Taig mill tachometer design

[iointerrupt]

About:

This is a spindle tachometer specifically designed to mount on a Taig milling machine with no machine modifications required. The tachometer outputs a pulsed signal, to be feed into a breakout board and then into a computer, which can then display the RPM readout.

Attached are all the various files needed if one is interested in making this tachometer. The original schematic and circuit board were done with DipTrace. Standard Gerber files are included for manufacturing the board, as well as DXF files for milling one. A rough bill of materials is also included. Many details of its operation are outlined below.

Design History:

The original credit for this circuit I believe goes to Scott Shumate who had a simple design for using a Alan Pinku's enclosure for the QRB1114. His enclosure was nicer then anything I knew I could make though, so I set out to find another way.

The QRB1114 sensor has a rather small operating range, so the standoffs on the board correctly set its position relative to the spindle, 5/16" length ones worked good for me. Like Alan's design, this board uses the two pre-existing 10-32 holes on the dovetail plate to hold it all in place, so no permanent machine modification is needed. Given the extra width of the board and standoffs, the stock .75" 10-32 screws have to be replaced with 1.25" ones.

Since I run my CNC setup with an auxiliary 12V power supply for peripherals, I added the LM78L05 to the circuit to drop the voltage down to 5V needed for the sensor and output signal. As outlined below, this is rather flexible. Once the regulator was added, I decided a power LED might be nice too.

Next came the tachometer LED, mainly for debugging purposes. It turns on once per revolution of the spindle. At RPMs higher then 1500 this causes the LED to appear to be continuously on, which makes it not terribly useful in practice. Good for testing that the tachometer is working correctly or not however.

The final piece of the puzzle is the open-collector output for Gecko G540 support, discussed further below. As I run my machine with a G540, having a tachometer that worked with it was a bit of a requirement for me.

Any feedback on the circuit is welcome, as I am a software engineer by trade and only do electronics as a side hobby.

Circuit Board:

The circuit board for MillTach is a simple single-layer board, with wide isolations. This makes it rather ideal for milling, which is how I made mine. Using your mill to make parts for your mill is the first step towards a future of robot overlords, which I for one welcome.

There are some things to watch out for when assembling a MillTach board, mainly part orientation. The QRB1114 should be positioned such that the text writing on it points up. The NPN transistor, Q2, is laid out on the board as having pins in the order CBE, make sure your transistor matches this.

Worse yet, is U2, the LM78L05. The pinouts on these regulators seems to vary, the National ones I got were the reverse of what the board has. On the board, input is on the left, output is on the right, just make sure you match that. A wise idea is to solder the power supply portion of the board first (D1, C1, U2, C2) as well as the power LED (POWER, R1), then apply power and confirm it is working before proceeding.

Lastly, do not accidently apply power to the signal pins. I can tell you this from experience. If anyone with more circuit design experience knows how to protect against this, I'd be interested to know.

Connector and Pinout:

Although this is board is designed to use a JST XH connector, most people will probably want to just solder wires directly to it. JST XH is very common on radio-controlled models as a battery connector, but requires a crimping tool to make wire connections.

Looking down at the board from above, from left to right the pins are:
Pin 4: Tachometer signal. See "Output Signal" below as this signal may be any of two different styles.
Pin 3: Ground.
Pin 2: Ground, again.
Pin 1: Power. See "Power" below for more information.

Also should be noted here, that MillTach should share a ground with your breakout board or whatever the signal is being feed into. The two ground pins are to facilitate this.

Power Supply:

MillTach can use any power supply from 7-15V DC, and will draw at most 20mA of current, but probably far less.

Optionally MillTach can be powered directly off 5V DC if one omits some power supply parts (D1, C1, U2) and wire power directly into pads 2 and 3 of U2 (the ground and output of the LM78L05 respectively).

Higher supply voltages, up to 40V, are possible as some LM78L05s are rated to this higher voltage. The status LEDs would need to be omitted in this configuration though, as they'd over stress the regulator.

Output Signal:

The tachometer signal from the board has two possible modes, depending on the OUTPUT jumper (JMP1 on the schematic).

TTL mode: Output signal is either 0V or 5V. Probably use this for most breakout boards.

G540 mode: Output is an open collector output, behaving much like a switch (high impedance or connected to ground). This is mainly useful for Gecko G540 which has built-in 12V pullups on the inputs, as it expects them to be home switches. TTL output does not work in this case.

If you aren't going to use G540 mode, Q2, R5 and JMP1 may be omitted, and replaced with a jumper wire between the upper two pads of JMP1.

