.

Here's a quick example of our battery voltage at timed intervals.

SV (Static Voltage) 8.42 volts (no load)
LOAD Applied - 8.09 v
10 seconds 8.00 v

150 seconds 7.41 v

300 seconds 7.21 v

and since our SHOOTOUT is a 33 lap race (with an estimated average lap time of 11.0 seconds. I went ahead and tested the packs to a full 6 minute drop.

360 seconds 7.01 v

SO basically over the 6 minutes we'll have one full volt of drop, minus the top charge, and this will consume approx 2150 mAh from the battery (Since it's a 6 minute run...it's easy math to see we'll average 21.5 amps)

The car WILL go faster with a HIGHER mAh battery, and a 2s2p pack will generate a LOT LESS voltage drop (But we don't allow them)

With the basics we use, most cars will see a 1 second drop off in lap times from FAST lap to LAST lap.

ie: 2nd lap is the fastest generally (First is from a dead stop) and we'll see a 10.0 or 10.1 -- and generally the FINAL lap will be 11.1-11.2

....I would send you a 21.5 to test, but I don't have any extras. (I'm actually ONE short right now...I fried our GOOD ONE trying to dyno @ 1:1

John Stranahan

iTrader: (27)

My apologies then. We are 1 and 1 for ruined motors druing this thread. Think of the Glory though (or the Sacrifice for Science). That slow roll on would apply on the dyno as well. Maybe that's what Nick has the knack with as during those speed runs he participates in (I participated one year) traction sucks and a slow throttle roll on is essential to get up to speed to get your aero devices working. I understand your voltage drop. This has always been the case. If I understand what you are saying is that a higher gear ratio will result in more voltage drop later in the race which negates the effect of running closer to maximum power.

Gearing the 10.5 for road course.
There is always a limitation and that is what we are trying to find. The best limitation is when we find the lowest lap times. The 10.5 starts the mod motor series. On the 10.5 if I put the peak power at the average speed the resulting pinion with a 72 spur is a 40 tooth and average amp draw is 60 amp. This motor is not capable of this average amp draw so then you have to gear down. What I have done is geared low enough that when I have medium traction, I don't break the tires loose very often during throttle roll on. When we get new asphalt, traction will go up, my pinion will go up 1 or two teeth. I have about a 40 degree temperature buffer right now before I get too hot, but those 140 F track temps are not too far away. Note that although we don't spend a lot of time at max power we do pass through it for a time every time we accelerate out of a slow corner. So in effect we are back and forth over that power hump every corner.

Our runline is 525 feet. Laps in the 20 second range are possible with this car under good traction conditions. Tires are 7.536 in circumference. Axle RPM for max power is 7523 RPM yielding a 1.8 ratio for 13,500 rpm at Max Power.

A 3.5 motor is capable of 45 average amps or more.

SystemTheory

SWTour,

Thanks for the photo and description of your battery limitations and lap times. If you can post information on the bank angle, turn radius, and length of the straights, I will generate a track map.

John,

I understand your gearing algorithm may work well on a variety of ovals. If so, I would like to understand why.

The Fantom power on 4 cells in your post #3 is much lower than the Sentry power on 4 cells for the 13.5.

If I can get the torque-speed curve on a graph or spreadsheet for the 13.5 on 4 cells it would assist with an acceleration study of the Oval versus Road Race lines. I can convert the torque-speed line to a force-velocity line to show how the Road Race cars will punch harder out of the corners than the Oval cars due to the different installed gear ratios.

I typically model a NiMH battery as 1.4 volts after peak charge and 1.5mOhm per cell in a Sub-C series stack. If you have better data please inform me.

The time to roll through peak power with the same flywheel should not change with increased voltage on the NiMH stack, but it does tend to increase with greater internal source resistance as the stack grows. This difference may be negligible or it may actually show up on the Dyno.

I'd like to look for predictability in your peak power measurements on NiMH if I also knew the open source cell voltage, cell resistance, torque, current, and radian speed at the peak power point for all three cases. If you have any brush motors of similar power to the 13.5 it would be interesting to study a similar comparison of outputs on the growing NiMH stack.

mattnin

iTrader: (44)

Isn't there another school of thought that would make sense to gear for max efficiency instead of max power?

