Paul & Sabrina's cheap DIY 144v motor controller

[MPaulHolmes]

I got a little 0.1watt speaker, and am going to hook it up to a breadboarded dspic30f4011. I'll listen to 8KHz pwm, and then test out the spread spectrum switching. This will let me fine tune how much variation in frequency is needed before the noise isn't irritating anymore.

[Astro]

Originally Posted by MPaulHolmes

I got a little 0.1watt speaker, and am going to hook it up to a breadboarded dspic30f4011. I'll listen to 8KHz pwm, and then test out the spread spectrum switching. This will let me fine tune how much variation in frequency is needed before the noise isn't irritating anymore.

I was just thinking, now you need to create a mode where the pwm is varied in such a way that the motor plays a tune.

That way it won't just be the F1 guys that can make their motors play music.

http://www.youtube.com/watch?v=1JPBdBIFGNQ

[Astro]

Originally Posted by MPaulHolmes

(if you want to get your own cable, or can find 40 feet, let me know)
Welding Cable 2 AWG 50
...
The last part is the end caps. I haven't gotten that yet either. I was thinking 1/4" or thicker of some plasticky sort of material. I hate this part of controller making. The dang enclosure, and interface to the outside world.

For the housing. Maybe add some through connects similar to what Curtis has on their controllers. Have the copper buses bend upwards at the end and so that they are vertical and in line with the ends so the through connects can screw one side directly to the bus and the outside have a stud to connect the lugged cables from the motor/batteries.That way you could use metal ends and screw it down making it weather proof and not have to mess about with plastic ends?
Is there a need for air flow or does the heat sink suffice?

[jmcginley]

Controller playing music? Maybe try for Smoke on the Water!

[MPaulHolmes]

Originally Posted by Astro

For the housing. Maybe add some through connects similar to what Curtis has on their controllers. Have the copper buses bend upwards at the end and so that they are vertical and in line with the ends so the through connects can screw one side directly to the bus and the outside have a stud to connect the lugged cables from the motor/batteries.That way you could use metal ends and screw it down making it weather proof and not have to mess about with plastic ends?
Is there a need for air flow or does the heat sink suffice?

I'm having trouble picturing this. Metal end caps? How would you isolate the bus from them?

[MPaulHolmes]

This was the first attempt at spread spectrum switching, and the 8kHz tone:
https://www.dropbox.com/s/8p7fidujso...2011.20.07.mp4

This is what it sounded like after trying to increase the randomization. It was noticeably more hissy:
https://www.dropbox.com/s/krmry0occ4...2011.39.27.mp4

It won't be this loud though while driving. I had the speaker blasting on maximum. All 0.1 watts of her. haha.

[MPaulHolmes]

I finally found a use for group theory! To take the array of random values of frequency, and just loop it, definitely doesn't sound too random. People are masters at hearing patterns. If you loop a random noise, you'll hear it if it repeats too often. So... You take the group <Z mod p, *>, but with zero removed, and P is a prime. Then, for any x in Z mod p,

x
x*x
x*x*x
x*x*x*x
...
x*x*x*x...*x (p-1 times) are all different values!! So, it gives you a way to generate every index in the array of random values. YOu can generate a different sequence of numbers from 1 to p-1, for every single element in Z mod p (except for 0 and 1).
For example, if p is 127, you can find
2
2*2 = 4
2*2*2 = 8
2*2*2*2 = 16
...
etc...

Then, you just choose position 2 in the random frequency array, then 4, then 8, then 16, ... eventually it comes back around to 1 (mod 127).

So, you can have truly white noise! Where it won't repeat until, say, every 4 or 8 seconds, but it's all so random, that you won't detect any white noise patterns that are 8 seconds apart!

EDIT: Oh ya, it works perfectly. There are no overtones of anything repeating. It's totally white noise. I'm going to cry it's so perfect.

[flores]

Originally Posted by MPaulHolmes

EDIT: Oh ya, it works perfectly. There are no overtones of anything repeating. It's totally white noise. I'm going to cry it's so perfect.

I lost you on the second sentence, but I glad you are glad.

[flores]

Originally Posted by MPaulHolmes

DHL Tracking number: 1077685674.

They are in LA now. Do you have all the parts already?

[MPaulHolmes]

Originally Posted by flores

The-1are in LA now. Do you have all the parts already?

Is a fish thirsty? Heck yes I do! Well, when I had changed a couple things, I had already ordered the parts... so I may need some 30 ohm resistors, but may not. One change was from 1 10ohm To 3 30ohm on the 3 inputs to the igbt drivers. By the way, back to group theory! It will be computationally cheaper to do
X
X+x
x+x+x.
Etc., since those will also all generate a distinct sequence of p objects.