Cub 3 and some kind of sensor help!!!

[Joe Gwinn]

I'd likely be using a wooden platen and ceramic magnet. But there are flexible epoxies, like G-2 Epoxy, which is used for aircraft. That what I use for guitar making. CA glue is also rigid.

The flexible epoxy ought to work. I didn't realize you had access to such epoxies.

I plan on using a Balluff inductive proximity sensor. It's similar, but not exactly like this part (they have different connectors)

Product Detail

I have six of them I got on eBay.

That ought to work, but beware the duty cycle problem.

[David Schwab]

The flexible epoxy ought to work. I didn't realize you had access to such epoxies.

G1, G2 Epoxy - Lee Valley Tools

It's good stuff.

That ought to work, but beware the duty cycle problem.

I'll keep that in mind. They were suggested to me by another pickup maker who uses a small neo magnet to trigger it. My counter has the inputs and power for the sensor.

I have yet to mess with the things though, or even power up the counter.

[RedHouse]

The flexible epoxy ought to work. I didn't realize you had access to such epoxies.

That ought to work, but beware the duty cycle problem.

If he's using a CUB4, 5 or 7 he shouldn't need to worry too much about the duty cycle thing, they have built-in Schmidt trigger input buffers and if you hook-up to the "HS" terminal it can go up to like 5Khz IIRC

Those CUB3's have small digits, the 4, 5 and 7 have bigger readouts.

[Joe Gwinn]

If he's using a CUB4, 5 or 7 he shouldn't need to worry too much about the duty cycle thing, they have built-in Schmidt trigger input buffers and if you hook-up to the "HS" terminal it can go up to like 5Khz IIRC.

The schmidt trigger input does not solve the duty cycle problem, but it does solve the reed switch contact bounce problem. In a sense, the schmidt and associated low-pass filter make the duty cycle problem worse.

The duty cycle problem is caused by a rotation sensor that activates on too small a portion of a rotation. As the rpm increases, the pulsewidth being sent to the counter becomes narrower and narrower, eventually becoming too short to be noticed by the counter. The solution is mechanical - one arranges things so the duty cycle of the pulses coming from the rotation sensor is around 50%, regardless of rpm.

[RedHouse]

The schmidt trigger input does not solve the duty cycle problem, but it does solve the reed switch contact bounce problem. In a sense, the schmidt and associated low-pass filter make the duty cycle problem worse.

The duty cycle problem is caused by a rotation sensor that activates on too small a portion of a rotation. As the rpm increases, the pulsewidth being sent to the counter becomes narrower and narrower, eventually becoming too short to be noticed by the counter. The solution is mechanical - one arranges things so the duty cycle of the pulses coming from the rotation sensor is around 50%, regardless of rpm.

So what one needs is an edge detector before the CUB's buffer that triggers on the edge of a pulse so the width isn't an issue, those are easy to implement.

I'm a little surprised it's even an issue Joe, say at 1000 RPM winding speed one would only be seeing ~16.6Hz are those hall sensors output that bad they can't maintain a decent pulse width at 16 or 17Hz?

[Joe Gwinn]

So what one needs is an edge detector before the CUB's buffer that triggers on the edge of a pulse so the width isn't an issue, those are easy to implement.

That would work.

I'm a little surprised it's even an issue Joe, say at 1000 RPM winding speed one would only be seeing ~16.6Hz are those hall sensors output that bad they can't maintain a decent pulse width at 16 or 17Hz?

Do the math. If the closure time is 5% of a rotation and the counter requires 10 mS pulses, then 5% of rotation period equals at least 10 mS, so rotation period is 200 mS, which corresponds to 5 rps, or 300 rpm. Above this speed, the counter will begin to miss counts.If the duty cycle is 50%, then that same counter can manage 3,000 rpm.

[JGundry]

Can't you just move the sensor and magnet further away from the axis of rotation to solve the duty cycle problem?

[Joe Gwinn]

Can't you just move the sensor and magnet further away from the axis of rotation to solve the duty cycle problem?

Probably not; probably reduces the duty cycle.

I've previously mentioned one way to solve the problem, the use of a pair of stationary parallel mild steel plates to conduct the magnetic flux from spinning magnet to stationary reed switch or hall effect sensor. The plates are perpendicular to the axis of rotation, and the magnet swings between the plates, with magnetic axis parallel to the axis of rotation.

[RedHouse]

...Do the math. If the closure time is 5% of a rotation and the counter requires 10 mS pulses, then 5% of rotation period equals at least 10 mS, so rotation period is 200 mS, which corresponds to 5 rps, or 300 rpm. Above this speed, the counter will begin to miss counts.If the duty cycle is 50%, then that same counter can manage 3,000 rpm....

Wow that's lame.

I'm glad I bailed on hall sensors for this application. Doing a 600-1100 RPM pickup wind is a walk-in-the-park for an opto and a CUB using the HS terminal. I've had my winder up over 2000 RPM doing round coils without a problem. Watched it on my o'scope to make sure it didn't skip a beat but then again my winder uses an interruper wheel that is more like 20% of rotation. As for me I can't see adding steel plates to crutch a hall sensor seems like a waste, but then I'm not there.

[David Schwab]

Do the math. If the closure time is 5% of a rotation and the counter requires 10 mS pulses, then 5% of rotation period equals at least 10 mS, so rotation period is 200 mS, which corresponds to 5 rps, or 300 rpm. Above this speed, the counter will begin to miss counts.If the duty cycle is 50%, then that same counter can manage 3,000 rpm.

Joe, you are a cornucopia of knowledge. Now I know what to avoid when I set up the sensor.

[Joe Gwinn]

I'm glad I bailed on hall sensors for this application. Doing a 600-1100 RPM pickup wind is a walk-in-the-park for an opto and a CUB using the HS terminal. I've had my winder up over 2000 RPM doing round coils without a problem. Watched it on my o'scope to make sure it didn't skip a beat but then again my winder uses an interruper wheel that is more like 20% of rotation. As for me I can't see adding steel plates to crutch a hall sensor seems like a waste, but then I'm not there.

If you are willing to provide power to a sensor, it's hard to beat an optical interrupter module. The advantage of a reed switch is that the CUB3 will provide the power to sense if the switch is open or closed. The steel plates don't have to be complicated or pretty or precise, and once built last forever. One can make them from electric box cover plates.

[Joe Gwinn]

Joe, you are a cornucopia of knowledge. Now I know what to avoid when I set up the sensor.

Thanks. I was a teenage engineer. Better than stealing hubcaps in the parking lot.

The quick test is a voltmeter to tell if the output to the counter is high or low. Rotate the bobbin by hand and see what fraction of the circle the signal is high.