-Matt

John Stranahan

iTrader: (27)

Mattnin-when batteries were more of a limiting factor maximum efficiency made more sense. I never used efficiency. I always maximised power in certain regions of the brushed motors power band. Max efficiency on these brushless motors seems to be in a higher rpm lower power region of the motor curve.


System Theory-The reason for the first low power reading in post #3 is stated in that post. It has to do with the "race start" algorythm in the Sphere speed control. Later, I bypassed this by goosing the dyno with a small throttle application braking it to a stop. A 5 second stop must be avoided. The second average is 67 W. This is power calculated by the Fantom software but using a 4 cell pack. The Sentry gives a value of 135 W or so. So the question is which is correct. I can vouch for the motor feeling better on four cell than our old stock motors were on 6 cell (135-140 Watts peak). I have to believe the 135 W. I had no control of the calcualtion of 67 W. I think we are getting good numbers on the powerful mod motors. We are not getting 900 and 1000W unbelievable readings. We are getting 560 W getting to the track. Note the first test on one 3.5 was low, but the charging technique was not in place and the motor saw less voltage.

I have complete control of the calculations on the Sentry. I have high confidence in it. We will have to look at some Robi numbers. They would be more comparable.

I still do a small pretest with The LRP TC control. I marked a neutral out on the servo tester. It is easy to reach. I brake the dyno with my finger on the flywheel to get set up for the first test.

I have Sentry files available. These Fantom runs did not produce a usable file due to the lack of the $75. Windows Fantom software. I'll see if I have some better comparison data.

"I understand your gearing algorithm may work well on a variety of ovals. If so, I would like to understand why."

for my simple gearing algorythm to work (gearing for max power at the average speed) you must have a motor that is capable of being run at max power for 5 minutes; it does not overheat. You also must have skilled competitors doing their best to attain a high average speed. We are looking for tiny gains by applying dyno tuning.
Turns out not having an overheating problem might be rarely the case or only with a 13.5 or 17.5 on a flat oval.

If you are putting out max power during times of acceleration your lap times will be lower. No majic here. Pass back and forth right over that power peak hump.

SystemTheory

Peter,

You could estimate cornering speeds with a stopwatch and measurements to find each turn radius. Check the link I insert at post #13 for ways to get good measurments, and download Google Sketchup to use the Arc Tool to draw a track map and get arc lengths. Obviously improvements in your setup will tend to increase usable traction and corner speeds. For a single speed gear choice in scale racing the tradeoff is simply to make more force coming off a slow corner, or less force off the same corner and more top speed in the long straights.

John,

Discrepancy noted. Sentry data will do just fine. You can post the graph to this thread or send spreadsheet. I would like to recalculate cornering speeds on the oval and map operating points on the 13.5 torque-speed line. If I get some good results in my acceleration models I'll post results here.

Matt,

SWTour describes a motor operating closer to the maximum efficiency point, due to battery limitations, as I understand it. This reduces the equilibrium speed on the big oval but saves energy to finish the race at competitive pace.

John's inside line on the Houston oval is also more fuel efficient, but for different reasons. The total energy draw from the battery is smaller per lap than the guys driving the outside line with a taller gear ratio. Thus in a long race the car on the inside line should generate more peak power due to a fresher battery and a cooler motor, other factors kept equal.

John,

I trust the Sentry over the Fantom at this point and with a torque-speed line and the times on your Houston track I can probably verify more than 67W required for accelerating the mass of the car in the time window.

John Stranahan

iTrader: (27)

Here is Jev UK's Robi output on a 10.5. I don't have access to the data file but the voltage is likely to be the same as my just peaked on the charger 5 cell run with the 10.5. Power in the 180-190 W range for both. Amp draw about 45-50 ampere for both. I believe my 6 cell tests just put out more voltage than the Robi. Here the difference was 30%. There is no Factor of Two difference however. Notice the Robi efficiency readings are crap when rigged up to do brushless.

http://www.rctech.net/forum/electric...ml#post4096497

petk

Hi John,

Thanks for testing the 10.5, and your thoughts on road course gearing.

Here in Denmark we are racing 5 cell/9.5, the closest track to my home is a 265 m asphalt track. http://www.vejleminirace.dk/default.asp?id=33&groupID=2

I put 2000 rpm on your 10.5 results, and calculated optimum gearings. Not very scientific, but I don't have a dyno yet. Track length is 265 m, and average lap time is 22 sec.

