Hour meter for laser cutter

[(*steve*)]

At the hackerspace I frequent we have a couple of laser cutters. We want to keep track of the laser time to track tube life.

Someone got inside and found a signal that is about 4V when the laser is off, and 0V when it's on. Some more investigation revealed that it can't sink or source much current at all.

The hour meter requires 12V to 80V and we have 24V available in the laser cutter, sharing the same ground with this signal.

The laser is controlled with PWM, and I don't know if this signal has that PWM on it or not, and we decided that we wanted to count laser time when engraving as the total time rather than the laser time (during engraving the laser turns on and off lots, and typically very quickly)

We drew up a few circuits, but I prefer what is essentially a pulse extender. The final circuit was this:

edit: this is updated in a post below.

I created a 5V rail from the 25V rail. The 4V input will appear as a logic HIGH to the Schmitt trigger inverters. There's a weak pull-up as well, and an input resistor to limit the current should the input voltage exceed the supply rails.

The input is inverted and extended. finally driving the gate of a mosfet. Whilst the mosfet isn't a logic level device, with 5V on the gate it can safely pass far more current than we need (which is a few mA for the hour meter).

Most of the complexity is for the various LED indicators, several of them being driven from spare inverters because there are 6 available, and I wasn't using all of them.

With the values shown, the pulse is extended by about half a second.

The whole prototype fits on a board about 5cm x 5cm with plenty of room around the outside for mounting holes and a terminal block.

 

[kellys_eye]

The idea for an hours-meter on the laser is great.

 

[shrtrnd]

Now submit a proposal to the manufacturer asking for compensation for designing an add-on to their laser-cutter that customers will want as a valued feature.

 

[(*steve*)]

And for amusement value, the schematic that was drawn on the whiteboard, which was the schematic I used when creating the PCB design!

The final circuit shown above wasn't drawn until after the PCB has been made and the circuit tested



The partial circuit below it was an initial option (with a transistor to invert the signal before driving the gate of a MOSFET) that was abandoned after some measurements of the current available from the signal and thoughts about handling real life aspects of the signal.

I'll post the PCB layout when I return to the PC it's on. And I'll tell a sorry tale of the paint ablation on the laser, and my need to eat dinner to get a vessel in which to etch the board.

 

[(*steve*)]

And here is the PCB layout. Blue is copper (top side) and red is silk screen (top side).

The silk screen layer is unused except as documentation.



Oh yeah, you'll note I have a 4.7uF 0805 cap specified here for the output of the regulator. I ended up using the 0.33uF, the current draw is minimal from this circuit.

 

[kellys_eye]

Nice, clear, smart and simple board layout. What software did you use to do it?

I'm looking for easy-layout pcb software that doesn't require you to first make a schematic, BOM, rats nest etc - just a simple click from A-B to drag the track and a library of pads!

and my need to eat dinner to get a vessel in which to etch the board.

LOL, Mrs k_e is missing a few (expensive) sandwich boxes that she swears have disappeared into my workshop!

 

[AnalogKid]

Assuming the LED Vf = 2 V, the LED current is only 0.6 mA. Seems a bit low, but I don't know the LED part numbers or the viewing environment. Also, while the 40106 high and low output currents are spec'd the same, in my experience the 40106 can sink more output current than it can source (for the same output stage voltage drop).

Since you have regulated 5 V, consider the 74AC14. Same pinout, same hysteretic inputs, but a very robust output stage that can source and sink 20 mA.

ak

 

[AnalogKid]

With the values shown, the pulse is extended by about half a second.

22K and 0.33uF looks like 7 milliseconds to me, less when you figure in D1's Vf.

ak

 

[(*steve*)]

22K and 0.33uF looks like 7 milliseconds to me, less when you figure in D1's Vf.

It's a 1M resistor. Well spotted.

I posted the wrong schematic. I'll post the correct (and slightly better formatted) one later.

Assuming the LED Vf = 2 V, the LED current is only 0.6 mA. Seems a bit low, but I don't know the LED part numbers or the viewing environment.

Surprisingly bright, and very easily seen in the indoor lot environment (in testing it was ok a well lit electronics bench).

As I think I mentioned earlier, they're almost a point source and appear very bright for a given current.

Can't tell you the part number, they're a "junkbox" part (a strip of unlabeled LEDs I got from somewhere I no longer recall). All I can tell you is that they're 0805 red LEDs with gold (!) coloured/plated contacts.

