R/C Tech Forums

Go Back   R/C Tech Forums > General Forums > Electric On-Road

Reply
 
Thread Tools Search this Thread
Old 05-04-2009, 08:47 PM   #121
Tech Fanatic
 
Join Date: Jun 2006
Posts: 872
Trader Rating: 35 (100%+)
Default

I have got a Reedy 5000 35C I can try out when I get back into town next week. I can try it out on a Novak or Hacker 13.5. I can compare it against a Tenergy 5200 25C LIPO.
__________________
High Desert Raceplace, Grand Jct CO, Sanwa Exzes Plus Stick Radio, Spektrum,Thunderpower 230g, Modified Tri-Nut Novak GTB2, Ballistic 4.5t 550, 5.83lb SCTE Ten, Novak Sentry Brushless Dyno, Crossweight Setup Station, Junsi 20A power supply Icharger 20A Charger,TP610C, 22B, Novak Edge, Novak 13.5, SC10, Havoc Pro SC XDrive, Ballistic 17.5 Matthew Joseph Cordova
mattnin is offline   Reply With Quote
Old 05-05-2009, 07:06 AM   #122
Tech Adept
 
Join Date: Nov 2007
Location: Vine Grove, KY
Posts: 228
Trader Rating: 4 (100%+)
Default

Gameover, that's absolutely awesome information! How about running the same tests with a Sphere, Tekin and GTB ESCs to see how they compare with the same motor? This is the kind of stuff that can be really useful and help us understand what's going on in these ESCs.

Thanks!
richard.bratton is offline   Reply With Quote
Old 05-05-2009, 01:02 PM   #123
Tech Elite
 
yyhayyim's Avatar
 
Join Date: May 2006
Location: Weston, FL
Posts: 3,424
Trader Rating: 26 (100%+)
Default

maybe someday someone will test Speed Passion GT 2.0 ESC and motors. Very popular and they also have the new "mini Dyno" for their ESC and motors, to determine which timing/gearing is most efficient...

the GT 2.0 ESC is rated at 800amp output and can run up to 2.5 motors...http://speedpassion.net/product.php?lang=&c=55

http://www.racing-cars.com/usa/produ...ecnumber=50627

Keep up the great work John!!!
__________________
Yohan Hayyim [Associated RC10B4.1]

"Release the Kraken"
yyhayyim is offline   Reply With Quote
Old 05-05-2009, 03:31 PM   #124
Tech Adept
 
Join Date: Apr 2008
Location: Melbourne, Australia
Posts: 237
Default

Quote:
Originally Posted by richard.bratton View Post
Gameover, that's absolutely awesome information! How about running the same tests with a Sphere, Tekin and GTB ESCs to see how they compare with the same motor? This is the kind of stuff that can be really useful and help us understand what's going on in these ESCs.

Thanks!
i have a Sphere TC-Spec and a GTB ESC, i also have a Novak 10.5t and a Tekin 10.5t.

I'm also going to test the LRP 17.5t with the SPX.

I think i also have some Novak 13.5's and i have an LRP 8.5t and 4.5t (X11's). I probably won't test these with my current flywheel.

I don't have any SP stuff...
gameover is offline   Reply With Quote
Old 05-05-2009, 07:19 PM   #125
Tech Adept
 
Join Date: Nov 2007
Location: Vine Grove, KY
Posts: 228
Trader Rating: 4 (100%+)
Default

Quote:
Originally Posted by gameover View Post
i have a Sphere TC-Spec and a GTB ESC, i also have a Novak 10.5t and a Tekin 10.5t.

I'm also going to test the LRP 17.5t with the SPX.

I think i also have some Novak 13.5's and i have an LRP 8.5t and 4.5t (X11's). I probably won't test these with my current flywheel.

I don't have any SP stuff...
I can't wait to see the Sphere and GTB results!
richard.bratton is offline   Reply With Quote
Old 05-06-2009, 10:39 AM   #126
Tech Adept
 
Join Date: Mar 2009
Location: Metro New York
Posts: 139
Default

Inertia of Rolling Car Mass Reflected to the Power Source

The power source is a single shaft from an electic motor or internal combustion engine. Assume an ideal gearbox with zero inertia, no heat loss, and full tire traction under power. These assumptions do not change the calculation of J because the Conservation of Mass and Energy apply to the ideal system.

