10S custom skate ESC: testers wanted!
- Thread starter austindavid
- Start date
Hi Benjamin,
Asking me questions when waiting for the VESC boards ...
Do you think the FET would benefit from having some more cooling ?
Maybe tying them to some copper or aluminium bar which would be to a aluminium profile ?
I see most of the other VESC are cooled and yours is not.
Is this something to be worried about ?
Asking me questions when waiting for the VESC boards ...
Do you think the FET would benefit from having some more cooling ?
Maybe tying them to some copper or aluminium bar which would be to a aluminium profile ?
I see most of the other VESC are cooled and yours is not.
Is this something to be worried about ?
Slithr Boards
10 µW
- Joined
- Feb 26, 2015
- Messages
- 6
Vedder,
I was wondering what your coefficient of friction was between the longboard wheels and the asphalt.
I am trying to calculate the necessary torque needed at certain grade hills, and this parameter, the coefficient of friction, has been uncertain.
Any thoughts?
Carl
I was wondering what your coefficient of friction was between the longboard wheels and the asphalt.
I am trying to calculate the necessary torque needed at certain grade hills, and this parameter, the coefficient of friction, has been uncertain.
Any thoughts?
Carl
akiraEC said:
AkiraEC,
From the different setups shown I think it will depend of the use you make of it. It is efficient so it won't warm as much as a standard RC ESC but the more amps you want to draw, the more current you're bound to use and then the more cooling solutions you will need.
I'll go from a visual estimation from pictures and say that till 6S you can go with a closed case as long as it has spare room between the PCB and the Cover. Past 6S you will want a fan plate or a true fan like in a computer.
Slithr Boards said:
Carl,
I think this is not relevant to Vedder ESC, thought somebody here gotta have the answer. It is relevant to your truck and driving system models, plus the drag of your motors. Of course you can add the quality of your bearings and wheels, the grip rating etc...
There is a lot of factors to take in consideration to know how to calculate the friction coefficient.
Now for hills the most important thing will be the motor you want to choose and the amount of power you will feed it with. You don't want to load a 1000Kv motor the size of a longboard wheel uphill.
At stand start and at say 20mph coefficient drag is negligible too. You will want around a 200Kv or 100Kv motor to climb without fear. You will also want sufficient voltage for voltage gives you torque.
For a good example to follow look at the DIYers here and also the commercial boards, they have good examples of boards.
onloop
10 kW
- Joined
- Jun 24, 2013
- Messages
- 600
Vanarian said:
i think as you go up in voltage there will be lower current & less heat, so likely less need for fan/cooling.... based on vedders video/testing i don't think a fan or large heatsink is required (i think he mostly use 10s or 12s)....but i guess we need to wait & see...
And of course it would not hurt to have it..
chuttney1
100 W
Vanarian said:
While it may not be necessary, I still think doing this type of calculation to factor in friction can help with calculating mechanical losses to help give a more accurate prediction of calculated speed using the equations provided from users on this forum when building an eboard. It can also help us learn what is the maximum KV value for a motor if the diameter of the motor's can did not matter. What is the maximum hill grade given a motor's size and KV and weight of the user?
Slithr Boards
10 µW
- Joined
- Feb 26, 2015
- Messages
- 6
Vanarian said:
The friction coefficient between a longboard wheel and asphalt is a constant and does play a huge factor in the torque calculations for hill climbs. Yea there might be friction from the bearings in the wheels, drive train, and the motors, but since vedder has done the calculations as well as the testing to choose the right motor, then design his esc, he might have a good idea for the coefficient of friction I could use for my calculations. I was thinking around .25 ...
Carl
onloop said:
Not sure about this. For a defined limit of power, if you go for the highest possible voltage and lowest possible amperage I totally agree with you, you build a lot less heat and you win efficiency. But we have limited options about V x A ratio on e-longboard sized parts, i.e. even with an ESC capable of 72V where do we find a 50mm or 63mm motor which can accomodate this much voltage? This is not widely available. Plus ESCs sized for longboards are more matched for 10 or 12S like Vedder is doing.
