Greetings all,

You probably don't remember but several months ago I asked about switching from Mach 3 parallel port to Kflop running from Kmotion and after yet another tragic event due to Mach 3 in the shop a few days ago, I finally pulled the trigger and Bought a Kflop. My plan is to get the Kflop up and running via Mach 3 as soon as is reasonable for me and then over the next few months, to switch to Kmotion CNC full time.

I've been searching the various forums and archives and have found a lot of answers to my questions but there are still some that I have that I was hoping some folks could help me out with here or perhaps just point me to the source of the answer. First, a little about my setup.

3 axis ganry router with a 4th axis that I haven't really used yet.

G540 for drives and I/O
VFD using 0~10v from G540 and start/stop signal.
3 hall effect home switches on one input.
Home made pendant for Cycle Start and Feed hold running into two inputs
Contour Design shuttleXpress for jogging and MPG (USB device, uses a plugin to Mach 3)

In addition to this, I'd like to add encoders to my setup, at least for the 3 primary Axes.

So, my questions are as follows:

Connectors: What type of connectors are used on the Kflop? They look like old disk drive connectors of some sort but I can't remember their type for searching online. Ideally I'd find them with a ribbon pigtail already on there.

Encoders: I'm thinking of getting these encoders as they're reasonably priced and come with a cable: http://www.automationtechnologiesinc.com/products-page/servo-motor-encoders/us-digital-shaft-mount-optical-encoder-for-nema-23-servos

Do I need to get some sort of BOB to add opto isolation for the inputs on these or can I run them straight into the Kflop?

With my G540, I get 2000 pulses per rev but these encoders after quadrature will only give 1440ppr. Will that work or do I have to search harder for encoders?

Can I have 3 axes with encoders and 1 without in Kmotion CNC?

I'll leave it at that for now - as far as I can tell, that's what I need to get ordered to get started at least.

In case you're wondering, the tragic event was that I loaded a very large program into Mach 3 and apparently, it sometimes has a rough time dealing with those that the timing gets screwed up. The machine stalled on a retract and ruined a customer's guitar body blank…$100 down the tubes for me.

Anyhow, thanks in advance for the advice and please be patient with my questions! I haven't written a shred of C in about 20 years!

Best,

-Andy Birko

 

Hi Andy,

That all sounds reasonable.  Except the USB MPG.  We don't support those.  See these posts:

http://www.cnczone.com/forums/dynomotion_kflop_kanalog/169803-kflop_pokeys_together.html#post1209926

http://www.cnczone.com/forums/dynomotion_kflop_kanalog/204522-lots_questions-3.html#post1417282

The connectors are standard 0.1 inch pitch IDC connectors and 0.05inch pitch ribbon cable available from many sources.  Winford.com sells them and simple breakouts.  Like this:

http://www.winford.com/products/brk2x13.php

Those encoders look reasonable.  I would purchase those with the differential outputs.  See:

http://dynomotion.com/faq.html#Encoders

Resolution is not critical or need to be matched.  Higher is always better unless it results in a higher count rate than 1MHz.

Optical isolation is always a good idea.  KFLOP itself only accepts 3.3V LVTTL signals.  Consider our Konnect board with 48 Opto I/Os.

http://dynomotion.com/Konnect.html

Not all axes need encoders.

I wish my tragic events only resulted in $100 loss :}

Good luck!

Regards

TK

Group: DynoMotion	Message: 9027	From: andy.birko	Date: 1/27/2014
Subject: Re: Finally Bought a Kflop!

--- In DynoMotion@yahoogroups.com, Wcarrothers Yahoo <wcarrothers@...> wrote: > > Wow. That is a lot. I could not read it all > > But I wonder what kinda accidents people have with Mach and why. I've been running it parallel, smooth stepper, and with kflop on different machines since 2006 and with no accidents than I didn't cause or were not properly caught by he limits What happened to me is something that's happened to me before but I could never figure out why. I'm running Mach 3 on an older machine but it's not ancient or anything. The guitar body had a lot of surfacing involved and the program was over 700k lines. My first operations were to face one side and the machine stalled while cutting at only 100ipm (my machine cuts just fine at 400ipm) which should have been my first clue. I proceeded anyway thinking it was just some weirdness. Later, while profiling the sides, I got a similar stall on a z-retract and then the machine tried to rapid and gouged the guitar body by about 1/8" ruining the project. As part of my investigation I ran the program in air and got stalling at the exact same lines every time I ran it. On some advice from a friend, I ran the program with the "toolpath display" set to off and the stalling went away .lovely

Group: DynoMotion	Message: 9029	From: andy.birko	Date: 1/27/2014
Subject: Re: Finally Bought a Kflop!