Apply a Set of Coherent Units (meter-kilogram-second, slug-foot-second)
m - car mass
r - tire radius
G - speed reduction gear ratio
v - forward velocity
w - axle speed with full traction
ws - shaft speed of source
Ja - inertia of car mass if reflected to the rolling axle
J - inertia of car mass if reflected to the source shaft

Kinetic Energy of Forward Car Motion
Kv = 0.5*m*v^2

Kinetic Energy of Inertia at the Drive Axle
KJ = 0.5*Ja*w^2

Reflected inertia Ja must have the same amount of energy as translating mass m.

Conservation of Energy
Ja*w^2 = m*v^2

Assumption of Full Traction
w = v/r

Substitute and Solve for Ja
Ja*(v/r)^2 = m*v^2

Ja = m*r^2 |reasonable result

Kinetic Energy of Inertia Reflected to the Source
K = 0.5*J*ws^2

Conservation of Energy
J*ws^2 = Ja*w^2

Speed Reduction Gear Ratio
ws = G*w

Substitute and Solve for J
J*(G*w)^2 = Ja*w^2

J = Ja/G^2 = (m*r^2)/G^2 **|inertia of car mass reflected to the source

The source "sees" the car mass as a direct flywheel load of inertia J (at full traction).

Oval Car Example (m = 41{oz}, r = 1.14{in})

m = 1.16{kg}
r = 0.028{m}

Car 1: G = 2.00 Oval Racing Line

J1 = 2.274E-04 {kg-m^2}

Car 2: G = 2.44 Inside Racing Line

J2 = 1.528E-04 {kg-m^2}

Fantom Steel Flywheel

JL = 9.89E-05 {kg-m^2}

Comments: this is the minimum load inertia reflected to the source, since it neglects the inertia of rotating parts in the driveline. The cars on the outside line obviously require more kinetic energy to reach higher velocities. This is energy taken from the battery and then scrubbed off on each lap to get back into the low speed corners.

Contrary to an opinion stated above, the inerital load on the track may be greater than that of a flywheel on the Dyno. However you only need to spin up the "track flywheel" partially, since the car begins with kinetic energy off of each turn. Much racing occurs at part throttle, and there is only one holeshot, so the system does not accelerate through a full run as on the Dyno. A bigger flywheel/taller gear takes more time to spin up, and draws more average current (not peak current) due to the extended spin up time.
SystemTheory is offline   Reply With Quote
Old 05-06-2009, 12:38 PM   #127
Tech Rookie
 
Join Date: May 2009
Posts: 19
Default

Quote:
Originally Posted by SystemTheory View Post
Contrary to an opinion stated above, the inerital load on the track may be greater than that of a flywheel on the Dyno. However you only need to spin up the "track flywheel" partially, since the car begins with kinetic energy off of each turn. Much racing occurs at part throttle, and there is only one holeshot, so the system does not accelerate through a full run as on the Dyno. A bigger flywheel/taller gear takes more time to spin up, and draws more average current (not peak current) due to the extended spin up time.
I was basing this mainly on my experience with own brushless vehicles, I'm yet to have a vehicle that can't 'get up to (full) speed' in a high-traction environment in 2-3 seconds or considerably less. Whereas the more ideal flywheel tests were taking 5-7 seconds, b'cos the way the measurement numbers were only coming back with 0.1 second intervals.

But continuing this train of thought, would it be possible to develop better dyno tests, or ones more representative of racing? I mean tests that more closely simulate what actually happens on the track (you describe it above). The sole focus from now has just been 'spin up from rest' tests. Even without a 'brake dyno', if you had a realtime display of the flywheel, you could spin the flywheel up to 10,000, and then keep it there with minimal throttle input. Then start the test (with the flywheel already spinning) while flooring the throttle, measure the actual power output. I'm guessing if you did this, you could well get different results than just doing a straight ramp from rest to max RPM (and looking at what happens around 10,000RPM). Just b'cos of how the ESCs firmware works, if nothing else. What do you think?
Access is offline   Reply With Quote
Old 05-06-2009, 02:44 PM   #128
Tech Adept
 
Join Date: Mar 2009
Location: Metro New York
Posts: 139
Default

Access,

I appreciate your direct experience with brushless systems, since any good theoretical model must be consistent with such observations.

** Mattnin's flywheel is ~3.45E-04{kg-m^2} and John captures about 2.6{s} for the Hacker 13.5 on 4 cell NiMH to produce a torque-speed curve. When I run this torque-speed curve as input in an engineering simulator with Js = 4E-07 plus JL = 3.45E-04{kg-m^2}, the response takes 2.6{s} and cuts off before output torque goes to zero, just as John did in the spreadsheet. I used this simulation to confirm the analysis below and my model J = (m*r^2)/G^2, which gives back the same response time if I set mass = 0 and add J to Js at the source. Oval cars geared on the inside line take 1.11{s} to accelerate from zero to top speed with 2.1 g start.