Max power ratio 4,38
Max efficiency ratio 4,94
Actual ratio 4,8 comming of the track after 5 min 65 C, so with current temperatures I could go up to 4,71.

As you say it seems that the motor temperature is the limiting factor. So the better we are at keeping the can cool, the closer we can get to running at max power?

Best,
Peter

John Stranahan

iTrader: (27)

Peter-Thanks for the post. I think you are on the right track with your motor gearing. Fabulous track. If your traction allows it gear up that last little bit.

John Stranahan

iTrader: (27)

Novak velocity series 6.5R

I used to race this motor in a 10 turn brushed class in a TC. A good 10x1 and this motor were very equal on the track when the 6.5r was geared to survive the heat. Notice the peak power is 398 W. Unfortunately this occurs at a 96 amp draw. This means you have to gear the motor much more conservately to keep it from getting too hot. It likely has 150W more peak power than a good 10x1 but a lot of it is unusable except in brief burst.
My opinion after testing a lot of the brushless motors in a TC on a big 1/8 scale track is that you might as well be using a 4.5 or 3.5 and not be pushing a lower powered motor so hard that it thermals the speed control off. Your best motor in the heat is going to be one of these two.

I checked the sentry software for anything that might cause a 2 fold error. I found none. On the Fantom you can select the inertia of the flywheel. I did not have this window available to change it. (the two flywheels are a factor of two apart in inertia) I think that this is probably not the problem, as Fantom numbers have always been on the low side compared to other dynos. I only had an Aluminum flywheel, which is the standard flywheel, in my box.

SystemTheory

Approximate Oval Performance Study, Images Attached:

1. Oval versus Road Race Lines

2. Torque-Speed Curve transformed to Force-Velocity Curves

Dyno data directly from John's spreadsheet page 2 for Novak 13.5 on 4 cells making 135{W} peak power. Assume ideal gearbox with no frictional loss and zero inertia. Actual machines would make less axle force. Tire radius 28{mm}.

3. Force-Velocity Curves transformed to g-Rate Acceleration Curves

Force equals mass times acceleration. Solve for acceleration and divide by standard gravity g = 9.807{m/s/s}. Assumes zero driveline inertia with chassis mass m = 1.16{kg} = 41{oz}. Actual machines have less g-rate due to driveline inertia, driveline losses, and air resistance.

4. Force-Velocity Curves transformed to Power-Velocity Curves

Power equals force times velocity, apply standard SI units to plot watts. I put up a conservative cornering rate aC = 2.2 g's to map corner exit velocity to the best case power curves at maximum voltage. I put up 3/4 power just to have an image of the impact of picking up the throttle. This shows both cars near peak power on full throttle going into the straight section.

Comments

Matching the 4.4{s} lap time is difficult due to simplifying assumptions. The Road cars pull more g's in the "straight" with lower gear to keep pace with the Oval cars carrying more speed and pulling less g's with taller gear.

The Oval car could accelerate from 23{mph} up to 32{mph} in 18{ft} at a constant rate 1g. The power P = Fv would go from 117{W} to 160{W} in a straight upward slope to do this. This suggests 135{W} on the Sentry OK. A faster corner exit is likely.

The battery stores energy as milliwatt-hours {mWh}. The First Law of Thermodynamics says energy equals work plus heat.

E = W + Q

The Road car takes less work out of the battery on each lap because it invests less kinetic energy in rotating parts. The Road car dissipates less heat against driveline friction and displacing air because it has less velocity. The average current draw is lower in the Road car (current flows to do more work). This explains the cooler motor on the Road car.