Thanks for the advice, Tom, --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote: > Optical isolation is always a good idea.  KFLOP itself only accepts 3.3V LVTTL signals.  Consider our Konnect board with 48 Opto I/Os. > http://dynomotion.com/Konnect.html The Konnect looks fantastic but the budget is very tight on this project and you know how costs can get out of control! While searching around, I found these: http://numato.com/opto-isolator-breakout which look like they might do the trick. What do you guys think? Otherwise, I'll probably buy a C10 or something from Automation Technologies and just remove the DB25 connector. Shame about no USB shuttle/MPG control. The shuttle wheel on the ShuttleXpress is really nice for lining up certain parts. I can live without it though, might just add an MPG through Kflop with some axis select buttons or something like that.

Group: DynoMotion	Message: 9030	From: Wcarrothers Yahoo	Date: 1/27/2014
Subject: Re: Finally Bought a Kflop!

Yes with super large programs you should turn it off. Although most of mine are small runs.  I have hardly ever turned it off.   Although.  All my Cnc machines have been pentium 4 boxes my first on machine 1 was a p3 machine which worked real good.   B On Jan 27, 2014, at 9:24 PM, "andy.birko" <andy.birko@...> wrote:   --- In DynoMotion@yahoogroups.com, Wcarrothers Yahoo <wcarrothers@...> wrote: > > Wow. That is a lot. I could not read it all > > But I wonder what kinda accidents people have with Mach and why. I've been running it parallel, smooth stepper, and with kflop on different machines since 2006 and with no accidents than I didn't cause or were not properly caught by he limits What happened to me is something that's happened to me before but I could never figure out why. I'm running Mach 3 on an older machine but it's not ancient or anything. The guitar body had a lot of surfacing involved and the program was over 700k lines. My first operations were to face one side and the machine stalled while cutting at only 100ipm (my machine cuts just fine at 400ipm) which should have been my first clue. I proceeded anyway thinking it was just some weirdness. Later, while profiling the sides, I got a similar stall on a z-retract and then the machine tried to rapid and gouged the guitar body by about 1/8" ruining the project. As part of my investigation I ran the program in air and got stalling at the exact same lines every time I ran it. On some advice from a friend, I ran the program with the "toolpath display" set to off and the stalling went away….lovely

Group: DynoMotion	Message: 9031	From: andy.birko	Date: 1/27/2014
Subject: Re: Finally Bought a Kflop!

--- In DynoMotion@yahoogroups.com, Wcarrothers Yahoo <wcarrothers@...> wrote: > > Yes with super large programs you should turn it off. Although most of mine are small runs. I have hardly ever turned it off. Um, yea, I know that now. I still think it's unacceptable for a CNC control package to crap out like that with no warning. I read the Mach 3 documentation cover to cover and even if the warning is there, I have no recollection of it. If the program is struggling, it should warn you *before* you ruin the part. Better yet, it shouldn't fail just because the program is large.

Group: DynoMotion	Message: 9042	From: andy.birko	Date: 1/28/2014
Subject: Re: Finally Bought a Kflop!

A few more questions --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote: > > Hi Andy, > > That all sounds reasonable.  Except the USB MPG.  We don't support those.  See these posts: I think I can live with that as I survived without any sort of shuttle for a few years. What types of pendants/shuttles/MPGs are folks using with the Kflop? > > Resolution is not critical or need to be matched.  Higher is always better unless it results in a higher count rate than 1MHz. > In looking further, it seems that a 1000cpr encoder can be had for not a tremendous amount of money which would give me 4000ppr or double my steps/rev. According to my calculations, at 1000ipm (which my machine certainly can't do) I'd be at 2564 RPM (10mm lead screws) which would give a pulse frequency of about 171kHz / axis. Multiply that by four and you get a count rate of 684kHz which sounds like it's safely below the 1MHz limit of the Kflop. That got me thinking: US-Digital has a big price break at quantities of 5+ and again at 10+. Would anyone here be interested in a group buy for these encoders? (hopefully that's ok to propose on this board). I'm looking at these encoders with the following specs: http://www.usdigital.com/products/encoders/incremental/rotary/kit/E5 1000 CPR (4000ppr) 1/4" bore No Index Single Ended Output Hole in Cover Default Base (figure out mounting yourself - either epoxy, double sided tape, etc ) For this configuration, the encoders come out to $52.14 each. $43.11 if we got to a quantity of 50 Additionally you'd need to install the encoders 1 x .050" hex wrench for <$1 1 x Centering tool ~$5 Cables (optional) at around $14 each for 6'. Connector (make your own cable) for $2.90 each. I'd handle the ordering, re-packaging and whatever else is necessary, maybe tack on $5/order to cover shipping to me. Shipping to you would be via USPS flat rate box, medium box is about $13. Any interest in such a group buy? -Andy Birko

