[misterstarshine]

I need to discuss details about the software by the way. All I've done is some g-code and simple cadding about 10 years ago. I've got a poor insight in the details of a CNC software. I must say I REALLY HOPE that they support feedback from the machinery. After all the PC software side dictate the timing (master) and the CNC machine is the slave in this sense. All CNC softwares should have the (optional) ability of operating with feedback from the CNC machine in order to guarantee that the machine catches up with the PC software.

This should be a concern for ALL motor drivers that are supposed to be considered suitable for CNC applications. What I still don't know is how expensive this is. I've got my worst fears since many low-budget CNC apps ignore the feedback approach.

Now is the question, SHOULD feedback be expensive and complicated? Judge yourselves,



Are ball mice expensive? Didn't think so either

/Tom

[Jojje Bergan]

misterstarshine: You can get everything from me. I can send you one of mine present controllers and 3 motors. Then you can use Mach3 Demo. I can also support you with a photoplotted film in right scale for the motor.

[misterstarshine]

So most CNC software are compatible with feedback then. I was thinking that the feedback process had to go through the stepper driver but when I think about it I realize this is not a necessity. If this would prove to be the situation then Jojjes, Markcomps and all the other plain and simple A3986 driver would already be capable for a feedback system.

As far as I can see then, the need of an open-loop anti-stall regulation is then only useful in the absence of a feedback encoder. Is this true or am I completely lost here?

/Tom

[Jojje Bergan]

I was thinking you could borrow one of mine A3977-controller for testing. I have ~30 controllers in stock now

http://www.bergan.se/page_1148574213203.html

[markcomp77]

ster3986 dimensions

thaks... there is problem of dimension
problem in library... or datasheet



and



sizes - scan printed on A4

3.6mm - x
1mm - 12mm => x = 0.3mm (leg) good

row of 19*leg
19 x leg -> 109,25mm (on print)
109,25mm - x
12 - 1 => x=9,1041666mm

datasheet x= 19*0.3 + 18*0.2 = 9.3mm
...

[riko]

I am measuring 6.5 mm from outside pin to ouside pin using my alipers on the chip.

The footprint in the documentation looks right since it would be 5.9 + 1.85 = 7.75.

Are you using Eagle PCB program?

Can you post the library you are using?

[pminmo]

thaks... there is problem of dimension
problem in library... or datasheet



3.6mm - x
1mm - 12mm => x = 0.3mm (leg) good

row of 19*leg
19 x leg -> 109,25mm (on print)
109,25mm - x
12 - 1 => x=9,1041666mm

datasheet x= 19*0.3 + 18*0.2 = 9.3mm
...

My library 38 pin tssop I believe was wrong also. I suspect it was created incorrectly. I think my prototype will be usable, but....not sure. should know today or tomorrow when I try to install the 3986..

[markcomp77]

Comparison of :
-real pcb
-real chip
-printed pcb
one scan together...

http://www.2a.pl/~mark/cnc/a3986/pcb_chip2.jpg

[riko]

Don't quite understand all of you calulcations.

But the pin spacing is 18*.5mm = 9mm

The part is too narrow.

[markcomp77]

I will make today... new tssop definition

[riko]

What library are you using

[markcomp77]

Don't quite understand all of you calulcations.

I use linear approximation...

I measure on printed scan... size of 1mm + size of leg (or ...)
measured size of 1mm is 12mm => 1mm - 12mm
measured size of leg is 3.6mm... and I don't know its real size => x - 3.6mm

So,

1 - 12
x - 3.6
=> x * 12 = 3.6 * 1 => x=3.6 * 1 / 12 = 0.3mm

Etc.

[markcomp77]

What library are you using

I used allegro definition of tssop38

[riko]

I will re-etch new board when it is ready.

Phil, I feel your pain.

Solution if width is too narrow, bend pins downward.

For bad spacing: Solder in using micro solder technique for whatever pins are good, Then use a small screw driver to bend the rest into place.

Here's what I am using to check the alignment:

http://www.radioshack.com/product/in...entPage=search

Illuminated Microscope

$9.99
Model: MM-100
Catalog #: 63-1313

[riko]

Is it the libaray that is bad or the chip documentation

[riko]

There are 25.4mm per inch.

So .02 inch spacing = .508 mm spacing

Alternatively .50 mm = .019685 innches

So what is the real spec on the part. Looks like Allegro dropped the significant digits (which come inot play on sucha small part)

[markcomp77]

I will re-etch new board when it is ready.

I also...

I will make new definition of a3986 tssop... wider etc...
Riko.... make definition of real tsso also too

[markcomp77]

Metric vs English

on:

on scale - left is mm, right is inch

[originator]

Here is a 5mm pitch 3986 libary I made, most Eagle Tssop's are .65mm pitch. In fact, the first in my library is .65 by mistake, use the one marked 5mm! .65 Tssop's are common.

BTW there is a prodcut I just discovered you guys may alreay knwo about, alloyed paste in a syringe, just run a bead and cook it. Kester brand, and a few others. Looks really simple from the reading. (hopefully)

[riko]

Originator, Great Post.

I guess its a metric pitch.

I am guessing everyone's been using an English TSSOP (ie .020" pitch).

I guess it's true you have to do the math.

Oh well, can't cry over spilled milk.

I'll just have to re-etch when the new board is ready form Mark77.