I estimate the Oval Cars geared for the outside line take only 1.7{s}* for a full throttle run up in an ideal (frictionless) system and require about 1.7 g's* starting rate of acceleration, small enough to prevent tire spin in a well balanced car. If I add friction in the driveline and air resistance the top speed is reduced and it takes less time to get up to full speed!

One might think the presence of friction adds spin up time. It actually reduces the time to get up to top speed and reduces the kinetic energy invested in system mass at the lower top speed.

However friction and air resistance add elapsed time on the track. The rate of forward acceleration is reduced and so is the top speed in long straights.

The simple answer to your excellent question is that such tests could be performed but are expensive in terms of time and equipment and thus not very practical. In your example, one could Plot the full throttle rpm, current, and power curves versus time as measured on the Dyno. Choose any rpm as the lookup index, and when you pick up the throttle at the index rpm, the current and power curves will evolve to the right as traced on your plot. Take the difference in time and that's the spin-up time for that system response.

To extend this to all track conditions would be impractical b/c tire radius and gear ratio change system inertia, and you can't build such a variable flywheel into the Dyno.

An engineering model takes the torque-speed curve at full throttle as measured on the Dyno, uses some reasonable method to modulate this curve with a throttle profile (validated on the Dyno and through track performance measures), and uses these as inputs to a well considered system model in a differential equation solver. Right now it is not clear to me the mathematical model of throttle setting versus Dyno response for a brushless system, and this is compounded by the problem of variable profiles in the speed control.

Last edited by SystemTheory; 05-06-2009 at 03:35 PM. Reason: * corrected values; ** new comment
SystemTheory is offline   Reply With Quote
Old 05-06-2009, 04:18 PM   #129
Tech Elite
 
Join Date: Aug 2002
Location: Houston, Texas
Posts: 3,777
Trader Rating: 27 (100%+)
Send a message via AIM to John Stranahan
Default

Novak Ballistic 3.5 Test coming soon
see a small report on the construction in my CRC thread.


The dyno run most closely resembles a start from a stop. We do this once per race and once per crash. Working with the dyno and actually having it on the radio trigger gives you a feel for the beast. You have more resistance right at first spinning that flywheel in spite of what inertia might be felt at speed. Then 3.5s seem to woosh up to speed faster than on the track after the start is over with. Even a stock motored rubber tired touring car on high grip surfaces will spin the tire accelerating from a stop whether you notice it or not. The driver that adjusts the motor acceleration to match traction available will accelerate the fastest. That sneaky speed control on my M8 radio will provide some outstanding acceleration on flat parts of a slippery dirt track, but will fail miserably in the turns from throttle lag and will fail on the ramps of jumps where you have more traction.

I believe if you runup a GTB from 10,000RPM to max you will get the same power curve. There is no significant tinkering going on in the normal profiles. The LRP probably will do little better on the flying start as it tends to stutter from a stop on the flywheel (note again never on the track). I would expect exactly the same peak power as this does not seem to be affected much by timing changes. The power band is only shifted.
Attached Thumbnails
Dyno, Homemade, Using a Novak Sentry Data Logger, Continued, The Experimental Thread.-novak-ballistic-3.5-motor-003.jpg   Dyno, Homemade, Using a Novak Sentry Data Logger, Continued, The Experimental Thread.-novak-ballistic-3.5-motor.jpg  
__________________
[email protected]
TeamCRC.com Mikes-HobbyShop.com

Last edited by John Stranahan; 05-06-2009 at 08:13 PM.
John Stranahan is offline   Reply With Quote
Old 05-06-2009, 07:19 PM   #130
Tech Rookie
 
Join Date: May 2009
Posts: 19
Default

Quote:
Originally Posted by John Stranahan View Post
I believe if you runup a GTB from 10,000RPM to max you will get the same power curve. There is no significant tinkering going on in the normal profiles. The LRP probably will do little better on the flying start as it tends to stutter from a stop on the flywheel (note again never on the track). I would expect exactly the same peak power as this does not seem to be affected much by timing changes. The power band is only shifted.
Well all I can say is try one or two running starts at some arbitrary RPM (if it cannot be measured) and see what happens. If it compares to the data so far then that's that. There's a lot going on in this firmware, probably more than we realise. I think a 'soft start' is pretty much a given for a modern brushless controller, and that is designed with normal driving in mind, not spinning up flywheels. How the ESC is transitioning from this 'soft start' to normal operation and then onto the higher RPMs where increasing levels of forward advance are used -- they could be using more feedback or making more assumptions (which may not be true in the dyno case) than we realise. You can do a lot more with brushless ESCs (variable timing advance, RPM feedback, etc.) than with the brushed ones.

