Super-PID new low-cost router speed controller

[RomanLini]

Hi, after weeks of development I'm finally allowed to discuss this project. As some of you may know I've been an electronics product developer for a lot of years, and a CNC hobbyist for not so many years.

A couple of months back I created a closed loop router speed controller for my home CNC router, my router is; (here.

The controller uses a reflective opto sensor pointed at the router shaft and a microcontroller to give full PID control. Seeing that the system could be made to work and work quite well, started a discussion with Val from VHiPe.com who does some small quantity electronics manufacturing for me (and I do some contract design for him).

Val is in the process of arranging a forum name and organising to become a site sponsor and the official spokesman for the SuperPID kit product, but he's an oldtimer, not really an internet guy so I will be holding his hand a little regarding product enquiries until he gets more savvy with the forum process. I will be available anyway to answer tech inquiries for the product operation.

After some weeks of testing and refining the Super-PID product is near completion and production underway.

My home PID prototype is seen below, I should discuss how it works and then the differences for the commercial model.

There is a small reflective opto sensor about 5mm diameter that points at the router shaft from about 2mm to 3mm distance. Then a white spot of paint is painted on the shaft. I just used liquid paper for testing (matte white on the black metal shaft) and it has worked without incident for many weeks now so I didn't bother buying proper white matte paint.

The control board detects the router speed and adjusts the power supplied to the router to maintain a constant speed using an industrial PID algorithm. Speed is varied by a single large knob. This means the router can be set from speeds 5000-30000 RPM and will maintain a constant speed even when cutting etc.

The PID algorithm I used is really PID+ as it has much in the way of modification and special additions to improve stability under the very difficult condition of low speed cutting with a router that has very low rotor inertia (low flywheel effect) and only one sensor pulse per rotation, AND only being able to control the mains power pulses at 100 (or 120) times per second. Adding to these issues was the fact that cutting tools like single flute put one huge load pulse on the rotor per rotation which is again a problem very hard to stabilise.

Anyway the weeks of programming and testing are done, the thing works very nicely on my 1/4" router and even a bad case like a little Dremel(!) and should work even better still on all larger diameter routers that have a higher rotor inertia.

My prototype Super-PID has an accurate LCD tacho, and so does the Super-PID. However on mine the "power bargraph" is a row of 10 LEDs, where on the Super-PID it is a much more accurate 48 bar bargraph on the lower half of the LCD. This feature shows the percentage of power applied to the router at all times. So you can see on my prototype at low revs (5000 RPM) there is only one bar lit, the router is being fed around 10% power to maintain speed.

If it starts cutting, the power bargraph will rise to maybe 20-30% depending on the depth of cut and this is shown on the bargraph even though the speed stays at 5000 RPM or very close.

The 2 main benefits of the PID system have been very quiet opration, many of my jobs I can now cut at 10000 RPM or less with the router almost silent, and the other benefit is that the low revs have significant power now so I can take deep cuts in low-melt plastics like polystyrene and leave a clean finish with nice large chips thrown off.

The Super-PID kit will be sold as an assembled and tested PCB with LCD tacho/power display and screw down terminal blocks for the wires. It has 120v and 240v models (to suit all countries) and currently it is planned to sell as "kit" which just means you need to put in in a safe box as it uses mains voltage for part of the circuit (the rest of it's circuit is fully opto-isolated so you can control it with Mach3 etc). The "kit" form with no box was chosen to allow a cheaper price (as this is a low-cost alternative to VFD Spindles).

I'm open for questions and hope this is all ok with the moderators.

There are no proper product photos for a week or so as PCBs are currently being made - product due mid Oct. My prototype that has seen about 6 weeks hard use is shown below;

[DeadTom]

That is just awesome, I think if the price is in my range I will need to get one for my new machine build in the very near future.
I have a question regarding the type of spindle that this will control: I am thinking of using a large brushless DC motor for my spindle motor can your SuperPID be made to work with the electronic speed controller that will be needed to drive the brushless motor?
Tom

[CarveOne]

This sounds like a very useful item. I'll keep an eye on this thread with an interest in buying one also.

Would it help to make the router run smoother at low rpm if there were two pulses per revolution? Instead of one white dot, use two short white lines on each side of the shaft that are wide enough to be detected reliably.

CarveOne

[JeremyFisher]

Looks promising, +1 on the interest in buying.

What sort of power can it handle?

J

[RomanLini]

Thanks for the interest guys.

DeadTom- The product target price was to be under $150 for the fully assembled unit, possibly it could be as low as $120 or $130 but costing is still waiting on one of the parts suppliers (quantity LCDs) and will also depend on some time trials of how long the parts assembly takes. The PCB is a top quality Australian made unit, Australian labour and parts top quailty (metal film resistors, over-rated power componenets etc) so this was designed for the quality end of the market, not bottom-end Asian price etc.

