Software suggestions and requests

[DoulosDS]

..... I have asked the seller if he has any more of the 24 VDC pumps left... I will let you know what he says.

He says the 24 VDC pump does NOT pump 4500 ml/min but only 450 ml/min... the label was misprinted (then why use it to illustrate the 12 VDC pump?) and he didn't say whether he still had them for sale (if anyone is wanting one.).

[DoulosDS]

Another comment on peristaltic pump accuracy: those driven by DC motors, rather than stepper motors, have gear trains to provide the necessary torque. The $12 pump I found on Amazon has a maximum rotor speed of 100 RPM, but the motor is rated at 5000 RPM (probably no-load speed), so I am guessing there's a planetary gear system between the motor and the pump. If there can be a sensor on the motor shaft, the precision of the pump output can be multiplied by the gear ratio. Assuming a gear ratio of 36:1, that would make the slice thickness accuracy much better. Given the torque load of the pump rotor, I expect the motor doesn't coast very long after the power is shut off, maybe one revolution or less. Instead of a sensor on the shaft, it may be possible to detect the motor armature transitions in the motor current or voltage. Depending on the number of segments in the armature, that would provide even greater precision.

[DoulosDS]

Anybody know how to get precise control of a DC motor? I'm thinking a microcontroller with a hall effect sensor on the motor shaft counting full turns. An A/D input would monitor the current or voltage and count the armature pulses during the revolution to determine angular position. Then it should apply a reverse voltage pulse for rapid stopping.

[CescoAiel]

Anybody know how to get precise control of a DC motor? I'm thinking a microcontroller with a hall effect sensor on the motor shaft counting full turns. An A/D input would monitor the current or voltage and count the armature pulses during the revolution to determine angular position. Then it should apply a reverse voltage pulse for rapid stopping.

Or a conductive disk turning around with the axle with a drag-contact and a non-conductive area... Ofc. that won't help much if it turns too fast...

[CescoAiel]

Another comment on peristaltic pump accuracy: those driven by DC motors, rather than stepper motors, have gear trains to provide the necessary torque. The $12 pump I found on Amazon has a maximum rotor speed of 100 RPM, but the motor is rated at 5000 RPM (probably no-load speed), so I am guessing there's a planetary gear system between the motor and the pump. If there can be a sensor on the motor shaft, the precision of the pump output can be multiplied by the gear ratio. Assuming a gear ratio of 36:1, that would make the slice thickness accuracy much better. Given the torque load of the pump rotor, I expect the motor doesn't coast very long after the power is shut off, maybe one revolution or less. Instead of a sensor on the shaft, it may be possible to detect the motor armature transitions in the motor current or voltage. Depending on the number of segments in the armature, that would provide even greater precision.

True, but you'd still need to keep in mind the number of rollers squeezing the pump-tube and their relative position to the end of the housing (ie where it stops pinching and allows the tube to return to normal diameter)

[DoulosDS]

Or a conductive disk turning around with the axle with a drag-contact and a non-conductive area... Ofc. that won't help much if it turns too fast...

Hall effect is much more reliable and I expect a less noisy signal... as well as able to handle higher RPMs.

[DoulosDS]

True, but you'd still need to keep in mind the number of rollers squeezing the pump-tube and their relative position to the end of the housing (ie where it stops pinching and allows the tube to return to normal diameter)

Right; the motor controller would just keep the pump stopping point from drifting or dithering. The pump would always turn an integer multiple of 1/n for n lobes. My idea of sensing the pump lobe position optically would allow too much slice-to-slice variation. If the armature modulation counter goes over the number of armature bars, the controller could report an error or even correct an error if somehow the motor revolution counter lost count.

[oninoshiko]

Hall effect is much more reliable and I expect a less noisy signal... as well as able to handle higher RPMs.

Hall effect sensors are also cheap. I'm doing some experimenting with MLX90363 units, but you really wouldn't even need something that complex for what you're proposing. something like a US1881 (which sparkfun carries) should work.

[DoulosDS]

I suspect the most likely problem will be how to attach a magnet to the motor shaft. The photos don't give me much hope at the commutator end of the motor, unless the magnet were very small and could be glued to the end of the shaft with a droplet of epoxy. I'll have to open it up to see if the gear train provides a better magnet mounting opportunity. Maybe the gear on the output shaft will have a suitable spot. It will be a few weeks before the pump arrives here from China, so I won't know until early May.

[CescoAiel]

I suspect the most likely problem will be how to attach a magnet to the motor shaft. The photos don't give me much hope at the armature end of the motor, unless the magnet were very small and could be glued to the end of the shaft with a droplet of epoxy. I'll have to open it up to see if the gear train provides a better magnet mounting opportunity. Maybe the gear on the output shaft will have a suitable spot. It will be a few weeks before the pump arrives here from China, so I won't know until early May.

Depending on the stator/rotor confuguration (and whether it is brushless or not) you may even be able to sense the rotor rotation from the outside, without attaching any magnets... Something you could test before even attempting to open it and attach a magnet anywhere...