On Friday 14 February 2020 11:14:18 Chris Albertson wrote: > Looking at the PDF, It seems I forget to ask one question about the > signal: Is the sin/cos signal single-ended or differential? In other > words is it referenced to ground or do we get pins with labels like > sin+ and sin-
The coils are usually isolated, with op-amp bias supplied to one end of the coils, so you get sin plus and minus and cosine + and - > If there are even 100 pulses per revolution and the spindle runs at > 100K RPM then we are talking about 10000K or 10 MHz. At that point I > start thinking about hardware quadrature decoding rather than software > or in the Linux CNC world they use FPGA chips. Where would one find a 100k rpm device with an encoder. > But without understanding the signal it is way to early to even think > about a design solution. Let's see what's on all the wires first. IDK about resolvers but the quad in a 5i25 shouldn't have a problem with a quad signal at 10 MHz. > On Fri, Feb 14, 2020 at 7:29 AM dave engvall <dengv...@charter.net> wrote: > > a bit overdone for your app but this should give you an idea. page > > down to fig 16. > > > > > > https://www.st.com/content/ccc/resource/technical/document/applicati > >on_note/group0/6a/82/83/37/61/69/4e/74/DM00497286/files/DM00497286.pd > >f/jcr:content/translations/en.DM00497286.pdf > > > > Dave > > > > On 2/13/20 11:01 PM, andrew beck wrote: > > > Chris. I'm still learning electronics. Could you expand a bit on > > > this please. > > > > > > Maybe draw a napkin sketch of it. Sorry to be a bit slow on the > > > uptake > > > > > > On Thu, Feb 13, 2020, 5:46 AM Chris Albertson > > > <albertson.ch...@gmail.com > > > > > > wrote: > > >> A sin and cos are 90 degrees apart. All you should need to do is > > > > threshold > > > > >> the signal and you have A/B quadrature. Many ways to threshold > > >> it but you want the one with least noise. > > >> > > >> A simple way to convert a sin wave to a square wave is to amplify > > >> then > > > > clip > > > > >> it with diodes. A comparator can also convert the signal. > > >> > > >> The point to remember is that sin/cos is quadrature and all that > > >> is > > > > needed > > > > >> is some signal conditioning. > > >> > > >> On Tue, Feb 11, 2020 at 10:46 PM andrew beck > > >> <andrewbeck0...@gmail.com> > > >> > > >> wrote: > > >>> Hi guys > > >>> > > >>> wondering if anyone has any ideas here. > > >>> > > >>> I have a heidanhain spindle motor that runs up to 10000 rpm and > > >>> has a > > > > 5v > > > > >>> sin cos encoder on it. I am currently controlling the motor > > >>> with a schiender vfd. I am talking to the support engineers > > >>> here in New > > > > Zealand > > > > >>> about buying a encoder card so I can get better low down torque. > > >>> If I > > >> > > >> run > > >> > > >>> the card in full encoder closed loop control in the vfd I can > > >>> get 200 percent of the torque right down to 0 rpm for 30 seconds > > >>> or so which is pretty useful. I am currently just running the > > >>> drive in Variable > > >> > > >> frequency > > >> > > >>> control which rapidly looses torque at low rpm. > > >>> > > >>> Anyway they have a bunch of cards I can use but don't have a > > >>> encoder > > > > card > > > > >>> that is suitable for sin cos encoders. I have no trouble > > >>> changing the encoder but am not sure if I can get a source of > > >>> encoders that spin up > > > > to > > > > >>> 10k rpm. I am spinning a $20 1000 line encoder on the rear of the spindle motor. That is a diff output I'm making into rail to rail sq waves with rs485 convertors. Wide open at the motor is still 4000 edges per turn. In high gear the spindle makes 3k revs, 750 more than OEM. The gear ratio in the head in high gear is a bit over 7.1 so at 3g indicated that 4000 edges = something in the 70k area, which is absolutely not a problem. This is with the A/B coming from the motor, but the index coming from an opto wheel under the drawbar bolt. gear ratios determined by counting A signals for 100 turns of the spindle, so I use 2 different scale values controlled by switches on the rim of the gearshift knob. That means everything is accurate to a very small fraction of 1%. And it works very well even at Pgains that totally hide any load induced slowdown, I am not aware of any overload until the currant limiter in Jon's (pico systems) pwm-servo which is what I am driving that motor with from a 126 volt home brewed psu that needs at least a 20 amp supply breaker. Its set for around 17 amps making the iron in the motor squeak, at close to double that motors nameplate FLA. Don't ask why I haven't stripped the plastic gears in the G0704's head, I don't know and they're now nearly 5 years old. Since that driver is a true 4 quadrant driver, I took advantage of the knob switches to run the motor at about 35-40 rpms when neither switch is closed, so when the spindle is running but in the wrong gear, I can reach up and change gears quite safely as when neither switch is closed the motor is down to that idle speed in less than 200 millisecs, the psu has absorbed the braking energy, jumping to around 150 volts, the the gears can slide into engagement with the next gear without any noises, and by the time the other switch roller drops into the slot, closing the other switch which allows to commanded speed back to the driver, its 99% engaged, and 200 ms later its back at the commanded motor speed but in the other gear, and that huge higher voltage is used up by being fed to the motor to bring it back to set speed and everything is back to normal. But don't try this with a switching power supply as it will be destroyed by the stored energy being pumped back into it. Thats an analog supply with enough microfarads for storage that I have to soft start it with a 100 ohm 200 watt resistor for about 3 seconds else it trips a 30 amp breaker when its turned on. 4 huge 600 watt toroids and 65,000 uf to charge up. Information overload? Probably but it demos that anything you want LCNC to do is just a matter of hacking up the hal code and wiring it up to do it. The machine its running is more of the limits than LinuxCNC will ever be. Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis Genes Web page <http://geneslinuxbox.net:6309/gene> _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users