Sounds correct but what is the difference between the old and new methods? The old one worked AND had limits in the regen current. What did the old firmware do to the rest of the energy and why can’t the new one do the same thing?
I’m quite certain I’m going to be corrected on this one but I feel like the old FW just used the min(battery_min, motor_min) as braking current limit, no matter duty cycle. Therefore braking would be linear.
Battery min was there to set a limit. So when this limit would be exceeded then the brake current is reduced. So at high speeds and a low value the brakes would still be weak but still would work at the battery min limit. And once the duty cycle is low enough it gains brake force and then works on a constant level.
To achieve the battery min brake limit in the new software you need to do a full brake. But when the duty cycle get’s too low you quickly brake too hard and the brake strength varies the entire time while the trigger is stable. So you are constantly trying to find the right mount of brake with your finger.
Long story short, the old solution fro braking was better and it is easy to fix in the new firmware if Benjamin decides to do so.
2 Likes
Funny thing, in the prototyping or thinking phase I would say that current implementation makes much more sense and it is the correct one. But the old version is much better in practice.
I’m going back to the old firmware, the new one has too many issues, the unpredictable braking and overcharging the battery are the most important things, the next one is “R is 0” which won’t allow me to run FOC on my Focbox.
This FW might work in the future but for now, it’s not ready
2 Likes
WOW just WOW 
just updated firmware from version 2.18 to version 3.28 on dual 4.12 hardware VESC’s on an TRAMPA using the VESC-Tool for the first time.
Setup was a breeze!!! Hover over virtually anything or look for a question mark and a help box appears!!!
-New firmware HAD to be updated before VESC-tool would allow full connection to the VESC’s. No problem, just BE SURE YOU KNOW WHICH HARDWARE VESC YOU HAVE, mine was 412…
-Ran motor setup wizard, Breeze!! Help boxes appear often to remind you what parameters are being set and what to prepare for and expect(spinning motors, keep clear etc. SO helpful!) -Ran controller setup Wizard, which asks whether the current VESC is a master or slave! So easy!! No controller setup was required for the slave!My thanks go out to BENJAMIN VEDDER 
and ALL who made this happen. 
A great thing got a lot better.
I run exclusively FOC for its quiet nature. So I cannot compare any differences in BLDC or sensorless. Also this is my opinion, since no real world figures were recorded.
Running just the Wizards for setup, using the same settings I ran before, Immediate differences were noted.
- More powerful Areas like light brush over 2ft was previously giving me trouble. Also slow starts uphill were difficult to finesse, slowing me until the motors were churning to a standstill. With the update, I tackled brush I would have turned away from, and Ive tackled some huge mounds of grass so far, creeping up them slowly to immitate an unexpected hill climb. This, as well as the following will help a great deal with the real rough trails I like to bash!
- Less cutouts.
Both VESCS previously experienced a timeout when jumping and holding the throttle. Also on fast street startups and fast sidewalk bumps. No more! Over 15 miles today on the 3.28 update and did not experience one timeout. This was plaguing my funtime and thought I just needed more HG2’s onboard…trailriding was causing major cutouts so I am excited to get padded up and head for the real rough stuff thats gives the belts hell
Finger temperature readings were as usual after long 33km/h stretches (this seems to produce more heat than anything else I do) VESCS are running room temp as always and motors were quite warm to the touch as always.
Trampa is holding Benjamin hostage. Frank is the only way he can communicate with us.
1 Like
Yeah, he is in a secret room with no window.
Drop your comments on the Vedder forum, easy. www.vesc-project.com Benjamin can’t get lost in conversation on multiple forums. There is simply no time for that.
Frank
That is what you experience in any brake, car, bike, motorbike. If you apply 50% brake at high speeds, braking is appropriate, while 50% at low speed is blocking your wheels. Kinetic energy is high at high speed and low at low speeds. In consequence the same brake power works against different levels of kinetic energy. When you use the VESC in an automated environment this allows precise computation of your brake force.
When you brake down your car, you release the brake pedal bit by bit while you slow down to get a constant negative acceleration. Battery min/motor min now set the absolute strength of the brake at 100% breaking, which is the correct way to handle it because that is what you’d expect it to regulate. If you want to tweak it, you can use the curves to adjust the behavior of the brake. The brake is now more linear (technically), while the throttle curves allow you to de-linearize it to your needs. Otherwise you’d deal with two curves, which would be a bit funny and not predictable, nor precisely adjustable.
All you need to do is set up the values correctly and use the throttle curves to get your desired braking behavior.
Frank
FYI: I’ve just sent Benjamin a FOCBOX with the R=0 error for him to investigate.
Cheers
6 Likes
I don’t want to start a argument here because my intention is to improve the controllability and safety.
But you are wrong. Also for the car. A constant pressure on the brake pedal results in a constant deceleration (negative acceleration). And for the VESC a constant brake torque (amps) results in a constant deceleration.
