An update... I spent the afternoon tweaking software and air pressure etc. Still haven't been able to get the 3/8" impact deep socket from falling off. The dimple in the drive is deeper and the set screw definitely locks tight but after a few cycles the screw is loose again. Eventually the socket flies off.
Restricting the air exhaust did slow down the full speed but not enough without some major mechanical changes. With that in mind I took another look at the PWM approach but this time I followed the suggestion to do PWM which was 10% and 20% of 100mS. The valve still has time to open and supply enough so the motor doesn't turn super fast (although too fast still) but also has torque. The problem is when using a torque wrench to set 18 ft-lbs the 10% wouldn't budge it. The 20% setting does but once it's a tiny bit loose it will turn 4 or 5 turns. It was reliable enough to not drop the R8 most of the time. But it would take a lot of tries at 1 second per cycle to move it a tiny bit, then another tiny bit, and so on. And if it's lifted a touch and the magnet isn't at the right height because it didn't sit fully on the head of the draw bar then it spins for the entire timeout period. Make that time too long and once again the R8 drops out. Would be better if the indexing was on the drawbar nut but that raises other issues with the quill moving down and the nut descending downwards into the center of the drive pulley. Reflective sensor on the round bottom of the nut would work. As it is the nut is getting pretty trashed so until the 12 point socket arrives I think this part of the project is on hold. John > -----Original Message----- > From: ertleyd--- via Emc-users [mailto:emc-users@lists.sourceforge.net] > Sent: November-21-21 12:04 PM > To: 'Enhanced Machine Controller (EMC)' > Cc: ertl...@yahoo.com > Subject: Re: [Emc-users] Closed loop control of air motors. > > > you could try putting the air valve "very close" to the inpact. > this will reduce the air volume betwen the valve and inpact. > a valve that closes when off, creating a vacuum in the inpact input line. > this vacuum may help control the inpact over run. > low input hose volume and a vacuum when shut off. > > the use of 2 valves in the inpact input line would be able to control the > volume > of air that is avable to releace the collet. you may need a small adjustable > accumulator or longer/larger line between the 2 valves to increace the air > volume > so as to start the inpact hitting hard. > > good luck > Happy Thanksgiving > And Be Safe > > dale > > On Sunday, November 21, 2021, 01:12:19 PM EST, John Dammeyer > <jo...@autoartisans.com> wrote: > > Thanks everyone.� Definitely something to think about.� As it turns out, > sitting on a shelf somewhere, I have an old RIGID 14.4V > impact driver with a failed battery pack.� When the pack failed a new pack, > only available by mail but not allowed to be shipped by > mail due to lithium issues, was more expensive than a replacement impact > driver kit on sale. > > The video here shows a slow motion behaviour of the internals of an impact to > drill adaptor. > https://youtu.be/f0gSJa3L_7c > > So that opens up possibilities. > > I will try my existing impact driver on the draw bar with a TTS holder and > then use a torque wrench to see what it takes to undo it. > > As I recall, using the larger 3/4" torque wrench 20 ft-lb was way more than I > ever tighten the system now which is why the stepper > motor and 10:1 planetary drive looked interesting.� Especially since my > electronic lead screw has a built in 3A stepper driver (Z axis) > and encoder index sensing circuit (Spindle Index).� > > I had thought about using my ELS as a remote Pendant for Drawbar Control and > Tool Changing etc.� One of the reasons I used this > GIM module as the tool changer controller as I now have indictors on the LCNC > screen showing the status of the draw bar all with > CAN bus and CANopen protocol.� And in place of the X axis drive my ELS has > CAN bus too but it didn't have enough outputs for relay > driving without an additional expansion board or BoB.� In fact I can send CAN > messages to my GIM tool changer to trigger the load > release feature too so it could be done from LCNC or the ELS. > > The GIM has two DIP relays and two relay drivers and an optically isolated > input connected to the PIC30F5011 interrupt input.� Plus > some inputs for the switches so all in all enough for the pair of pneumatic > valves and the other I/O.� It lets me test it and program it > on the bench without the need for LCNC running or connected.� Being able to > set break points etc. is very handy for embedded > systems development. > > And I do have the Pi4 with LCNC and a Lawicel CANUSB to test the CAN side.� I > used the CANUSB rather than a HAT for the Pi4 so I > could also use CAN on the Mill PC in the shop. > > Really the only reason I went with the pneumatic butterfly wrench is because > I bought it and paid for the drawings document over > 10 years ago.� That's where the basic design came from.� I just redid them in > Alibre and then cast the parts and machined them with > LCNC after first testing with 3D printed parts. > > John > > > > > > -----Original Message----- > > From: Gene Heskett [mailto:ghesk...@shentel.net] > > Sent: November-21-21 7:54 AM > > To: emc-users@lists.sourceforge.net > > Subject: Re: [Emc-users] Closed loop control of air motors. > > > > On Sunday 21 November 2021 03:24:24 John Dammeyer wrote: > > > > > Actually that is along the lines of what I've asked.