Also, while the 40106 high and low output currents are spec'd the same, in my experience the 40106 can sink more output current than it can source (for the same output stage voltage drop).

Since you have regulated 5 V, consider the 74AC14. Same pinout, same hysteretic inputs, but a very robust output stage that can source and sink 20 mA.

The 40106 is what I have, and that's what I used

Since I use spare gates for driving the LEDs, the current sourcing ability isn't a huge issue.

Likewise, I'm not switching the MOSFET rapidly, so the gate current is not important. Similarly, discharging the small capacitor isn't hugely time critical. And if I am discharging a capacitor, an output stage that can't sink lots of current isn't necessarily such a bad thing.

The main reason for the 5V supply is to bring the range of input voltages within the valid input range for the logic.

Nice, clear, smart and simple board layout. What software did you use to do it?

Sprint layout

I'm looking for easy-layout pcb software that doesn't require you to first make a schematic, BOM, rats nest etc - just a simple click from A-B to drag the track and a library of pads!

It's all that, except the vast library of pads. However designing your own pads varies from trivially simple to easy, and adding them to the library is quick and easy.

LOL, Mrs k_e is missing a few (expensive) sandwich boxes that she swears have disappeared into my workshop!

Fortunately, in this case it was just a takeaway container.

 

[(*steve*)]

Here is the updated schematic

 

[(*steve*)]

Here is one of the boards. It got slightly overetched, and you'll note that the board looks a bit twisted(?). The cooling air moved the PCB as the pattern was being applied.

 

[Engineer_Paul]

If you were looking for easy to do the job, I would probably use a light sensor to turn on the timer, as all gas lasers do produce visible light when in operation, even if the output is 10.7uM such as with CO2. Helium gas is a good source of white light as a plasma. Then you do not need to get into the logic (unless that is how you want to measure the time). I have always found that the simplest solution that does not interact with another circuit is best when measuring signals. Also isolation is a key thing when working with anything solid state around tubes. A tube can handle a high voltage spike, that will turn even 2400v hockey puck SCR's into a short circuit failure.

 

[(*steve*)]

Yeah, we had a flashover of the high voltage to wiring to the Z motor (which for some reason was placed in close proximity to each other. Fortunately the damage was limited to the Z controller.

After rerouting the wiring to the Z motor and replacing the cable to the anode of the tube, we hope never to see that again.

We also don't want to route any wires into the tube enclosure that we don't have to.

If I were going to use something to detect the light from the tube, I'd probably use a very similar circuit anyway. Even if I used only a single op-amp or schmitt trigger inverter, the number and type of packages would be similar. Maybe I could eliminate the voltage regulator. Considering the parts cost of this is essentially zero (I could add it up, but I'd be surprised if it exceeds $0.50) I'm not sure what I would save.

40106 $0.11
78L05 $0.03
0.33uF $0.002 x 3
2N7002 $0.009
LEDs $0.07 x 5 (maybe I should have picked the $0.004 ones?)
1N4148 $0.0045
5.11k 1% $0.0017 x 3
1M 1% $0.0012 x 2
22k 5% $0.0016
10k 1% $0.0011

Total cost... $0.5152 -- but hey, $0.35 was on those expensive LEDs. I'll pick the cheaper ones for the next one and my cost will be $0.1852.

Oh, hang on, the board cost me something (I can't recall, I got a lifetime's supply of off-cuts for bugger all)

And the laser time was about 6 minutes (for 4 boards)... That's $0.60 (although it's rounded up to $1)

The paint for the board is expensive (about $10 per can), although this was donated for the project and I kept the rest of the can. So about $0.20 (to be overly generous about the cost) out of the $10 can...

So add another $0.20 for the board

Yeah, I cheated and used the etchant at the makerspace, so I didn't pay for that.

All up cost for the one board was $0.72

Hmmm, my time. Let's assume I'm paid $x per hour (yes, they pay me that much). I probably spent 20 minutes chatting about the problem, another 10 minutes coming up with the circuit in my head, maybe an hour from blank sheet to design ready for the laser cutter, About another hour pulling out the parts (those bloody 1M resistors were hiding). Add another 3 hours traveling, painting, engraving, etching, and building the prototype, and we have 5.5 hours, so a huge opportunity cost of $5.5x. Oh hang on, another 1 hour making the pretty schematic to post here. $6.5x

So, $6.5x + $0.72 for the first board.

Yes, China would have been cheaper (if you assume a reasonable value for x), but waaay slower too.