Let's say I aim for 2000W only for a motor rated for bigger power so I have room to be always safe. At 6S (22.2V) I need 90A. Not really the best choice. I go straight for 10S (37V), I will only need 54A. Now we're talking a better V x A ratio, thus you are definitely right, less heat comes up with the bigger voltage.
But since there are many guys here like me who will aim for something more dangerous, let's choose 3200W power for one motor. Not many options become available because it is a bigger figure, motors sized for 50mm or 63mm have limited rating as far as maximum voltage and amperage go. For 10S (37V) I need already 86A to reach my 3200W goals, if I can go 12S (44.4V) I can grab some more efficiency and only need 72A. But I won't be gaining more than this because longboard parts are sized for 10-12S max.
In this last case not much can be made, we're still close to the 90A figure and we added more Volts on top of that. Just because I wanted bigger power and I cannot get a better V x A ratio, I'm bound to get more heat than say somebody who only want 2000W. Another variable is to go past the manufacrturer's ratings of a motor, since heat has less impact and causes less losses it could do the trick. But then the question is : will the motor be able to support the iron losses if it has low amount of iron connections inside?
But as long as the VESC goes, you are right, it does seem to manage heat very well
And who does ride his board to its maximum peak power all time anyway?chuttney1 said:
This question is hard to answer. To be able to give you a more or less accurate answer, I would need to estimate the precise minimum power amount needed to climb the steepest hills at a defined speed, maybe at least 5 motors wound from 50Kv increments with the same power limit, ideally one board where I can swap the motors and finally do a meeting with many testers from different weight (say ranging from 50 till 100 Kg). With all these I could be able to tell you what Kv rating and what Voltage/Amperage ratio one would need depending of his weight.
If any one of you want to give it a try, You should know that a human in "sprint" effort can provide 400W if he's in athlete condition, 600W if he's a champion. So for the lowest test figure I'd go for 800W minimum.Slithr Boards said:
It does help to get more accuracy but there are two limiting factors in both the quality of numbers you can obtain from the friction coeff and also real benefit from estimating it :
-Friction coefficient is not a constant because the ground itself is not constantly equal. During your travel you will meet different materials under your wheels. For every polyurethane models this will be a major drawback in friction coeff. For rubber tyres ones, this could be more linear due to a better absorption of irregularities. Unless you decide to go for dirt grounds. Again, friction coeff highly variable.
-Real life use. Everybody here agrees to say that 30mph on a longboard is already dangerous, not many riders commute at this speed actually because there are too much risks of meeting obstacles like walkers, cars, etc... Only performance users will go faster than this. Even as a performance user, once you get a first estimation you will give more credit to real life tests to know more or less what you will achieve with your board. So in the end at lower speeds, even when climbing hills, you won't care about the coeff friction.
What I mean is you can't really rely on friction coeff for daily usage because there are too many variables, plus the efficiency of electric systems is so that as long as you have proper manufacturing of your wheels, bearings, drivelines and motors, you will counter the friction. Only best next thing is 0 friction drive, with 0 wheels for example.
In my opinion you will definitely get better results if you calculate your body's drag coeff instead. You stand on your board, your body provides more resistance to wind than your board will ever produce friction with the ground. And this measures constantly, be it for daily or performance use. :wink:
onloop
10 kW
- Joined
- Jun 24, 2013
- Messages
- 600
Honestly I have no idea what you guys are trying to work out exactly... or why you would even need to know this to build an electric skateboard... Maybe useful for F1 cars... Also i assume there is no precise answer to this question because there are too many variables. Namely the surface quality of road & the density of the urethane..
If a particular setup doesn't get you up the hill do this.
1. Increase voltage
2. Change gearing ratio, keep dropping motor pulley teeth until you get to the top.
3. Increase teeth on wheel pulley.
4. Add additional motors.
5. Decrease wheel diameter.
6. Go lower KV motors.
7. Lose weight.
Fu*king around with formulas ain't got nothing on real world testing.