Have you raced with one of the new 'ballistic' motors yet? How do you like them?
Access is offline   Reply With Quote
Old 05-07-2009, 01:17 PM   #131
Tech Elite
 
Join Date: Aug 2002
Location: Houston, Texas
Posts: 3,777
Trader Rating: 27 (100%+)
Send a message via AIM to John Stranahan
Default

Battle of the Big Dogs

The results are in. I used the Sentry dyno to compare the just released Novak Ballistic 3.5 with the recently released LRP X12 3.5. I used the SMC 6000 mA-h 28c.

Manners of the LRP motor on the dyno were improved. There was no stuttering with this better battery. It may be able to deliver more early amps causing smooth motion. The Novak motors were always completely smooth. I have my voltage lead working this time. I charged my pack to 8.25 V. The resting voltage just before a run was 8.214 V. My voltage reading agrees with Matt's Sentry to the .001 place.

I ran several test. First an early Novak Velocity 3.5 motor that I had hooked up but had not tested previously on this battery. Followed by the Novak Ballistic series 3.5 Followed by the LRP X12 3.5 all run with the GTB which was controlled by the servo tester control.

The Results
Clearly the LRP X12 has more power than the Ballistic Series Novak 3.5 R. 699 W for the Nova 760 W for the LRP. See the first graph.

If you look at the efficiency line farthest to the bottom, the LRP in red is slightly more efficient near 40,000 RPM. It may be giving you more power without a gain in heat and ampdraw. It does make this power at a higher RPM and I found on the track it needs to be geared quite low. 10.6 overall in a TC.

I crown the LRP X12 the king over this pair.

The new Results
And then I plotted up the Novak Velocity (early) 3.5. 800W. So I have a new king. See the second graph. The word early needs some explanation. I have owned 3 Novak velocity series 3.5. The first had huge power. You could hear strong electrical induction noise on the straight as the pan car revved up to Nitro speeds. I eventually after long use burned this one up. I bought a second. Not so strong. The noise was more subdued. The power was just off a little. I suspected Novak had made a running change to reduce amp draw and reduce the strain on the GTB. I wrote about this in my Pantoura Thread. I noted the new motor would do better on a slippery track. Well soon after, Novak came out with their light blue series of velocity motors. Just detuned with smaller wire I would guess. I sent my burned up 3.5 in for rebuild and it came back re manufactured and was identical to the original. I was so pleased. And here is that re manufactured motor kicking serious ass a year or two later. I don't know if you can buy this early version any more.

This Early version Novak 3.5 motor motor is the King of all my test so far. In the pan car it is just superior. In the TC it may overheat and lack early exit punch.
Attached Thumbnails
Dyno, Homemade, Using a Novak Sentry Data Logger, Continued, The Experimental Thread.-novak-ballistic-3-5-vs-lrp-x12-3-5-smc-6000-28c001.jpg   Dyno, Homemade, Using a Novak Sentry Data Logger, Continued, The Experimental Thread.-novak-early-3-5-smc-6000-28c001.jpg  
__________________
[email protected]
TeamCRC.com Mikes-HobbyShop.com

Last edited by John Stranahan; 05-07-2009 at 01:28 PM.
John Stranahan is offline   Reply With Quote
Old 05-07-2009, 01:38 PM   #132
Tech Adept
 
Join Date: Mar 2009
Location: Metro New York
Posts: 139
Default

Current Limiting in the Electronic Speed Control

The principle of current limiting in a brushless phase motor is discussed in this article:

http://powerelectronics.com/power_ma...ses/index.html

but I'm not sure how different software/firmware profiles apply these principles. In a robust hardware design the MOSFETs or IGBT transistors must be self-protected for maximum voltage, current, and operating temperature limits. This circuitry indirectly "senses" the heavier flywheel load since it will try to draw more average current, and trigger self-protection. The equation below does not show it, but maybe the flywheel has an impact on resonance so Driver's throttle control has two challenges, resonance for some loads and no control over the "robot" current limiter.

Notice Equation (2) is duplicated below and Figure (2) is attached.

I = ((V - Vbe)/R)*(1 - e^(Rt/L))

see the article for details.

