2168 A3144 Hall effect sensor board

[otherguy]

I've been prototyping a hall effect sensor board during the holidays with a bunch of components purchased off ebay.



This kind of sensor can detect the presence/absence of a magnet. They don't actually need to contact the component in motion.

The A3144 sensor has an input voltage range of 4.5 - 24vdc. The board is designed to work off a 5volt supply voltage.

I ordered some 15mm (outer diameter) N45 neodymium magnets with a hole through the center that you can rivet/screw through. They trigger the sensor at about 0.5" away. Sensing range will depend on the strenngth of the magnetic field produced by the magnet.

If anyone needs/wants a BOM let me know and I can put something together. Nothing real special here though. All discretes are 0805 surface mount components.

• 1x 10uF capacitor for decoupling.
• 2x green LEDs
• 2x 180ohm resistors
• 1x female right angle header (0.1" pitch)
• 1x A3144 hall effect sensor.

The board can be ordered from OSHWPark here.
I haven't tested this PCB out, but have built a number of boards by hand with the same schematic.

All source files are here.

Note, if you're interested in this kind of sensor, but don't want to build your own, there's one sold on WCP from FRC971 here (not sure where the description and images went - new site?).

[asid61]

I'll keep this in mind when I need a hall effect board.
Do you have a schematic available so I can see where to plug stuff in? Or a labeled version of the PCB? I'm sorry, I just can't tell which way the hall effect faces/ where it plugs in. Once I see the labels it'll probably be really obvious where everything goes.

[otherguy]

Sorry, there wasn't much room on the board for component labels. The schematic is, as you guessed, very simple.

All the schematic files are at the link above (see: "all source files are here").
They're all for Kicad.

I'll add a PDF version of the schematic so that you don't need to install Kicad if you haven't already.

[craigboez]

Can someone who has used the 971 board or something similar like this comment on how accurate and repeatable they are? What would you use them for and where would you avoid using them?

I'm happy to see a DIY version of what can be purchased from WCP.

[philso]

You may want to add one or two (better) mounting holes. Otherwise you can only attach this with cable ties, velcro or double sided tape. Your application may benefit from there being a repeatable and consistent spatial relationship between the object being sensed and the board.

[MichaelBick]

Originally Posted by craigboez

Can someone who has used the 971 board or something similar like this comment on how accurate and repeatable they are? What would you use them for and where would you avoid using them?

I'm happy to see a DIY version of what can be purchased from WCP.

I haven't used them, so I'll let somebody else comment on actual usage, but the reason 971 developed the sensor is because they found that regular mechanical limit switches accidentally triggered. They are essentially made to replace limit switches.

[mman1506]

Why did you choose to use a through hole limit switch over an SMD one? Also aren't the A3144 discontinued?

[electroken]

Originally Posted by mman1506

Why did you choose to use a through hole limit switch over an SMD one? Also aren't the A3144 discontinued?

The A3144 is indeed discontinued. Allegro Microsystems A1104 should drop right in for our purposes.

[otherguy]

Originally Posted by mman1506

Why did you choose to use a through hole limit switch over an SMD one? Also aren't the A3144 discontinued?

The only reason I used through hole components was because that's what I have on hand. I'm making these boards for testing purposes. If we find we link the sensors and could use more, it would certainly make sense to use all surface mount components as it would prevent there being any conductive components on the bottom of the board, making it easier to mount.

You're right, the A3144 component is a discontinued part, but they are widely available online (search ebay for hall effect sensor). I'm assuming they are clones from China, but the parts I've been using all seem to be workign correctly. If the supply actually dries up, its easy enough to spin a new version of the board. Obviously if I were trying to build a product for sale, I wouldn't be so recklessly sourcing components.

You may want to add one or two (better) mounting holes.

There is a hole right down the middle of the board. I agree though two holes would be preferred. As it is currently implemented, the board may rotate, but I'm betting friction will be sufficient to keep this from moving. We will see.
I was playing with the idea of having two hemisphere cutouts on the side of the board, allowing two screws to sandwich the board between them, but didn't have time to verify that the way I knew how to do it in Kicad would actually meet the fabrication guidelines on OSHWpark. I didnt want to add another screw hole with keepout as it would increase the board size. I was trying to keep the board as small as possible.