If a particular setup doesn't get you up the hill do this.
1. Increase voltage
2. Change gearing ratio, keep dropping motor pulley teeth until you get to the top.
3. Increase teeth on wheel pulley.
4. Add additional motors.
5. Decrease wheel diameter.
6. Go lower KV motors.
7. Lose weight.
Fu*king around with formulas ain't got nothing on real world testing.
torqueboards
1 MW
Yeah, what's the point. You couldn't even stay on the board instant full throttle on a 6kw or 8kw board. Nobody would even stay on the board.
The torque limiting feature is win win for skateboards. There is no point for too much of it with tiny wheels from a dead stop just for climbing hills or achieving better acceleration at higher speeds, if you do not fall off even then from it.
As for the heat I don't think vedder has yet tested the new fet's that run cooler then the previous fet's he used last year.
Ben, is there any kind of guide for learning how to program code for the esc? I would really like to try to learn.
As for the heat I don't think vedder has yet tested the new fet's that run cooler then the previous fet's he used last year.
Ben, is there any kind of guide for learning how to program code for the esc? I would really like to try to learn.
This is a long reply that I wrote in a short time, so I hope that it isn't too confusing.
The current makes a huge difference and the voltage, kv and gearing decide what current you need for a given amount of torque. Since there is a square relation between current and heat generation, running at for example 80 A produces 4 times as much heat as running at 40 A. So if you have a given setup designed for a given top speed with 6s, changing to a motor with half the kv and going to 12s will give the same performance and produce only a quarter of the heat in the ESC for the same operating conditions. This is why I encourage to go for higher voltage and lower KV. A well-designed (motor kv, gearing) 10s-12s single motor setup with a top-speed of 35-40 km/h with VESC will not require additional cooling even if you climb hills, but a setup with 6s and the same top speed will require additional cooling for sure since 4 times more heat is produced in the FETs. Notice that the top speed is an important factor and also notice that the top speed I'm talking about is at the maximum speed the motor can reach with the available battery voltage (running high voltage and only using half throttle does not help at all). If you design for a top speed of 20km/h or less, you can probably get away with 6s and no cooling.
Even if you have a board with very high theoretical power output, the top speed is what decides how much torque you need for a given acceleration, and the gearing and motor kv decides how much current you need to generate a given torque. If you have a 6kw setup designed for 40 km/h, you will have a very high acceleration at full throttle, but there is no way that you can draw that much power continuously (meaning that the high current and thus high heat generation only come for a couple of seconds at most). The only way you can output that much power continuously is to design for a very high top speed, and then you need high voltage and/or dual motors and/or cooling. It is also important to understand that if you gear for high top speed, the losses will be higher even at low speed since current is what causes the heat, and at low speed you will have higher current and lower motor voltage for a given torque than you would with gearing for lower top speed.
My friend has done some calculations with wheel friction and air resistance, but the numbers I gave above are based on me actually testing the board. Since the heat generation depends so much on the current and the current is very difficult to estimate for an "average" ride, calculating accurate numbers is difficult. Also, when it comes to thermal properties, it gets even more difficult. An uneven load profile with a certain average power output will produce more heat than an even load profile with the same average power output because of the square current-to-losses relation. Also, when the FETs get warm, their resistance increases and causes even more losses for the same current. Another thing is that the higher the temperature difference between the pcb and the ambient air is, the more heat will be "radiated away". Yet another thing is the influence of cooling: just a little bit of airflow makes a huge difference, so maybe a small fan or air canals in the ESC box will be better than trying to add a heatsink. This is why I think onloop is right: experimentation and knowing the factors that you can change is the best way to go. Regarding the factors that can be changed, the current/voltage/top speed relation is very important to understand.
I wrote a post about that here:
http://vedder.se/2014/10/chosing-the-right-bldc-motor-and-battery-setup-for-an-electric-skateboard/
In that post I give an example where you decide for a given top speed and adapt the other system parameters accordingly to get a well-matched setup. I really suggest that you read it. When I have time I will write a post with some more calculations (friction, hill grade etc) to make a more theoretical analysis, but I don't know how useful it will be. Real-world experience and knowing how to match the parameters of the system to get the best possible performance under the given conditions will be more useful.