John Stranahan

iTrader: (27)

I thought I would add some more description. The road race car was actually an oval car. It became that solely because I moved the batteries to the far left side. It was a home made prototype with a 3-link rear suspension and dual independent A-arm front suspension. This is different than your normal pancar. The line I took with this car was optimized by years of experience to suit this particular car and get its lowest lap times. In other words, it had better cornering g's as well as better forward grip available. This allowed a tighter line. A very similar line was run at the carpet nats because the carpet gives you better cornering grip and forward grip. We used different gear ratios somewhat because of the difference in the line length. Other consideration for the oval guys were tuning out corner exit ills with a high gear ratio which numbs the car. Possibly the thought that hotter is faster with a 13.5 just like with the more powerful mods

System Theory-Nice Effort. I think maybe the last graph you posted needs a little more explanation. What exactly is it telling us. The vertical dotted lines are important. Radar measured top speed in the 30-33 mph range on the straight. I see the complete oval simulator coming. You plug in your dyno numbers, gear ratio, track dimensions, runline, coefficient of drag, frontal area and it calculates a lap time. Then you tinker with the gear ratio and see what happens. This might be an interesting tool for oval racers.

I expect a new digikey amp sensor very soon. On its arrival I will send Matts dyno back with many thanks. I have asembled a duplicate dyno with the Aluminum flywheel to do some 3.5 testing and speed control testing later in the week.

SystemTheory

John I'll try to better explain that last graph. I am interested in working on a lap time simulator model and may email you for feedback on the project.

Post #37 shows oval run lines approved by John and my cornering power model:

http://www.rctech.net/forum/5694203-post37.html

Applying these assumptions at 3 g's results in too high an average speed given the 4.1{s} fast lap and 4.4{s} average lap.

This graph assumes aC = 2.8 g's on each line in the tight corners. A part throttle setting takes the power down to enough to push air and overcome friction in the slower corner segment. There is no power invested into the acceleration of mass and rotating parts during a constant velocity corner.

Here John's car on the inside line would exit at 21.4{mph} and he rolls on the throttle, coming up to full throttle before the middle of the diamond segment. A throttle profile is not available, and I've shown a jump to the peak power on the RED power-velocity curve to show where his car is geared. As he says, this car can better transition from side bite to forward bite due to tuning, and it takes some driver skill to apply the throttle profile.

If cars can pull 2.8 g's on the outside line they exit with 25.7{mph}. The exit is into a straighter line and less throttle skill is needed. A jump to full throttle would take the car up to peak power on the BLACK power-velocity curve due to how these cars are geared.

If each car averaged the corner speed all around its line, it would make lap times listed in the graph. John says top speed is 32{mph} on a speed gun, and I assume this occurs on the outside line, and it requires acceleration in the straight off a corner slower than 25.7{mph} or else lap times would go below 4{s}.

I suspect John's car rolls through and over the hump in peak power on the 46.75{ft} segment on the RED curve, and the outside cars roll over the peak power hump in the short 24{ft} straight for a burst of acceleration, with cornering power on the order of 2.0 to 2.5 g's.

SystemTheory

John,

This model runs in an engineering simulator to estimate electric car acceleration. I've used it to study the full scale Tesla Roadster and the 1/10 scale X-Ray T2 007, but with only ballpark system parameters available. I use a similar model to approximate rocket powered cars and cars rolling downhill.

It solves the differential equation of motion with T(ws) as an input source adding power and energy to the mass and inertia elements, with D(v) + b(w) dissipating power as heat due to velocity feedback losses. It obeys Newton's Laws and the Conservation of Power and Energy.

The gear ratio G and tire radius r operate as ideal power and energy transformers. Assumes no heat in the gearbox and no tire slip. The net force at the axle flows into mass m (like an electic current) toward ground, and this mass obeys Newton's Second Law F = ma.

The simulator solves for forward acceleration a and takes the integral to find the forward velocity at node v (solving the impulse-momentum integral for the net driving force). An integrator provides forward displacement at node x.

1. The torque-speed source is a look-up table from the Sentry Dyno.
2. Js is the estimate of the rotor inertia in the electric motor.
3. G is the overall gear ratio (speed reducing or overdrive value).
4. Je is the estimate of driveline inertia at the rolling axles, less wheel inertia.
5. J is the moment of inertia of each tire/wheel assembly.
6. r is the tire radius.
7. b is the driveline damping.
8. m is the mass of the car.
9. D is drag from air density rho, frontal area Af, and drag coefficient Cd.

All system values in standard SI units to keep power in watts in all elements. I write code to convert and plot outputs in United States Customary Units.

For the Novak 13.5 Oval study I need reasonable estimates of items 2, 4, 5, and 9. If you want to work up some numbers for any of these missing values I'll adjust the model and post up full throttle acceleration curves. I have the motor current here and may include current analysis too.