Use of a PTC in Series with DC motor

[orionrobots]

I am a long time lurker and eevblog youtube fan. But this is a first post here.

I've been working on CnC motor control boards - after fixing my own, one of my own viewers has sent me theirs. Now there are a a number of cheap-ish CnC machines and controllers from China. The CnC mill itself is nice and rigid and made with nice solid aluminium extrusions. The steppers seem good, and the spindle motor is alright.

However, the electronics is another story. The controller for the steppers seems to use fairly standard Toshiba branded stepper controllers - with heatsinks, 1 per channel. There are opto's for those, for limit switches on this board.

The other board on it is a power distribution and spindle motor control board. I've a crude schematic I dug up here - https://github.com/orionrobots/CnCNotes/blob/gh-pages/galleries/2014-04-13-diagnosing-the-cause/YOC-PW3024-E-Diagram.jpg.

Power comes in via a huge toroid transformer, with multiple taps - both rectified, one regulated to 12v, and the other rectified but left unregulated for the stepper - . This is switched through a FET.

The FET is controlled via a 555 (powered by the 12v) and a pot adjusts the duty cycle of the 555, supply PWM to the FET to drive the low side of the motor. So far, so good.

The CNC spindle motor is being dragged through material. It has been a failure point of my machine twice, and one of my viewers machines too. The motor gets stalled, or a pot too high, and woops - overcurrent. There is a fuse on the motor control, and a current sense set up. The default set up of the machine is that the current sense optocoupler output goes to nothing though. The fuse on mine has blown twice, and both times, only after the magic smoke from other stuff went.

My first failure case - I managed (foolishly - a Z axis error) to embed the chuck in balsa - which is well clamped and allowed it to push in, but then held it fast. The FET blew, and the Fuse went after. Replacing the FET and Fuse had me back in business in a day or two.

The Second failure - the pot was dicky. It managed to drive the 555 to a continuous drive, which led the motor to pull too much current. This time was slightly spectacular. The FET failed - deadly and shorting, the 555 smoke started to come out, the 12V regulator went bang with sparks and black smoke (sorry - I didn't roll the camera for that), and both fuses (for 12v and 85v) blew. I had to replace a lot, and grew concerned that a dodgy pot could cause such a failure.

The third failure was not mine, but a chap who got a machine extra cheap - without working spindle. It turned out that his 555 was also blown, along with the FET, the current sense 2W resistors, and worse still - the fuse replaced with a screw  . I replaced the 555, put in a real fuse, and the FET, it took a while to realise the current shunt/sense resistors had gone, which probably meant it was still being run after that fuse should long have blown.

All of this is repairable, but it leads me to the conclusion that the board is slightly fail happy. My thought is to try and specify a PTC that would limit the current across this circuit if the motor starts to stall. Does this sound sensible? If I could place that in series with the fuse there, it should make the failures somewhat less catastrophic. The fuse is rated at 5Amps fast acting, the FET datasheet (https://www.vishay.com/docs/91031/sihf630p.pdf) suggests 5.7A to 9A. I'm suspecting I want a PTC that would reach high resistance at something lower than that - not quite sure how to account for a motor that will be under load, but should not be allowed to stall - I could pluck 4A out of the air, but that sounds a bit wrong.

Looking at PTC's on Farnell - that 85v with high currents looks far above what they are supposed to handle - so I do question if it is a goer. Is this an obvious thing that has been "value engineered" away on these cheap boards, or is there a fundamental reason (pardon my ignorance) why that won't work?

Thanks for any help and advice,
Danny (Orionrobots)
 

[Gyro]

Hi, Welcome to the forum 

PTCs are very good for protecting motors - where the thermal time constant of the PTC is faster than that of the windings. They're going to be less useful for protecting semiconductor junctions though, which typically have very short time constants. You can get FF rated fuses for semiconductor protection, but even those depend on the amount of 'headroom' the semiconductor ratings have.

If you want to give PTCs a try then you'll probably need to experiment. Go for the lowest current value that you think you can get away with and then see if it nuisance trips during operation (duty cycle and loads will be a factor). The datasheet curves are a start, but no substitute for physical testing.

The risk you run is causing additional stalls due to the PTC tripping, which might cause more mechanical damage.