[tenmetalman]

Henrik,
as you stated earlier I seem to be one hell of a trouble magnet ! sense I last e-mailed you I managed to take a lightening strike while working on my mill. It killed my mill PC, blew a fuse on my second power supply, & damaged one of the modules you sent me. I have a old slow (1.3Ghz) PC that I've loaded XP & Mach3 on but have had to slow Mach3 down to 25000 vis 45000 where I was running. I have all three axes running, Z using a Uri chip & Y & Z using your modules. I'm rapiding at 45 inches & feeding at 25 inches to keep the computer from locking up while ghosting through the roadrunner program. I'll now go back to interfacing the spindle drive into mach3.... I'd like to send the damaged module back to you for repairs if thats OK. My X axis has always wanted to be jerky at any feed rate above ............ 10 inches or so & working under the module & 25 inches it's smooth as silk. in another e-mail I'd like to send the module settings I'm using, describe the pluses & minuses of the servo performance & see what you suggest. I'll get off my rump & order a motherboard & CPU that run in the 2.7/3 Ghz range & assemble it in the mean time
Awaiting your reply
Paul

[tenmetalman]

Henrik,
sense writing the above post last night I added a M47 to the end of the roadrunner program & let the mill loop through the program for well over a hour.
it preformed without complaint till I needed to go to bed before I fell off my stool.
to answer a earlier question on the list, I'm not using a smoothstepper. I've screwed up so many pieces of hardware I just didn't want to add a additional piece to the mix. as Henrik has pointed out to me I've brought allot of problems on myself by just scabbing things together in the past. sense Henrik has taken me under his wing I've spent hours cleaning up & cutting old wiring out my electrical cabinet. somewhere along the line I'd gotten terribly lazy & having developed bad back problems three years ago didn't help. so anyway the message is do it right or not at all.......
I also had to replace the spindle controller shortly after I got my mill home, had this point, maybe 4 plus years ago. I got the new phase converter up & running the spindle & unhooked it while I sorted out my drive/servo problems. today I plan to put the power back to the spindle, work out the control issues so that the analog spindle control segment on my C-11 BOB controls the spindle within Mach3.
Henrik,
I have modules #011, 012 & 013. I'll double check this morning but I believe it was #013 that the lightening messed up. it still indicates that when powered up all is well via the led & will talk to Hyperterminal. when I command a axis move in either direction (on whatever drive I plug it into) it seems to accept the first movement command. (axis does not actually move) if a second movement command is sent it faults out. I do my posting on Fridays & will package the damaged module & ship it then.
Paul

[H.O]

Hi Paul,

the message is do it right or not at all.......

Yeah, I tried to dress it as much as I could but apparently you read thru the lines and got the message ;-)

Seriously though, with the kind of trouble you've been having with hardware not only working properly but actually breaking, resulting in you having to throw more money on it's better to take a step back and sort one thing out at a time - properly. I'm glad it's all starting to come together for you although I'm sure my part in it is miniscule.

Weird problem with that damaged module...If there's been an overvoltage condition on the suply rail I would expect it to be dead, simply. Anyway, I'll have a look at it when it comes back, in the meantime I'll send you another as a replacement.

If you want me to take a look at your settings go ahead and post them. If you've managed to capture a move with the data recorder you can post that as well - it will give some insight to how the PID is operating. As always, let me know if there's anything I can do to help.

Thanks!
/Henrik.

Oh, you'll HAVE to post a picture or two of the machine and your setup, please!

[tenmetalman]

H.O. & all,
the silence on my end is due to my upstairs heatpump quiting over the weekend.
it's taken me a couple of days to figure out the problem, find out that I had to order parts from California (the other side of the nation, naturally !) & get them ordered this afternoon.
I'd been working on getting Mach3 to control the start/stop, direction & speed of the spindle using the analog 0/10vdc speed controller on my CNC4PC BOB. per the directions I'm using a optical voltage isolator to power the speed control. I'm using a NOS phase converter which that was made in Korea which probably isn't helping figuring out the interface. Mach3 thinks it's controlling the spindle but so far no go ! that's where I'm at..................... I'll be back with the the current settings I'm using with H.O.'s modules & a update on the spindle situation shortly
Paul
BTW, the new motherboard, a FOXCONN A74MK-L. the CPU a AMD Sempron 2.7Ghz single core processor arrived today. I assembled them, added 2 gig of ddr2 800MHz memory & will put it in the case shortly. I'm waiting on the Parallel card, a PCI Express x1 which still is in-route. the above is the core for the replacement PC for the mill controller.

