Scratch wheel

[cimo]

Rasteri, thanks for posting the diagram, but you probably disabled some layer in photoshop as I don`t see any Kryptonite part. We all know how importance of Kriptonite in scratch controller designs are, but some neeb can be confused.

EDIT: Not sure did you know, but Kryptonite is recently discovered in serbia (where I live)  ;D

http://news.bbc.co.uk/2/hi/science/nature/6584229.stm

in this case i would trade 0.0005% tolerance cut resistor leads for Kryptonite, or, for what it matters, it could be Kryplateintheevening, i am not in a hurry you know...

I ll PM you with details

[Sasha]

OK man, we have a deal.  ;)

[rasteri]

Have some code.

(Note the encoder now needs to be on pins 4&5, and it requires a PIC with comparators).

main.c

[rasteri]

Some pics :

(http://picasaweb.google.co.uk/rasteri/Scratchtroller)

Not pretty looking, but pretty damn functional :P

[Sasha]

I see.. you want wear your records no more.  :)

How do you place it... horizontally or vertically. I ask because of fader position.

I`m not an expert, but to me it feels most natural fader position would be something like 45°. What do you think?

Thanks for the pics.

[stryd_one]

Nice one rasteri!

Why the comparators I wonder? I take it that the levels coming from the sensors didn't always hit 0?

to me it feels most natural fader position would be something like 45°.

that's an interesting concept! My split-second was reaction was "of course!" because its seems more ergonomic; it will leave your arm/wrist/hand in a more natural position (straight line), and allow you to move some of the rotation into your shoulder joints (and therefore away from your wrist).

Then I tried it (on my air-mixer. it's very light.) and realised - that's good for a long set and all, but no good for fast movement. This is because the fastest movement you could use for the fader is a radial rotation of the hand, which uses the two bones in your forearm as an axis, and then your hand would be at a 90 degree angle, converting the rotation into lateral motion... yaknow, same way you always see DJ's hold their hand when scratching fast...it's natural for a reason :) To allaw the forearm to rotate fast, yu want to have your elbow right in next to your ribs, which means your forearm will point straight out away from you. When you do it that way, the 90degree angle that is standard(xfade at the front) makes the most sense.

That said, the upfaders would benefit from being at a 45degree angle in both situations, so that they are closer to perpendicular (like the xfader is)... if you rock the upfaders with both hands that could be tricky though......

Woops back to topic ;)

[Sasha]

I agree. It makes sense. I lack some experience, and gear to test it. Actually, only once I tried to use real crossfader that is not from my DIY stuff.  :-[

[rasteri]

You can adjust the angle of the fader relative to the user just by rotating the entire device. The scratch wheel doesn't care what angle you come at it from. The device wasn't really designed with any "ideal" orientation in mind.

Anyway, If the fader wasn't parallel with the front edge of the case, it would mean it could only be comfortably used by either left-handed OR right-handed DJs, not both. If you get me.

And yeah, the levels from the sensors didn't always hit 0. Probably fixable with circuitry tweaks, but being able to adjust the threshold is useful anyway I guess.

[cimo]

wicked, i ll make one for Sahira

2 different categories but this goes along with sasha s Ableton controller, nice work!

Simone

[nILS]

[me=nILS Podewski]orders one[/me]

When's the foolproof docu gonna be available? ;D

[rasteri]

When's the foolproof docu gonna be available? ;D

I've started to write a howto in the wiki (it's still a long way from being finished) : http://www.midibox.org/dokuwiki/scratchcontroller

[Sasha]

Thanks alot rasteri.  ;)

[Flexinoodle]

Looks like a MINI ITX would fit in that box for a true portable practice wheel

Looking forward to building one

Problem is i'm very big and have very wide shoulders and there is no way that fader position would do me so i'll have to use a bigger box

Rasteri your a bad young brother :)

Best thing i have seen in years

Flex

[dstamand]

http://www.usdigital.com/products/e2/body_index#description

Take a look into this miniature hollow shaft optical encoders.

[Sasha]

Looks great. But these thinks are usually pretty expensive.  :-\

[dstamand]

Looks great. But these thinks are usually pretty expensive.   :-\

I have a couple of these, i bought them in 1999-2000 when i was working on a precision remote control PTZ camera. It cost around 65$ each. I'm gonna try this sometime with the scratch wheel  :)

[MFDOOM01]

Hay rasteri ., i saw ur project on scratchworx ., and i think its wicked what you are doing!! just wanted to ask if u or any 1 else out there reckons it would be possible to use this scratchbox with rane serato ???

Im looking foward to seeing the complete guide on how to build a scratch box., ive already started harvesting parts etc for 1!!

keep up the good work!!

regards!