Apart from the DRV8302 missing, it looks like the wrong shunts are used and that the sense pads on the shunt footprints are not connected at all. This will give either zero or a random value for the current readings, so as soon as you connect a motor and battery and try to run the motor, the ESC will blow up for sure. If these wrong shunts are fine and have the correct resistance it can be patched easily with a small piece of wire, but the current reading accuracy will definitely be affected by this.
Since it looks like they tried to save money on the shunts, I wonder if they also replaced the ceramic capacitors with cheaper ones. This can potentially limit the maximum operating voltage (if the voltage is lower than I specified) or make the ESC reboot all the time and/or die (if the capacitance is lower).
That is not true at all.
The current makes a huge difference and the voltage, kv and gearing decide what current you need for a given amount of torque. Since there is a square relation between current and heat generation, running at for example 80 A produces 4 times as much heat as running at 40 A. So if you have a given setup designed for a given top speed with 6s, changing to a motor with half the kv and going to 12s will give the same performance and produce only a quarter of the heat in the ESC for the same operating conditions. This is why I encourage to go for higher voltage and lower KV. A well-designed (motor kv, gearing) 10s-12s single motor setup with a top-speed of 35-40 km/h with VESC will not require additional cooling even if you climb hills, but a setup with 6s and the same top speed will require additional cooling for sure since 4 times more heat is produced in the FETs. Notice that the top speed is an important factor and also notice that the top speed I'm talking about is at the maximum speed the motor can reach with the available battery voltage (running high voltage and only using half throttle does not help at all). If you design for a top speed of 20km/h or less, you can probably get away with 6s and no cooling.
Even if you have a board with very high theoretical power output, the top speed is what decides how much torque you need for a given acceleration, and the gearing and motor kv decides how much current you need to generate a given torque. If you have a 6kw setup designed for 40 km/h, you will have a very high acceleration at full throttle, but there is no way that you can draw that much power continuously (meaning that the high current and thus high heat generation only come for a couple of seconds at most). The only way you can output that much power continuously is to design for a very high top speed, and then you need high voltage and/or dual motors and/or cooling. It is also important to understand that if you gear for high top speed, the losses will be higher even at low speed since current is what causes the heat, and at low speed you will have higher current and lower motor voltage for a given torque than you would with gearing for lower top speed.
My friend has done some calculations with wheel friction and air resistance, but the numbers I gave above are based on me actually testing the board. Since the heat generation depends so much on the current and the current is very difficult to estimate for an "average" ride, calculating accurate numbers is difficult. Also, when it comes to thermal properties, it gets even more difficult. An uneven load profile with a certain average power output will produce more heat than an even load profile with the same average power output because of the square current-to-losses relation. Also, when the FETs get warm, their resistance increases and causes even more losses for the same current. Another thing is that the higher the temperature difference between the pcb and the ambient air is, the more heat will be "radiated away". Yet another thing is the influence of cooling: just a little bit of airflow makes a huge difference, so maybe a small fan or air canals in the ESC box will be better than trying to add a heatsink. This is why I think onloop is right: experimentation and knowing the factors that you can change is the best way to go. Regarding the factors that can be changed, the current/voltage/top speed relation is very important to understand.
I wrote a post about that here:
http://vedder.se/2014/10/chosing-the-right-bldc-motor-and-battery-setup-for-an-electric-skateboard/
In that post I give an example where you decide for a given top speed and adapt the other system parameters accordingly to get a well-matched setup. I really suggest that you read it. When I have time I will write a post with some more calculations (friction, hill grade etc) to make a more theoretical analysis, but I don't know how useful it will be. Real-world experience and knowing how to match the parameters of the system to get the best possible performance under the given conditions will be more useful.