[H.O]

Hi Paul,
Thing do tend to get in the way don't they ;-)

As for the spindle, eariler I sent you a schematic of how I have my C11 wiried, you can see how I did it there and perhaps get some pointers. You mention an optical voltage isolator to power the speed control, do you mean the isolator that is already ON the C11 board or do you have another one in the mix?

/Henrik.

[tenmetalman]

H.O.
see below, I messed this post up !

[tenmetalman]

H.O. ( I didn't think about the list reformatting all my numbers SORRY)
the replacement module arrived today. looks safe & sound. with everything else I've had going on your other module didn't leave here until yesterday ! In answer to your question I put a 12vdc to 12vdc isolator between the 2nd pc power supply I'm using and the analog spindle controller............ I've not spent much time in the shop sense my last post. I did go out there & write down the setting I'm using for the axes modules & Uli chip.
X axis
Module #00012
K= 159 (0)
I= 2 (0)
D=15900 (0)
T=1
V= 0
A= 0
X= 1
L= 6
E= 50000
C= 760
F= 0
O= 0
Y= 2.5MHz
J= 2
K= 0
Peak error (abs) 0

Steps per
31425.21
Velocity
42.102
Accel.
11.1932

Y axis
module #00011
K= 159 (0)
I= 2 (0)
D=15900 (0)
T=1
V= 0
A= 0
X= 1
L= 6
E= 50000
C= 760
F= 0
O= 0
Y= 2.5MHz
J= 2
K= 0
Peak error (abs) 1

Steps per
25240.96386
Velocity
33.696
Accel.
11.1932

Z axis
Uli chip #?
P= 1425
I= O
D= 250
H= 14250
X= 250
T= 255
I= 106
E= 30000
M= 1
Y= 70
0= 362
steps per
9000

All of the steps are rough settings using Mach3 & measuring with a pointer in the spindle to a machinist's rule so I'm sure their off by .070" or so in 5"

Step & Direction Pulse
2 & 2 All axes
this morning I was looking at any & all information trying to get a handle on the wiring & settings required between my VFD, the BOB & analog speed controller. If you look on the CNC4PC site at his C6 variable speed controller & then at the "Sample wiring of a C6 to a WF250Y Inverter" that's what I'm trying to sort out. In addition all of the VFD settings are setable ( if thats a word) I'm going out to the shop a give it another try using this information as a loose guide. "I'll be back" begining to sound like a threat Yet ;^O

[H.O]

Hi Paul,
First, regarding your settings. As long as it works to your liking there really are no wrong settings but here are a couple of things I noted.

You have the (K) parameter set to 0. What this means is that when the motor is within the number of encoder count specified by (J), (2 in in your case) the P, I & D gains are reduced to 0 with similar results to simply disconnecting the motor from the drive. When the error is becomes larger than 2 counts the gains are bumped up to the original values again.

If this IS indeed what you want then it's OK but I sugest you set to 256 to begin with. If you experience vibrations/dithering when the motor is stationary you can try lower this value in steps. The values displayed within paranthesis after P, I & D are the "scaled" values applied when the motor is within the number of counts specified by the (J) parameter.

(The (K)-parameter should default to 256, please double check that on the new module when you power it up).

Setting (E) to any value above 32767 (you have 50000) disables the following error trip limit meaning it will never fault no matter how far off target the motor is. I don't think this is what you want.

Apart from that and as long as it moves smoothly and positions correct I don't have any objections - although (I) seems pretty low.

If you command a long move at say 1/4 of full speed and then use GE a couple of times as it is moving it will show you how much it lags during the move. You could then try to add a little velocity feedforward (V-parameter) and see if you can get the "lag" reduced.

As for your spindle drive I'm afraid I can't be of much assistance without having the docs for your drive. I don't know how the C6 compares to the C11 (you are using the speed control built in to the C11 aren't you?).