When throttle is maximum it calls for a longer pulse width "input" to the ESC, raising the average DC voltage V. When shaft speeds are small, generally below the rpm of the peak power point, the reverse voltage Vbe is small and current draw is large. Add more inertia to the system and the large starting current wants to flow for more time, since the mechanical time constant is greater.

The ESC self-protects by cutting back the pulse width as shown in the bottom of Figure 2 to keep transistor current from over-shooting the rated current. I think John is getting some low current readings when flywheel is just starting in some Dyno pulls, and I think the DC voltage into the ESC should also be smaller, since Vbe is strictly a function of shaft speed and k.

The Tesla Roadster is essentially a giant brushless single speed car, but no magnets in the motor, it uses rotor coils (AC induction motor).

The old torque-speed curve is published here:

http://www.teslamotors.com/performan...and_torque.php

this curve looks just like one of the constant current-limited Roby Dyno curves posted on this site, with torque constant and power in a straight line ramp, then a dog-leg down in torque and power transitioning to a parabolic.

The discussion of the electro-mechanical driveline improvements to get 30%more shaft torque from better IGBTs is discussed here:

http://www.teslamotors.com/blog4/?p=67

Power Electronics Module (PEM) Update
Attached Thumbnails
Dyno, Homemade, Using a Novak Sentry Data Logger, Continued, The Experimental Thread.-pet-current-limiting-figure02-1205.jpg  
SystemTheory is offline   Reply With Quote
Old 05-07-2009, 02:24 PM   #133
Tech Fanatic
 
Join Date: Jun 2006
Posts: 872
Trader Rating: 35 (100%+)
Default

John, if you ever get a chance to test another Ballistic 3.5, please do. I have noticed a slight difference between Novak 13.5 motors myself, as much as 2000 rpms!
__________________
High Desert Raceplace, Grand Jct CO, Sanwa Exzes Plus Stick Radio, Spektrum,Thunderpower 230g, Modified Tri-Nut Novak GTB2, Ballistic 4.5t 550, 5.83lb SCTE Ten, Novak Sentry Brushless Dyno, Crossweight Setup Station, Junsi 20A power supply Icharger 20A Charger,TP610C, 22B, Novak Edge, Novak 13.5, SC10, Havoc Pro SC XDrive, Ballistic 17.5 Matthew Joseph Cordova
mattnin is offline   Reply With Quote
Old 05-07-2009, 04:18 PM   #134
Tech Rookie
 
Join Date: May 2009
Posts: 19
Default

Quote:
Originally Posted by SystemTheory View Post
principles. In a robust hardware design the MOSFETs or IGBT transistors must be self-protected for maximum voltage, current, and operating temperature limits. This circuitry indirectly "senses" the heavier flywheel load since it will try to draw more average current, and trigger self-protection. The equation below does not show it, but maybe the flywheel has an impact on resonance so Driver's throttle control has two challenges, resonance for some loads and no control over the "robot" current limiter.
For 'soft start', my guess is that not every ESC has the necessary built-in or dedicated current sensor(s) to do this; they are doing it by using RPM feedback (ie. at this RPM range, only this much PWM duty cycle is allowed), or a simple timer / forced throttle ramp. Or they just depend on battery dropout to limit the current. Some of the industrial ESCs actually have a 'current limit' parameter you can set based on what your system can handle, but I've never seen this in a hobbyist ESC.
Access is offline   Reply With Quote
Old 05-07-2009, 05:28 PM   #135
Tech Rookie
 
Join Date: May 2009
Posts: 19
Default

Quote:
Originally Posted by John Stranahan View Post
Battle of the Big Dogs
Do both motors have the same (fixed) forward advance? Can this be changed mechanically for both motors?

Is it possible to graph efficiency either independently or on the right axis (independent scale) so it is not so compressed? It looks like wherever efficiency is better (for either motor), power is also better, or some approximation thereof.

Last edited by Access; 05-07-2009 at 05:39 PM.
Access is offline   Reply With Quote
Reply

Similar Threads
Thread Thread Starter Forum Replies Last Post
KARZ Calgary xrayroooahhhh Canadian R/C Scene 19114 12-01-2017 01:26 PM
CRC Battle Axe, GenXPro 10, 1/10th pan, Brushless, Lipo,4c, Road, Oval,TipsandTricks John Stranahan Electric On-Road 1776 03-01-2016 01:18 PM
Novak Sentry as Brushless Dyno mattnin Electric On-Road 132 01-29-2010 05:53 PM



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off



All times are GMT -7. It is currently 04:40 AM.


We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.
Powered By: vBulletin v3.9.2.1
Privacy Policy | Terms of Use | Advertise Content © 2001-2011 RCTech.net