[Arbo]

Just plug a USB cable in the back, and bundle it with the parallel port cable??

What kind of amp draw do you get for power off of a USB port?

Another thought on LED's. I didn't use them, but may make more switches that do, you must consider what your source voltage is going to be. If you are using a G540 input line for the 'input' of the switch, it is 12v, so you will need a bigger resistor to knock that down or you will instantly fry most LED's by sending 12 volts through it.

Here is an online calc for figuring out what resistor you need. Put in supply voltage (12v for G540), drop across the LED (ie, it's operational voltage), and LED current (what it will draw) and it will tell you what resistor you need.

R = (Vs - Vl ) /I

How to calculate resistance in ohms

[Lapin]

I know how to calculate resistors but im unsure which value i need to get it to play with the pullup resistor on the sensor. On page 4 the values changes if u use a LED so thats what im concerned about.

[ger21]

I used these values. 560 and 1500
http://www.cnczone.com/forums/758009-post39.html

[Lapin]

ok ill try with that then. I guess its not thet critical. Thanks.

[RomanLini]

The resistor values of 560 ohms and 1500 ohms in the diagram in post #39 of this thread (that Ger21 linked to) should be fine for +5v setups and +12v setups.

Lapin- You asked about some super high brightnedd Blue LEDs. Those will also be fine with the resistor values above.

Those high brightness LEDs are *really* bright, and it's possible you might find them distracting or annoying especially as blue light is the colour the eye is most sensitive to.

If you want to make the LEDs less bright you could make both resistors 1500 ohms instead of 560 and 1500. Seriously, some of those superbright LEDs will shine right through your hand and can cause eye damage if you stare right into them up close. Most have eye warning labels on the bag.

I don't want to alarm you, generally I prefer the high brightness LEDs especially for warning lights or if they are mounted deep in equipment and normal LEDs might be hard to see. It's still a good idea to do a test of one sensor and see how bright it is before you make all of them and fill them with epoxy!

You also made a good point about people may need to order parts and not know exactly which ones to order! Here is a;

Parts purchasing guideline;

Hall switch = Honeywell SS441A ('A' means the longer leads so it's easier to solder)
Resistors = 560 and 1500 ohms if you are using a LED, 1/4 or 1/2 watt is fine (1/4w are smaller and easier to fit)
Resistor = 2200 ohms 1/4w or 1/2w if you are not using a LED.
LED = pretty much any type is fine (high brightness good if it's buried deep)
Cable = shielded + 2 centre wires is best, usually sold as "stereo microphone cable" (all decent mic cable is rated for high flexibility life)

[ger21]

Parts purchasing guideline;

Hall switch = Honeywell SS441A ('A' means the longer leads so it's easier to solder)
Resistors = 560 and 1500 ohms if you are using a LED, 1/4 or 1/2 watt is fine (1/4w are smaller and easier to fit)
Resistor = 2200 ohms 1/4w or 1/2w if you are not using a LED.
LED = pretty much any type is fine (high brightness good if it's buried deep)
Cable = shielded + 2 centre wires is best, usually sold as "stereo microphone cable" (all decent mic cable is rated for high flexibility life)

For those in the US, I got my cable from McMaster Carr, $0.17/ft.
71055K71

It has a foil shield with drain wire that I used for ground.

[RomanLini]

Just for the record, foil shielded cable won't do as many flex cycles as woven braid, and non-woven braid is the best for max flex cycles. All 3 types provide equivalent shielding quality for this application so I would choose the cable suited for best flexibility.

Normally foil shield is used more for fixed installations, although if it is a decent quality and not flexed through too small a diameter it should still be ok for thousands of flex cycles.

Microphone cable (especially a good one) is rated for very high flex cycles and usually has toughness components ie; insulation plastics chosen for toughness as people tend to trample on it.

[Arbo]

I just ordered a lot of cable, can't beat it for .17 a foot.

Confused about the resistor thing though.

I have used the SS441A switch, one wire per leg. + to a power supply, output to an input line on the G540, and ground to a common ground.

Tested this on the G540, seems to work fine. No resistors used at all.

Of course, if using an LED, you'd need one to ensure the proper V across the LED so as to not burn it up.

[jsantos]

I wimped out and ordered this instead. It's got both the sensor and the LED in it.

[Lapin]

Its funnier to build yourself

[Al_The_Man]

There are quite a few commercial miniature types made for detecting a magnetic ring on pneumatic piston to detect position.
These are often 2 wire hall, as well as reed type.
The hall type have an indicator led.
Al.

[RomanLini]

Interesting, and thanks for the info.

I do like the fact that the open-source hall sensor can be setup as non contact and slide-by, that gives benefits of being immune to grit fouling cuasing errors and immune to over travel problems (and it can be mounted mid travel).

Also the honeywell hall switch we use is temperature compensated and uses a 4 hall internal bridge so it is extremely accurate and has good immunity to noise. It's about the top of the range for hall switches for accurate sensing especially when set up with the magnet types and angles suggested in this thread.