/Henrik.

[tenmetalman]

H.O.
My "settings" are like everything else I do anymore, close, but not finalized.
I'm going to have to work with the module settings again & make notes to write from !. I set the big E settings to make large moves without the drive faulting
let me go through the process's again & work up some real settings. the drives do move extremely smooth, but I need to take the time go get the actual distances on the button. I was hoping to have the 2.7Ghz computer back on line, set the kernel speed to at least 60000Hz where I hope to run mach3. go through the process one time & be done with it. HA, the motherboard I ordered arrived DOA........... can you believe it.
Trouble Magnet, that's me & I compound it by trying to take short cuts....... more later

[tenmetalman]

H.O.,
Yes I am using the CNC4PC C-11 BOB revision 9.3. & yes I am using or trying to, the analog speed controller thats part of the C-11

[tenmetalman]

H.O.,
I only worked on axes with the modules on board all three drives. I ran into an anomie that made me wounder if' I've made a assumption that's incorrect. I've been making adjustments to the various settings P,I,D, etc and making moves in hyper terminal to test the results. While working on the Z axis almost nothing I tried ( hyper terminal ) would not result in an error or reduce the perceived abs error under 4000 with each move. before quiting I tried making a rapid move in mach3 & it worked without faulting. making adjustments & checking the results with moves in mach3 got the Z axis successfully going, not finished, but going. should I be testing in mach3 long before I have been ?
Paul

[H.O]

Hi Paul,
I must admit I don't fully understand what you mean so I'll simply write a few notes, hopfully something will come out of it.

Using hyperterminal to command the module to move the motor "bumps" the setpoint or target the amount you specify, ie if you command +1000 the motor setpoint will be increased 1000 count "in one go". This means that if you now check the Peak Error it will show 1000 because that's how big the error became when you "instantly" moved the setpoint 1000 counts "ahead". This value will then only be updated if an error larger than 1000 is detected or after you issue an reset which will reset the Peak Error to 0.

In short, the Peak Error value shown by the module is a pretty useless indication of performance when doing step response moves, ie when the moves are commanded by Hyperterminal because the moves aren't "trajectory planned" so there is no acceleration or deceleration.

Does that answer your question? If not let me know and I'll try again.

/Henrik.

[tenmetalman]

Henrik, my energy for the last few days has & is being directed at getting the new mill computer up & running. the 2Gd of new 1000Ghz memory would NOT work with the new motherboard even though it should have. I substituted 2Gd of used memory & was able to get to the post and eventually wipe the drive ( a 2.5" 80 g Toshiba drive) load XP Pro on it & add a 2 parallel port express 1 card. I loaded all the required software, drivers, settings here in the house before ever taking the PC out to the shop. I had to change one setting in Hyper terminal to get key repeat in the terminal window & it was up & communicating with the drive via serial port. the table axes jump when power is applied & mach3 started or stopped. I can see the step pin LEDs turn on & off with my key strokes but the X, Y & Z axes are not moving. so I figure my next move is to check the output of the parallel port pins. see if the signal is there at the right times & what the actual voltage output through the cable. never dull minute. I'll keep you posted. if all else fails I put a 3 1\2" drive......
what I was trying to say earlier was that when setting my Z axis the responses were quite different from I was in hyper terminal & mach3 three. to gain control of the axis I had to make change minimize hypertermimal and check a move in axis Z. If I tried a move in hyper terminal the axis would attempt to run away. if all else fails I put a 3 1\2" drive

[tenmetalman]

Henrik,
I've been slugging away & made some additional progress. when setting up my "new PC" I had installed a PCIe1 parallel card to give me the option of a second parallel port. I removed it. the motherboard has a parallel pin set & supports one parallel when a port is physically attached via a cable. I know thats not very clear, what I'm trying to say is that the actual female parallel port is not built in nor included.
anyway after removing the second parallel port card the computer, mach3 & the drives once again started communicating. I took the time to set the multiplier on the modules to 2 & adjusted the counts on all three axes to get the actual movement within +- .01". I t dawned on me after I came in that I need to slow down the acceleration/deceleration down on all of the axes & that will help with the error count & overshooting thats occurring, especially with the Z axis. my weekend has some commitments thats going to limit my time in the shop. after the 12th other things are going to be demanding all of my time for probably a month so I'll get done what I can before then. I hope to implement Henriks settings & give the spindle another try before I have to quit. here's a few photos per his request

