Alt. firmware and GUI for Thanos's AMC1.5 & AMC1.6

[prilad]

Hi, Speedy.

I think, I understand what you mean. And let me give the following arguments:

-- 1 --

that is the start up inertia of the motor ... time consumed for the fastest motor in starting from zero to it's full RPM is [in hundreds milliseconds] ... that's very very slower than a digital pulses ...

Yes, that is true... but...

simply that motor doesn't need that accurate pulses because it can't follow them up ... the motor is blind ... he can't see them ... he just see the overall result of them ...

... This is not true. Motor in this case works as an integrator, and not react to every single pulse. And in this case for motor better higher PWM frequency. Good example for this case - TV or Computer monitor. The human eye is also integrator. But if you set monitors refresh rate less then 50Hz, you can see the "blinking" screen.

-- 2 --

from the motor side of view ...8-bit is better than 10-bit ... in term of accuracy ... simply the motor doesn't need that accurate pulses because ... it can't see them actually ... follow them up precisely ...

...Let's make brain experiment and bring it to the absurd. You say that 8-bit is better than 10-bits. Let's go further - why not 6-bit? It's better than the 8-bit .... and so we get to the 1-bit PWM. Very simple mode - set motor ON if it not in position, and OFF when it at destination position

What do you think now?

I'm again saying that the for motor 8-bit PWM is not better than 10-bit PWM. But my English is not so good for an explanation of the thought.

Regards, my friend
Ale

[tronicgr]

http://www.x-sim.de/forum/viewtopic.php ... t=20#p2456

Alhough it has to do with refresh frequency, the same applies to the PWM "RESOLUTION" !!!

The less PWM resolution you are using the more "STEPPING" you get trying to set correct speed for the motor. And correct speed is very important in case you are using PID algorithms for POSITIONING!!! You don't want your motor bouncing back and forth trying to reach a position that is not simply there because of the low resolution of the position sensor or the motor overshooting because of wrong speed setting (100kmh instead of 80kmh for example!!!)

Thanos

[tronicgr]

To give a more understandable example, I'll use the modern cars cruise control.

Usually to manually speed up to the desired speed, enable the cruise control and the car "tries" to maintain this speed by automatically injecting gas simulating your leg pressing the gas pedal. If goes over that speed it lets go until setpoint speed is reached again. The setpoint increments are usually within a digit precision so you can increase the cruise speed from 75mph to 76mph by just pressing a button. Lets say that this system uses 10-bit resolution...

Lets try it again with another cruise control system that has 8-bit resolution. The minimum increments in that system is 20mph (too much but just trying to keep the analogy)... So you can only set speeds of 20mph difference... You are cruising with speed 65mph and you press the button for the next increment... The car automatically accelerates until it reaches 85mph... imagine that you can't cancel that until it actually reaches that speed... What if there is an obstacle ahead... then what?

Brake? yes... but as they teach us in the driving school, the braking distance is a lot more when you brake at 85mph than if you brake with 65mph...

I hope my example made things clear, if not more confusing...

Thanos

[speedy]

Hi Prilade :

Thank you Thanos for interesting ... if you want me to be convinced ... I want you to give me different answers for the same questions that I gave Prilad ... And I can't ignore the point you both referring to Accuracy is higher in 10-bit ... that's true Dear friends ... that's right I agree totally ... I promise that I'll never forget that .
========================================================================================================================================

I didn't realize There is any error in H-bridge until Prilad opened my eyes and gave me simple calculations of voltage variations and the real time of it .
And that is exactly what I'm trying to feed back to you guys ...

Now Let's look to the motor controller output -->> to the motor with these [small changing, small pulses ] ... there is a problem :

Because these steps are not real steps in motor voltage ... they are On's and Off's which are combined together to give you a RMS summery of a let's say 4 of these 10-bit steps to give 48mv voltage variant -->> the same as 8-bit single step does ... which is really NOTHING to 12v motor point of view ... with near zero motor speed difference ... plus that the magnetic field build up in coils has been interrupted four times by now for nothing ... Why is that ?

what can a 48 Milli volt affect in the smallest 3V toy motor ... scaled to ... mmm ... a AA size 1.5V battery


due to above statement the final over all curve of the voltage applied to the motor in both [8-bit control OR 10-bit control ] cases will be the same plus a problem ...
Here is the problem ... A suffering, Heated, up used, misused, misjudged, H-bridge from worthless ON's and off's .killing magnetic field in motor without purpose ...