Mach 3 Configuration:

Setting up Mach 3 to display the tachometer in the Spindle DRO is probably the simplest part of this. Under Ports and Pins, set the "Index" signal to Enabled, the Pin Number to whatever input on your breakout board you used, and that's it.

If one experiences bad data coming from the tachometer at slow speeds, also set "Index Debounce", under General settings, to 1. Higher denounce values will cause high RPMs to read incorrectly.

Otherwise...

That should cover just about everything I can think of, let me know if you have questions. Once again all the files for this project are avaliable here.

[noisillator]

Looks like a nice accessory for the mill. I was planning an order with Mouser anyway, and this would only add a few dollars. Thanks for taking time to post it!

[Jeff-Birt]

very good job. I especially like the fact that your using the screw hole as the target; I like the simplicity. Mounting it in a small plastic box might be a good idea to prevent metal chips from causing problems.

You could also mask the sensor off and give it a few shots of spray polyurethane as a makeshift conformal coating. I'm not aware of the long term consequences of using polyurethane as conformal coating would be on the PCB or components but suspect that it would be fine for a hobbyist environment. You can get small batches of conformal coating from places like Mouser; even in dispenser pens which is nice for touch ups and might coat a whole board of this size.

[iointerrupt]

noisillator - Isn't that how all these projects get started. I figured a single DigiKey order, but it took another two DigiKey orders, plus another to Jameco and another to McMaster-Carr before the parts list had settled.

Jeff - I haven't been cutting conductive material, so I only gave it passing thought. I briefly considered milling a second blank PCB to put behind the main one to cover the traces a bit. But a spray-on conformal coating would probably be the best and easiest thing here.

[mcphill]

Any idea what max speed it will handle? I would be interested in putting it on a router with a max RPM around 25k - would it go that fast?

[noisillator]

Originally Posted by mcphill Any idea what max speed it will handle? I would be interested in putting it on a router with a max RPM around 25k - would it go that fast?

The detector is the slowest component in the onboard path, and the Fairchild datasheet states rise/fall times of 8uS. That should be more than sufficient to read 25K rpm.

http://www.datasheetcatalog.org/data...ld/QRB1114.pdf

The question is whether the PC interface and Mach3 can keep up.

[noisillator]

Originally Posted by Jeff-Birt I'm not aware of the long term consequences of using polyurethane as conformal coating would be on the PCB or components but suspect that it would be fine for a hobbyist environment.

The brush-on conformal coating that we use for military-related RF electronics is essentially a polyurethane with a fungicide added. I wouldn't hesitate to use standard polyurethane wood finish for protection of PCBs at home. It's a bit of a pain if you have to unsolder something though.

[Astroguy]

never mind

[mcphill]

Originally Posted by noisillator The detector is the slowest component in the onboard path, and the Fairchild datasheet states rise/fall times of 8uS. That should be more than sufficient to read 25K rpm. http://www.datasheetcatalog.org/data...ld/QRB1114.pdf The question is whether the PC interface and Mach3 can keep up.

True, guess I could route the signal through a counter and divide by 10 to scale it down a bit. That should work fine, as the resolution isn't that important.

[iointerrupt]

mcphill - Yeh, as noisillator points out, the sensor should be able to handle 25K RPM, up to 60K RPM perhaps?

I don't know what your router looks like, but you'll need a pretty reflective surface for the sensor though. Possibly some shiny tape could do the trick.

[mcphill]

Thanks to this post, I did something similar on my router. I did it in a much simpler (IMO) way. Since I have 5v available in my power supply, the circuit is very simple.

Electrical Parts:

Emitter/Receiver from Fry's:
http://www.ecgproducts.com/specs/310...df/nte3102.pdf
25 Ohm resistor

Mechanical parts:

I made a target on my rapid prototyping machine to break the emitter-receiver beam. It screws on to the router spindle, under the collet nut. The window on the target is 30 degrees "wide".

... That's it...

+5V (from power supply) to 25 Ohm Resistor to "+" on emitter; "E" on emitter to ground. "D" on receiver to Ground, "+" on receiver to Pin 15 on the parallel port (my breakout board has pull-up resistors built in on the inputs).

I set up the Index function as described below for pin 15.

Works GREAT! Router runs 15k to 30k, and the readout looks like it is tracking perfectly.

Thanks so much for the info in this thread, and I hope others can use the info as well.

Next step is to pull out the speed controller from the router, and do a closed-loop feedback circuit that I can control via GCode.

[Astroguy]

I just finished making 4 of these and they work great! Thanks for the plans.
I have 2 extra if anyone is interested.