[H.O]

Hi Paul,
I'm glad you got the LPT-port working and outputing pulses to the drives, don't worry I understnad perfectly what you mean with the port being available on the board but not on the backplate - my "office machine" is the same way. The reason for the PCI-e card LPT-port not working might have been that you didn't enter its correct port adress in Mach3? It's usually 0x378 for an onboard port but for add-on ports it's usually different and must be changed in Mach3. Just a thought.

Axis runaway are serious business and not a good thing. However, as I've tried to stress in the manual IF you disable the following error limt (which you did be setting the E-parameter to 50000) and then commanding a large move the internal calculations can overflow and cause unpredictable results. When doing step-response moves from Hyperterminal you want to make small moves, 10, 50, 100 counts or so.

If the drives are current limiting as you accelerate then reducing the acceleration will help reduce the following error, it will also reduce the overshoot. However there should be no need to (as long as the drive is NOT current limiting) trade acceleration for less overshoot. Adding more D should reduce any overshoot, so should lowering I but you're already very low on that.

Now, I don't know much about your machine but somewhere around 500ms acceleration time is probably a good place to start. I'm guessing you should be able to accelrate twice as hard (250ms) once you have the tuning done.

Finally, did you change back the antidither settings to their defaults? (J=2, K=256)

Don't rush on my behalf, you know how I feel, take your time and do it properly ;-)

/Henrik.

[tenmetalman]

Henrik, Just a quick reply before showering & starting my day. Thanks for the input. as far as the I & J yes, I've set them back to your recommended values. I was very careful to install the parallel card per it's instructions, load & apply it's software and check all of the port addresses in control panel afterward. in the past on other PC's I had LPT#1 settings show up different after adding one or more additional LPT ports. later on I'll give it another try. for now I'm trying to keep my eye on the brass ring. I'll work on the overshoot problem & post the current settings & checks so you've got more information to work with....
Later
Paul

[tenmetalman]

Henrik,
BTW, axis runaway is defiantly a serious problem along with when the Z drive faults & the Z axis drops. I'm thinking I'll add a relay that applies the Z axis brake when the drive faults. I've got a indicator thats now "art" due to being altered... I found out my reflexes aren't what they one were
;^(

[LZ1TWB]

Hello all, what happened with the upgrade module? Any update on the work last month?

[H.O]

Hi Todor,

Well, I'm afraidf I don't have much news for you.

To be honest I'm a bit frustrated as I've sent out 7 modules (one of which vanished in the post), free of charge, including shipping at my expense This far, one person and one person only has given me any sort of feedback. (Thanks Paul!)

I'm not whining, we all have busy schedules and things to do but I was, quite honestly, hoping for a little more in return. (The first module was sent out almost 3 months...)

I've also just recently sent out two modules for which I got paid. In that case I certanly don't expect the user to spend time providing ME with feedback. Yet I feel confident that is just what he will do, and I thank him advance for that.

With that said I've ordered 100 PCB's and parts to complete 25 modules, should arrive in two weeks or so. I've also added a feature or two to the firmware and worked a bit on the docs. I feel pretty confident about it.

Again, I'm not whining, just letting you know how it is.

Thanks!
/Henrik.

[LZ1TWB]

OK, no problem, just wanted to be in touch with the thread. I guess the summer everyone is loaded with work, so there is little time for cnc if it's not the main moneymaker. But they should fulfill what have promised I think.

Anyway, a little OT about my luck here, last week I've had my first dead parts inside the control box since the build-up, 10 months ago. The soft-start power resistor has disconnected itself inside and without knowing it we used the machine till the 40A diodes shorted eventually charging the 40000 uF bank. Thanks god I put fast fuses on the primary, which saved other things on the row.

I was curious and broke the 20W resistor to see what has gone wrong. It seemed that around the blown part of the wire the ceramic fill-up has not been put well and there was an air bubble.

Todor