4 Likes
Exactly. When I’m driving my car and depress the brake pedal, the deceleration is constant. Lifting my foot off the brake pedal would cause my car to decelerate less and less…
@trampa
You know that I always supported you and Ben all along this project.
But these last days/weeks, I can’t really understand your reaction.
I mean, we are involved in an open source project which means that people can bring contributions.
These may be bugs correction, features coding or simply beta testing.
Now we receive feedbacks of many people having issue with brakes.
That’s a statement, and it’s kind of a serious stuff as it makes them going back to FW2.xx (and keep them away from VESC Tool).
But another contributor @Ackmaniac figured out what causes this issue.
Perfect Open Source spirit, isn’t it ? 
So why do you hide behind theory (and as an energy engineer, I can say that it’s even not true) instead of simply take it into account and release a beta FW with Ackmaniac’s proposal just for testing ?
If it’s worse, back to original and you’re right.
If it’s better, everybody will be happy to enjoy VESC Tool again and will thank you for having taken into account their request. 
4 Likes
That is all possible Pimousse, but the best way to to contribute is to write something about it on the vesc-project.com forum, or if your skills allow it, do a pull request and suggest some alternative code. Benjamin is happy about it and will look at it and respond and debate it! It’s can take a while sometimes, so patience is required.
http://www.vesc-project.com/node/110
I just try to explain why people face weak brakes and Ackmaniac did the same. Sure it’s all debatable and hints are taken serious. If you understand the matter, it’s easy to adjust your settings appropriately. The VESC-Tool is being used in all sorts of environments, not only skateboards. It’s multi purpose use. Skateboarders love this, robotic guys want that. In consequence Benjamin is aiming at technically correct solutions, which allow the same adjustments of parameters as before, but in a more sophisticated and precisely adjustable way. People have different demands. You want it this way, another user wants it that way. The only way forward is the technical correct way allowing appropriate settings for all/most demands.
There is a square relation between kinetic energy and speed! If you apply the same brake force at higher speeds, you gain a lot less negative acceleration (brakes feel weaker) because you work against a lot more kinetic energy. Same for acceleration. You can now adjust that behaviour through the use of a curve to fine tune to the desired brake force/acceleration at a certain speed. So you can now compensate things actively, which wasn’t possible like that before. You set your max brake strength first and then you adjust the curve to the desired response over the range of speed you usually travel at. So you now actively de-linearise it to your needs (which might feel more linear to you personally), which gives you better control over your brakes. Options are: ploynominal, normal and exponetial. You just need to know the feature and use it accordingly and experiment a bit.
Benjamin already split the throttle curves for brake and acceleration, so you can have a very soft acceleration, while adjust for a lot harder brakes. I had no issues to find sweet settings. He did that because it was requested by multiple users and because it makes sense.
Anyway, we will do some test runs next weekend and will debate it all, and maybe it results in a FW/software tweak, additional feature etc. Benjamin wants to make some videos about correct setups, which will be the way forward for many users. He aims at perfection and that will be the result very soon. So your input is important and in consequence there is a ToDo list online, so you can see that your hints/demands are on the list and being worked on.
http://www.vesc-project.com/node/110
I don’t want to fight anyone’s ideas, it’s about understanding the matter. Still opinions can vary, which doesn’t mean that your opinion isn’t taken into consideration.
Frank
And that is where you are wrong. Because with the paddle at a car you apply pressure to the disc brake. And the faster the disc rotates the more energy is transfered into heat. For electric motors it is the same. The faster the Motor rotates the higher is the voltage and with a constant current it results in higher watts.
But I am sure Benjamin understands and will identify this as an issue as well. And it also doesn’t make any sense for robotics. There the non linear throttle could be a issue but not a linear brake.
1 Like
Thank you for your answer, it reassures me.
However, I’d like to raise your attention on an important point. You said :
You’re now selling VESC6 through your shop.
I can tell you that some of your customers even don’t understand the difference between batteries in series or in parallel.
For my part, on the french forum, it consumes A LOT of time to help them, although I wrote tutorials and so on.
So you definitely can’t assume that people will figure out which parameter they have to “cheat” to reach the expected brake behaviour. 
VESC Tool is a huge improvement thanks to wizards to help novice getting started with VESCs,
Keep it intuitive af. 
This is why video tutorials are the way forward. If you know the FAQs, you can answer the questions.
Frank
1 Like
If only a day can count more than 24h, I would be pleased to do so.
ATM, every answer to a novice question is added to the tutorial.
But it’s already complicated to explain why you put this in motor max, that in battery min, etc.
So if now you have to say “Ok, your battery can handle 20A in regen but there is a trick for having usable brake, so put 34A instead…”.
What a mess. 
But that’s at least the truth. Your battery has to swallow what you shoot towards it. So a builder can make the decision to go for higher P counts and higher S counts, better cells etc.
Frank
1 Like