Can someone who has used the 971 board or something similar like this comment on how accurate and repeatable they are?

I cut up a bit of video I took of one of the prototypes in action. It's pretty accurate, but you can see for yourself if you think it's good enough.
http://youtu.be/6Dsre-IwP_I

[Travis Schuh]

Originally Posted by craigboez

Can someone who has used the 971 board or something similar like this comment on how accurate and repeatable they are? What would you use them for and where would you avoid using them?

I'm happy to see a DIY version of what can be purchased from WCP.

We used the ones from WCP (http://www.wcproducts.net/sensors) all over our robot last year and never had an issue about accuracy or repeatability. I know the chip has a spec somewhere, I will add them if I end up finding them. I remember the spec being within a couple thou, we decided it was good enough for our zeroing needs. We used them for zeroing our arms, range limits on our arm, zeroing our shooter, and sensing state in our shooter (particularly of the pneumatics). We have used them to replace limit switches for sensing position and find them easier to integrate, more robust, and more reliable.

We have used bare sensors in the past (http://www.adafruit.com/products/158), but have found the status lights on the WCP boards to be immensely valuable (this is why we worked with WCP to bring these to market). We saved probably 10+ hours of debug time (often at the most critical times...) by having those lights on the board.

[craigboez]

Originally Posted by Travis Schuh

We used the ones from WCP (http://www.wcproducts.net/sensors) all over our robot last year and never had an issue about accuracy or repeatability. I know the chip has a spec somewhere, I will add them if I end up finding them. I remember the spec being within a couple thou, we decided it was good enough for our zeroing needs. We used them for zeroing our arms, range limits on our arm, zeroing our shooter, and sensing state in our shooter (particularly of the pneumatics). We have used them to replace limit switches for sensing position and find them easier to integrate, more robust, and more .

How did you guys mount these- just double stick tape, or did you 3d print a case, or something else?

[Travis Schuh]

Originally Posted by craigboez

How did you guys mount these- just double stick tape, or did you 3d print a case, or something else?

We bolt or pop-rivet them onto plate, with a hole cut out for the connector (or in our case, we remove the connectors and solder directly). Double stick tape like the 3M VHB could also work. We used to mount them this way before we had boards with mount holes in them. You can check out our 2014 robot CAD to see exactly what we did if you are interested in more details.

Poprivets are nice because they are quick, but we will probably move to 4-40s exclusively next year because we can replace the boards if needed. We had two minor failures last year due to one of the wires breaking off on the end of our shooter (in-proper strain relieving in a high robot-to-robot contact area) and one of the boards being mounted too close to the magnet and getting damaged due to contact with the moving part. We ended up replacing the pop rivets with 4-40s when servicing these, so it would be nice to start with 4-40s.

[Mr. Lim]

It's been a while since we've had a sensor application for stuff like this.

Quick question:

What are the pros/cons of using a reed switch for a similar FRC application?

Are they just too fragile? Even the molded-over ones?

[otherguy]

A reed switch is still a mechanical device.
My guess is that a magnetic field overcomes the force applied by some spring steel or something?

I'll walk through my rational, but state up front that I have very little experience with reed switches. I could be writing them off prematurely.

While the reed switch is non-contacting, it would still be susceptible to conditions that would cause the reed to move when the actual physical component it is mounted to has not (i'm thinking shock event).

The hall effect sensor on the other hand has no moving parts (it may be more accurate to say the MEMS components that do move have such little mass that they won't accelerate significantly do to shock). So the only time they should indicate a state change is when the magnet (and the component its connected to) actually moves away from the sensor.

That said... looking through digikey, there are some pretty tiny surface mount reed switches. It might be a viable option. They are in the same price range as the hall effect sensors I used.

At the end of the day the hall effect or reed switch is just another tool in the toolbox for us. It's not going to be the end-all sensor. I like to have options on hand so that if we have ideas we can prove them out quickly. Hopefully they end up on the field, but if they don't it was a pretty small investment. I ordered electrical components off ebay, with the custom PCBs, these things come in at around $2.50 a board.

[otherguy]

Boards are in, populated, and tested out good.