Group: DynoMotion	Message: 9046	From: Tom Kerekes	Date: 1/28/2014
Subject: Re: Finally Bought a Kflop!

Hi Andy, If you follow the posts listed above they describe a common MPG used with KFLOP.  here they are again: http://www.cnczone.com/forums/dynomotion_kflop_kanalog/169803-kflop_pokeys_together.html#post1209926 http://www.cnczone.com/forums/dynomotion_kflop_kanalog/204522-lots_questions-3.html#post1417282 Checking your encoder math I get: 1000ipm = 25400mm/min = 423 mm/s @ 10mm/rev = 42.3 revs/sec With encoder 1000 cycles/rev = 4000 quadrature counts/rev 43.3 x 4000 = 169,000 quadrature counts/rev (much below KFLOP's 1MHz max) You seem to have an extra factor of 4. Unfortunately encoder manufacturers are not consistent in terminology so it is very confusing.  I think when USDigital refers to "pulses" they mean quadrature counts.  Most manufacturers refer to pulses as the same as cycles.  Maybe I confused you by using that term in that way earlier. You are welcome to promote group buys here in the group. HTH  regards TK Group: DynoMotion Message: 9048 From: andy.birko Date: 1/28/2014 Subject: Re: Finally Bought a Kflop! --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote: > You seem to have an extra factor of 4. Thanks Tom, I wasn't sure if the 1MHz limit was per channel or grand total so I multiplied by 4 just in case (for 4 axes) anyway you slice it, it's well below the speed limit of the Kflop. I'm curious though, obviously more resolution is "better" but what exactly does that mean? The cheaper encoders I linked to earlier had 1200 ppr (I remembered incorrectly in my first post). What kind of machine performance difference would you see between 1200ppr and 4000ppr? Group: DynoMotion Message: 9049 From: Tom Kerekes Date: 1/28/2014 Subject: Re: Finally Bought a Kflop! Hi Andy, All KFLOP's encoders can count at 1MHz simultaneously. Regarding resolution: it all depends on your system and requirements.  If you are just going to use the encoders to detect an Stepper miss step or stall it probably would make no difference.  A single stepper miss-step will result in a 4 full step shift.  But if you intend to use the encoder position for feedback positioning then the resolution would be more critical.  Ideally you would like to have resolution at least 10 times higher than what you are trying to position.  Imagine you are trying to position to within one mil when your resolution is one mil.  If the encoder says you are off by one mil it might really be off by only a tiny faction of a mil or almost a full mil.  And when trying to move the motor you have no clue if, how fast, or what direction things are moving until the encoder changes.  But now say the encoder reads tenth mils this give far more information on position and speed as the target is approached. I'm confused on your term ppr.  But assuming 300 cycles per rev = 1200 quadrature counts per rev and your 10mm pitch that would be: 10mm/1200 = 8.3um per count = 0.33 mils per count If mechanical issues are much greater than this then improving this will make little difference.  The point is higher resolution never hurt as you can always ignore it if you wish. Regards TK Group: DynoMotion Message: 9050 From: andy.birko Date: 1/29/2014 Subject: Re: Finally Bought a Kflop! --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote: > > Hi Andy, > > All KFLOP's encoders can count at 1MHz simultaneously. > > Regarding resolution: it all depends on your system and requirements.  If you are just going to use the encoders to detect an Stepper miss step or stall it probably would make no difference.  I figure that if I add encoders, I may as well use them for feedback positioning! > I'm confused on your term ppr.  But assuming 300 cycles per rev = 1200 quadrature counts per rev and your 10mm pitch that would be: > I got the term ppr from the US digital site, it's pulses per rev - basically 4 x Counts per rev. Thanks for the info though. It seems that 1200 would be good, but 4000 would be better. I cut a lot of very fine inlay and am finding some slop in the finished product that I'd like to get rid of as best I can. Not sure exactly where the error is coming from but some positional feedback should at least eliminate that from the search.