Sorry no it is not suitable for BLDC motor, and your BLDC controller should include speed regulation anyway? With a router you can run lower speeds (like a BLDC) and fairly quiet.

CarveOne- There is little point running a finer resolution speed sensor. For one, it causes issues if the 2 dots are not perfectly 180 degrees apart, and that makes it hard to do. Options like multi-slot encoder disks are hard to fit! Also the main low speed issues are that the router inherently has low power at low speeds AND that the mains power comes in low speed pulses (100Hz or 120Hz) - neither would be fixed by encoder resolution.

All that is irrelevant now as I have refined the PID system to work well within it's design parameters 5000-30000 RPM and the extremely easy to utilise reflective opto sensor system of 1 pulse per revolution.

I can possibly talk Val into giving a few units away to long term forum members who post good photo build logs (like yourself) so I may be able to get you a free one in exchange for a review/testing. That's pushing the limit of forum advertising etc so I don't want to go into that yet until Val sorts out the forum sponsorship and hopefully gets some feedback from the moderators etc as to the best way to proceed.

JeremyFisher- It is rated for continuous use at up to 1000W (120v USA model) or 1200W (240v Europe/Australia model). Heatsinking needs are minimal, but if you're like me and like to go extra on reliability and plan to use it at higher power levels for production a small additional heatsink and/or small fan is recommended.

Thanks again for the interest guys! Later today or tomorrow I plan to pull my routers vac foot off and take a photo of the spindle opto sensor and how I mounted it, that will help to show just how easy it is to attach this to your existing router setup.

[CarveOne]

Ok, I wasn't sure about whether the circuitry could handle 2X the pulse rate, and if there were any advantages if it could handle it.

Thanks for the offer. I may be interested if you need the help. PM me when you are ready to discuss it. Many of your sales may be from the US and other countries. You'll need to figure out how to handle that.

CarveOne

[Khalid]

I can possibly talk Val into giving a few units away to long term forum members who post good photo build logs (like yourself) so I may be able to get you a free one in exchange for a review/testing. That's pushing the limit of forum advertising etc so I don't want to go into that yet until Val sorts out the forum sponsorship and hopefully gets some feedback from the moderators etc as to the best way to proceed.

This is a very nice gesture..

[ger21]

Will this work with variable speed routers, or only fixed speed?

[brtech]

Seems like your next step is to have a modbus interface to Mach so gcode can set the speed of the spindle.

[RomanLini]

CarveOne- We expected to sell to the US as a main part of the market. Australian airmail is reliable and quick enough, it is typically 5 to 10 days to get a small parcel to the US.

Khalid- Thanks.

Ger21- It it designed for routers that have no inbuilt speed control.

In the early testing I removed my router cover and checked, its electronic speed control has a switch pot, so it varies up to full, then the switch goes "click". At that point the switch bypasses the electronic speed control so there is no problem. That is probably a common setup. Also, I actually did some cutting with the router's internal speed control on, and that seemed to make no difference to the operation of the PID unit, they did not interfere.

All in all I would recommend turning the routers internal speed control off, or bypassing it, but I do plan to do more tests. Thanks for reminding me.

Brtech- Thanks for the suggestion. Currently it will accept speed control from the pot, or from a 0-5 volt control source. I think some of the break out boards etc can produce a 0-5v speed control signal.

One thing we have decided since the PCBs went to press is to use a socket on the microcontroller, so it will be possible to add features later and get users to mail the chip back for an "upgraded" chip.


Photos of my protoype speed sensor.

Nothing special to look at, just hurriedly applied white paint (liquid paper) I painted on half of the shaft, and a tiny reflective optical sensor beam pointed at it;

[RomanLini]

I have been told today the PCBs are finished and with the courier.

So within a couple of days I can get a unit assembled and take some photos to show what the product actually looks like, and maybe borrow a vid cam and do a youtube of the little beastie in action.

In the meantime I did get time to add a few nice firmware display features and improved the soft start-up RPM ramp.

[JeremyFisher]

I was actually going to ask if you could implement a start-up ramp.

I have another question about the unit though. I have a weak understanding of electric motors so excuse me if this is a silly question but I'm sure others are wondering the same:
Can running a motor (which is designed to run from mains) at lower voltages (or on PWM current or however your unit works) have a detrimental effect on it?

The main concern I have is overheating of a motor which is running below its rated RPM and therefore has less airflow running through it while outputting the same power.

Have you run your router under load for extended periods at 5000 RPM?