If you buy ready-made sensors you save some work, but you might end up with something less accurate as many brands of sensors are not temp compensated and don't have 4 hall internal bridges for high accuracy.

And as Lapin said; "Its funnier to build yourself "

[jsantos]

Here's the specs for the sensor I bought. It's got temperature compensation but I don't know about the other stuff. Would appreciate your input into these specs.
I'll find out the accuracy when I get them installed and let everyone know.

I agree "Its funnier to build yourself " but being the clutch I am when it comes to small stuff, i think I'd rather buy it if I can. Besides, I'll still have the fun trying to make it work :-)

[RomanLini]

I just checked out a couple of Hamlin datasheets, they do list temperature compensated so that is excellent.

The specs also look as if the Hamlin unit may contain the Honeywell sensor, although there is no guarantee.

But it is possible and the Hamlin units (with and without LEDs) do look to be a nice sensor with the same basic 3-wire connection.

[jsantos]

I just checked out a couple of Hamlin datasheets, they do list temperature compensated so that is excellent.

The specs also look as if the Hamlin unit may contain the Honeywell sensor, although there is no guarantee.

But it is possible and the Hamlin units (with and without LEDs) do look to be a nice sensor with the same basic 3-wire connection.

Thanks, Roman. Will post and let people know how it turns out. I just received them today. Now I have to find power for them.

[jscavitto]

Hey Joe, Did you get those Hamlin units mounted?

[jsantos]

Hey Joe, Did you get those Hamlin units mounted?

Not yet. I acquired some linear rails from eBay so I will probably install those first before changing my home switches.

But I did test the sensors and they are 100% repeatable to 0.01MM. I'm saying it's repeatable to 0.01MM 'cuz that's the resolution of the caliper I'm using. It's probably better than 0.01MM as what other people have measured in this thread.

The attached pictures show the setup I did for measuring. The second picture shows a macro of the magnet and the sensor. The LED is quite small and partly covered by the clamp. I'm using a neodymium magnet that I got from ebay. The distance between the sensor and the magnet is 7mm.

I also found some magnetic reed sensors of the same brand that are NC. I will use those as my end limit switches attached in series with my e-stop for those times when I accidentally turn off my soft limits. :-)

[RomanLini]

That's a very clever test setup for something so quick to put together. All it took you was 2 wood clamps, a vernier and a big brain.

My machine testing showed that 0.01mm repeatability seemed quite possible although after repeated tests it was soon picking up things like thermal expansion of the stepper motor shaft and leadscrew length. Once you can home your machine to 0.01mm you'll start to see all the thermal issues...

It would be good to hear people's input on that. With my new machine build for accurate pick'n'place of tiny electronic parts I contemplated 2 systems;

1. Use 2 home sensors (each at the far ends of an axis), then compensate the axis between both sensors so it scales the movement based on the total thermal expansion.

2. Use one home sensor in the centre of an axis travel, then just regularly home to this sensor and spread the thermal error +/- from the centre of the axis.

I ended up choosing system 2 for simplicity and faster homing procedure and because with PNP all the coords will increase with machine expansion... Unlike CNC cutting where the leadscrew expands as it gets used but each new job placed on the table doesn't.

[jsantos]

That's a very clever test setup for something so quick to put together. All it took you was 2 wood clamps, a vernier and a big brain.

My machine testing showed that 0.01mm repeatability seemed quite possible although after repeated tests it was soon picking up things like thermal expansion of the stepper motor shaft and leadscrew length. Once you can home your machine to 0.01mm you'll start to see all the thermal issues...

It would be good to hear people's input on that. With my new machine build for accurate pick'n'place of tiny electronic parts I contemplated 2 systems;

1. Use 2 home sensors (each at the far ends of an axis), then compensate the axis between both sensors so it scales the movement based on the total thermal expansion.

2. Use one home sensor in the centre of an axis travel, then just regularly home to this sensor and spread the thermal error +/- from the centre of the axis.

I ended up choosing system 2 for simplicity and faster homing procedure and because with PNP all the coords will increase with machine expansion... Unlike CNC cutting where the leadscrew expands as it gets used but each new job placed on the table doesn't.

Thanks, Roman.

I did see your post sometime back about putting the home switch in the middle and I meant to steal, I mean borrow, that idea. :-)

Getting to 0.01MM accuracy is way too much of an overkill for me as I work mostly with wood, but it's nice to know that you can achieve such accuracy. :-)

[Greolt]

Roman

Thanks for a great thread.

My little PCB mill now has home switches on X and Y.

I used Alegro A3144 sensors. Got em off ebay, 5 for $2.

Data sheet looks like they are pretty much identical to yours. Temp compensated etc. etc. Unipolar (south pole activated)

Went with tiny 6mm dia rare earth magnets. Mounted what I think you call "slide by 90 degrees". Only about 1.5mm apart.

Sees a north pole until magnet passes when it sees a south pole and activates. They are very accurate like this.

Didn't bother with LEDs. Couldn't see the point.

They work great.

Thanks

Greg