That's why a MOSFET with high specifications when used in H-Bridge ... visually doesn't give what is expected ... !!! ... and for what is expected to be known ... just connect the 12v dc motor directly to a smaller 6v dc battery and see the performance RPM and torque compared to 50% PWM [that will give also a 6v overall output ] ... Yes the PWM over all 6v is not like battery 6v ... there is an interrupted magnetic field and back EMF due to that ...

with the best of the best regards ...
spe edy

[vicpopo]

Hi Prilad,
I make a little break in this expert discussion just to post a video from your FW in amc1.6 with NMofset h-bridge on Xsim3.
Thanks to :
@prilad for your FW
@sirnoname for xsim3
@tronicgr his amc1.6
@speedy for his Nmofset h-bridge.

Test without loaded frame with Race on with just effect 27 longitudinal force.

Best Regards

[prilad]

Hi, guys!

OMG!!! Our discussion is endless

@Speedy:
Why do you think that 10-bit PWM make "shorter" or "smaller" pulses than 8-bit PWM? For the same % of voltage both PWM type make ABSOLUTELY equal pulses.

Not

...plus that the magnetic field build up in coils has been interrupted four times by now for nothing ...

Not and not. For 50% PWM (and RMS voltage) both 10-bit and 8-bit PWM produse ABSOLUTELY equal pulses width. You're wrong if you think that the 10-bit PWM always produces shorter pulses and four times more than the 8-bit PWM. The only difference is that the 10-bit PWM allows you to adjust the pulse width with greater accuracy.

Therefore, I believe that the 8-bit PWM can not be better than the 10-bit PWM.

@Thanos:
Yes, very good example. But let me back to the digits.

Suppose the maximum speed of two identical cars = 102.4 mph. Let the first car has a 10-bit control, and the second car has a 8-bit one. Then the minimum possible speed adjustment of the first car will be (102.4/1024) = 0.1 mph, and for the second (102.4/256) = 0.4 mph.

This means that at a given speed, for example, 60.2 mph the first car it can travel with the speed of 60.2 mph, and the second car with the speed = 60.4 mph (or 60.0 mph). Granted, not so bright and big difference as in your example.

Therefore, I think the 8-bit PWM is quite sufficient accuracy for x-sim projects.

@Vicpopo:
Thanks for video.

@all:

Therefore, I believe that the 8-bit PWM can not be better than the 10-bit PWM.

Therefore, I think the 8-bit PWM is quite sufficient accuracy for x-sim projects.

You can say that I contradict myself Ok, Let me make a general conclusion.
I think that 8-bit PWM is NOT BETTER than the 10-bit PWM, but it will still have sufficient accuracy for use in x-sim projects. (IMHO)

Regards,
Ale

[speedy]

Hi Prilad :
Теперь я понимаю, мой друг ...
Ho capito il mio amico ...

For 50% PWM (and RMS voltage) both 10-bit and 8-bit PWM produse ABSOLUTELY equal pulses width. You're wrong if you think that the 10-bit PWM always produces shorter pulses and four times more than the 8-bit PWM. The only difference is that the 10-bit PWM allows you to adjust the pulse width with greater accuracy.

the same frequency aaaah ...I'll read again to get what I missed ... excuse me my friend ... I don't understand ARABIC

thanks for your understanding and patience ...
regards
spe edy

[tronicgr]

5v /1024 = 0.0048v

0.0048v * 12(v) = 0.0585v
0.0048v * 24(v) = 0.1152v


5v / 256 = 0.0195v

0.0195v * 12(v) = 0.2345v
0.0195v * 24(v) = 0.468v


See the motors are not running in 5v... through h-bridges they run on 12v, 24v or more... A quarter or half volt step means a lot in motor power terms...


Thanos

[prilad]

5v /1024 = 0.0048v

0.0048v * 12(v) = 0.0585v
0.0048v * 24(v) = 0.1152v

No, my friend. Unfortunately you're wrong.

Minimal voltage adjustment step for 12V motors is:
For 10-bit PWM = 12V/1024 = 0.0117V = 11.7 mV
For 8-bit PWM = 12V/256 = 0.0469V = 46.9 mV

Minimal voltage adjustment step for 24V motors is:
For 10-bit PWM = 24V/1024 = 0.0235V = 23.5 mV
For 8-bit PWM = 24V/256 = 0.0938V = 93.8 mV

Regards,
Ale

[speedy]

Hi Prilade: The Nuclear digital Mind
Hi Thanos : The Site Angel
with my total respect as my big brothers ...
===================================================

Final question is the start question ... so we can come to some conclusion ... and EOD "end of discussion".

Q- [in English please ] ... What is the biggest deference between 8-bit and 10-bit PWM in real simulator motion effects ? .. What will be missing at the end ? the game data follow up speed will be effected ? the very fast movements and bumps will be missing or misrepresented , Road ripples sucks , ... what ?

regards
spe edy ... forgive me ... The Little Talking Machine