[rasteri]

I don't think serato supports external midi controllers for scratching. I think the best it can do is map midi buttons to functions in the software, but don't quote me on that.

As for a guide, there's enough info on the wiki page (http://www.midibox.org/dokuwiki/scratchcontroller) and other midibox pages for you to figure it out. I realise this might be hard if you've never built a midibox before, but the people on this forum are unbelievably helpful, and I'll try and help you as best I can if you give me a shout.

A complete start-to-finish guide will be written eventually.

[stryd_one]

Any of you turntable encoder DIY freaks wanna try this quadrature decoder?

// Top of file.. declare as registers
register unsigned char sensor1,sensor2,oldsensor1,oldsensor2;
register int sensordirection;

....
....
....

void yourfunction(whatever){

unsigned char direction = 0;							// Default to an error/no movement

.....


  if (!((oldsensor1 ^ sensor1) & !(oldsensor2 ^ sensor2)) ) {	// If this was fwd/back
	direction = -1;												// Default to reverse
	if (oldsensor1 ^ sensor2) direction = 1;					// Set if fwd
  }
  
  sensordirection += direction;
  
  oldsensor1 = sensor1;
  oldsensor2 = sensor2;

Compiles to

_00115_DS_:
;	.line	97; main.c	if (!((oldsensor1 ^ sensor1) & !(oldsensor2 ^ sensor2)) ) {	// If this was fwd/back
	MOVF	_sensor1, W
	XORWF	_oldsensor1, W
	MOVWF	r0x01
	MOVF	_sensor2, W
	XORWF	_oldsensor2, W
	MOVWF	r0x02
	MOVF	r0x02, W
	BSF	STATUS, 0
	TSTFSZ	WREG
	BCF	STATUS, 0
	CLRF	r0x02
	RLCF	r0x02, F
	MOVF	r0x02, W
	ANDWF	r0x01, F
	MOVF	r0x01, W
	BNZ	_00119_DS_
;	.line	98; main.c	direction = -1;												// Default to reverse
	MOVLW	0xff
	MOVWF	r0x00
;	.line	99; main.c	if (oldsensor1 ^ sensor2) direction = 1;					// Set if fwd
	MOVF	_sensor2, W
	XORWF	_oldsensor1, W
	MOVWF	r0x01
	MOVF	r0x01, W
	BZ	_00119_DS_
	MOVLW	0x01
	MOVWF	r0x00
_00119_DS_:
;	.line	102; main.c	sensordirection += direction;
	CLRF	r0x01
	MOVF	r0x00, W
	ADDWF	_sensordirection, F
	MOVF	r0x01, W
	ADDWFC	(_sensordirection + 1), F

Which is around 140 instructions faster than the old way = you can scratch faster = leaves more room for other fun stuff on the chip :)

If you want to see my working you can check out the attached spreadsheet. You can thank narwhal/karnlund for convincing me that I was correct, that this could be done with logic, thereby making me actually try it ;)

Edit: Whoops! My excel equations were correct, but my pseudocode i made from them, had an error... Fixed.

quadrature.zip

[rasteri]

awesomeness, I'll give it a go.

[stryd_one]

Thx man, let me know :) I'm dying to know if it actually works... looking at the encoding and the equation I'm sure it will, but I lack the hardware to really find out...

[Narwhal]

I'll be giving this a try myself as soon as I can, but my excel work shows that the directions need to be reversed:

direction = 1;												// Default to forward
	if (oldsensor1 ^ sensor2) direction = -1;

I haven't confirmed this other than in excel, but it's something to keep an eye out for.  For right now, Excel seems to indicate that it _should_ work :-)

Anyone want to double check the excel work?  The logical equivalent of XOR in excel is OR( AND(NOT(A),B), AND(A,NOT(B) )

EncoderXORtest.xls.zip

[stryd_one]

my excel work shows that the directions need to be reversed:

Go with that then mate... Unlike you, I got the binary values in my spreadsheet from some webpage, so it's not exactly guaranteed ;)

Your working looks good to me. I'm really glad we decided to try this, at 10 times better performance it allows room for the pic to do other handy stuff.

Has anyone considered recording/effecting/playing back their scratches? I mean the movements of the turntable/encoder, not the audio it modifies...

[Narwhal]

Has anyone considered recording/effecting/playing back their scratches? I mean the movements of the turntable/encoder, not the audio it modifies...

Traktor 3's Native Mix files are essentially recordings of all the controls you interact with.  It's a really great idea, I think I drooled a bit when I first played around with it.

But are you thinking about applying effects to the movement data itself?

[stryd_one]

Traktor 3's Native Mix files are essentially recordings of all the controls you interact with

I meant doing it in hardware though... I guess if all the apps support it, why bother eh?

But are you thinking about applying effects to the movement data itself?

Bingo.