� Some sort of > > > brake that could limit the speed. > > > > > > If you go back to steam engines the two weights swung out and shut off > > > the pressure to the engine to keep it from running away.� I'm not sure > > > I could fit something like that in that space but I had an idea like > > > that. > > > > > > Occasionally, until I solve that problem the socket goes flying off > > > too.� The square shaft of the wrench is hardened and I've used a > > > Dremel to create a dimple but it's not deep enough yet.� Adding a > > > guard around the socket for safety is a good idea then.� And if > > > centrifugal force causes two arms with brake pads to fly out and run > > > against the guard to slow it down that would work.� But seems overly > > > complex. > > > > > > Using that same guard idea but with pads that apply load to the socket > > > to keep it from spinning away might be easier.� If you think about the > > > sound of an impact wrench it spins up and only when it hits a load do > > > you get the hammering sound. > > > > > > I've been considering this as a solution instead. > > > https://www.aliexpress.com/item/4000220757109.html > > > > > > John > > > > > Even at 10/1, I doubt it could generate enough closing torque. The air > > wrench is probably too much anyway. It can probably strip the drawbolt > > > > I use a 20 volt battery powered Porter Cable impact screwdriver which has > > plenty of closing torque due to its impact, and my reflexes are fast > > enough to prevent the overspin and r8 collet drop. I don't see why that > > idea couldn't be timed to turn it off when the magnet comes by the > > second time. An H bridge such as the $7 ebay thing I used to drive that > > BS-1 clone can switch 25 volts at 43 amps, shouldn't have a problem > > replacing the trigger switch and would also serve as the reversing > > switch without having to punch the direction button. Total electric > > conrol by a pair of signals from the bob, crowbarring the motor to stop > > it quick. Controls for that are a 30 minute edit and test in linuxcnc. > > Unless busier than that cat on that famous tin roof, charge/swap the > > battery about monthly. On my g0704, about a seconds hammering and theres > > no tool slippage, 10 seconds would likely strip or break the drawbolt. > > Put the H bridge right on the tool so 10 gauge leads are short for low > > loss. Since you have the air rigged to drop/lift it, keep that but I'd > > tally the down position and interlock the logic so the only thing you > > could do when its up, is position it to put the magnet in the correct > > position for the next unlock cycle. > > > > Your pix gives me an idea to make the engagement slider out of an old > > plunge router base, printing a collar to adapt the impact tool to the > > router base.� All I'd need is the giddyup to get off my duff and do it. > > But first, straighten the crooked post on that BBLB POS. Until I do that, > > its looking to be replaced and used for salvage. > > > > > > -----Original Message----- > > > > From: Roland Jollivet [mailto:roland.jolli...@gmail.com] > > > > Sent: November-21-21 12:14 AM > > > > To: Enhanced Machine Controller (EMC) > > > > Subject: [Emc-users] Closed loop control of air motors. > > > > > > > > It's not what you asked, but what about a centrifugal > > > > clutch/limiter? Above say 200RPM it could swing outwards and lock, > > > > only releasing again once the torque is removed. > > > > > > > > On Sun, 21 Nov 2021 at 09:45, John Dammeyer <jo...@autoartisans.com> > > wrote: > > > > > Has anyone ever tried any sort of closed loop control of an air > > > > > motor similar the one on this butterfly impact wrench. > > > > > https://www.penntoolco.com/52-424-9/ > > > > > In order to not have both the TTS and R8 collet drop out of the > > > > > spindle I'd like to be able to limit the number of turns to 2. > > > > > > > > > > What I've found however, is that 10mS pulses to the valve > > > > > supplying air for CCW rotation barely move it until it's unloaded > > > > > and then it immediately spins up very fast and does at least 4 to > > > > > 8 turns.� That's way too many. > > > > > > > > > > I've only got a one pulse per rev hall sensor detecting a magnet > > > > > on the socket.� But I suspect between reaction time of the valve, > > > > > the compressibility of air and the rotary inertia of the motor > > > > > that unless there is some sort of constant load that control is > > > > > pretty well impossible. > > > > > > > > > > With servos, if the motor is loaded and then suddenly totally > > > > > unloaded, it too might turn a number of revolutions before the > > > > > control system could bring it back down to the original speed. > > > > > > > > > > Short of adding some sort of mechanical brake am I trying to do > > > > > the impossible?� Oh and to avoid another 1000 words to explain > > > > > I've added a picture. > > > > > John > > > > > > > > > > _______________________________________________ > > > > > Emc-users mailing list > > > > > Emc-users@lists.sourceforge.net > > > > > https://lists.sourceforge.net/lists/listinfo/emc-users > > > > > > > > _______________________________________________ > > > > Emc-users mailing list > > > > Emc-users@lists.sourceforge.net > > > > https://lists.sourceforge.net/lists/listinfo/emc-users > > > > > > _______________________________________________ > > > Emc-users mailing list > > > Emc-users@lists.sourceforge.net > > > https://lists.sourceforge.net/lists/listinfo/emc-users > > > > > > 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, 1940) > > 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 > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