Apart from the DRV8302 missing, it looks like the wrong shunts are used and that the sense pads on the shunt footprints are not connected at all. This will give either zero or a random value for the current readings, so as soon as you connect a motor and battery and try to run the motor, the ESC will blow up for sure. If these wrong shunts are fine and have the correct resistance it can be patched easily with a small piece of wire, but the current reading accuracy will definitely be affected by this.
Since it looks like they tried to save money on the shunts, I wonder if they also replaced the ceramic capacitors with cheaper ones. This can potentially limit the maximum operating voltage (if the voltage is lower than I specified) or make the ESC reboot all the time and/or die (if the capacitance is lower).
vedder said:
This is quite worrying.
Will the boards be checked by anybody who has the knowledge before beeing shipped ?
Wouldn't it be better (if he agrees) to let Benjamin check that the implementation is correct before sending the board to everybody ?
If the assembler messed up something, it would be better to know it (and to return the boars) before the bards are dispatched.
What do you think ?
onloop
10 kW
- Joined
- Jun 24, 2013
- Messages
- 600
Please note: This is an unfinished, part assembled board the board in the photo is currently being bench tested by the electronics engineer.
As ben has mentioned & as you can clearly see there are some things that are not yet finished Or present.
As i understand you will all appreciate these things don't magically appear from thin air.... this item gets built from parts that must be sourced from several different suppliers from across the globe and assembled together with many complex processes taking place. This project has taken many hours of (unpaid) work from many people to get to this stage....
these photo have been provided so that VESC BETA investors can see the project is moving forward & prove that it is being worked on each day.
Please rest assured that all the vesc boards will be bench tested and have firmware loaded Before shipping.
As ben has mentioned & as you can clearly see there are some things that are not yet finished Or present.
As i understand you will all appreciate these things don't magically appear from thin air.... this item gets built from parts that must be sourced from several different suppliers from across the globe and assembled together with many complex processes taking place. This project has taken many hours of (unpaid) work from many people to get to this stage....
these photo have been provided so that VESC BETA investors can see the project is moving forward & prove that it is being worked on each day.
Please rest assured that all the vesc boards will be bench tested and have firmware loaded Before shipping.
vedder said:
Thank you for the correction, now at least this is precise explanations on what to account for when you calculate heat.
I have a misconception in that department, I tend to think only of battery / peak power relations whereas you need to estimate everything
my laziness got me. Can you confirm that a simple airflow tunnel can do the trick ? If so I could just funnel the air through my boots under my feet and stick the VESC inside the base of the boot
Two birds killed with one stone, VESC is protected from collisions and also some space is gained.agraham
10 W
vedder said:If you start going downhill and brake a lot on a full charge, the battery can take damage. There is not much to do about that. In fact, it is difficult to make non-regenerative braking on electric motors
How about pulse the braking around 5 Hz in this case. That will cut the damage to the battery to less than half and still give the user half their braking power so they can avoid going under that bus. The pulsing will serve as a message to the user: "please turn around and go uphill for a minute to burn off some charge".
I know it's very unlikely that someone would take a fully-charged board to the top of a mountain and bomb down, but if someone ever did it would be nice if they could survive. I do live at the top of a two-block long hill with a major road crossing at the bottom and when I get on my board fresh off the charger and go down that hill I can't help thinking "what would happen if my braking cut out right... NOW"
Stielz
100 W
Awesome stuff! are you guy still looking for testers?
Ive got a dual Turning SK 6374 powered longboard collecting dust cause my Hobbyking 150A sensored ESCs died in firery death so I'm in the market for some proper skateboard ESCs
Edit: just found the link to the beta membership website
Ive got a dual Turning SK 6374 powered longboard collecting dust cause my Hobbyking 150A sensored ESCs died in firery death so I'm in the market for some proper skateboard ESCs
Edit: just found the link to the beta membership website
chuttney1
100 W
Look who got some PCBs made from Hackvana. The quality from Hackvana exceeded my expectations and very well made, though this is my first time to order from this fab house. I ordered 10 and got 1 extra. Uses 2oz of copper. And Mitch from Hackvana is super helpful and awesome. 
