lol that sounds a bit on the slow side. Does it match the speed stamped on the motor case?

Thought I'd just share this. When I went to C.E.F yesterday saw this on the wall:



Clearly someone there has a sense of humour!

Today:

# Pumped/Wiped out all the remaining water in the base of the machine. Didnt want to leave it there to go rusty/mouldy.

# Put in the bolt that hold's the spindle motor shaft extension in place. (left it at home yesterday by accident)

# Wired up the motor to the VFD

#Wired up the VFD power to the mains supply rail

# Removed the now redundant DB15 rutex breakout board.

# Made up a new DB15 connector providing a direct 24v connection to the rutex motherboard.

# Tweaked the firmware on the rutex drives, raised the current limit and enabled shut down on error detection.

# Added 10nf caps to the limit and e-stop inputs to reduce noise problems


After wiring up the motor I powered it up (by the buttons on the inverter) and it roared into life. Everything seemed to work ok until I walked over to the pc and glanced at the monitor to see the display dancing all over the place as if someone was sliding a very large magnet around the edge.

Turned off the inverter and it went back to normal. 'Oh dear' I thought .. major noise problems ahead. So I started unplugging things to see if I could trace where the noise was getting into the monitor. Eventually I got to the monitor extension cable and unplugged that and then looked at the OSD which was totally stable. So I figured the noise was coming through the monitor signal cable rather than the mains OR the osd wasnt suseptible to the noise cause it was being generated internally. So semi confident i'd traced the proplem i plugged the monitor data lead back into the extension expecting the noise to return but again the display was stable.

Turned the inverter on and off a few more times and varied the speed but couldnt replicate the extreme effects that had appeared before. I thus reached the conclusion that while the inverter was indeed emmiting a degree of noise, the extreme effects were the result of a loose conection between the monitor and extension (data not power) rather than something more serious.

So I powered up the machien again and mach 3 and the inverter. With the capacitors on my limits I wasnt getting any spurious noise triggering things it shouldnt but then when jogging the axes my y axis suddenly locked up. Rebooting the servo drives fixed the problem but it quickly happened again. Investigating further i realised the y axis drive was only tripping while the spindle was running and would be fine the rest of the time.

My guess is that noise on the encoder lines is being interpreted by the drives as pulses which is then tripping the newly enabled feature which is designed to throw up a fault if the number of actual encoder pulses differ from the number of commanded pulses. While this does mean the machine is currently unuseable i'm not too concerned as I have not yet properly terminated the shielded encoder cables. It its all a bit of a mess atm as I was in a rush to get everything wired in so i could see things moving but I have now bought some expensive shielded RJ45 cables allowing me to cut down the encoder cables as short as possible and then plugging them straight into the drives without going via any break out boards.

Once this problem is solved and I can operate the machine with the spindle motor running and nothing tripping that shouldnt then I'll attempt to wire up the VFD to parrallel port outputs so I can have mach 3 control over spindle speed and direction and hopefully i'll manage to acheive this without inducing any more noise in places that noise shouldnt be.

Something that is starting to annoy me a lot now is the accuracy of my table. I havent got the numbers for my mach 3 steps per unit quite right so all the DRO's are out by quite a long way. In english a commanded 1mm on the computer is something much bigger in actual machine movement. I think one of the next very high priority tasks is to borrow a dial gauge from someone and calibrate the axes properly so the machine actually becomes accurate.

At least then I can try importing some gcode and drawing on paper with a pen in the collet and then move onto bigger stuff once the interference probs get ironed out.


**Note to self before ink rubs off back of hand***

Next time bring rivet gun and hoover.

Today:

# Riveted conduit back onto the wall of the rear cabinet (Had previously drilled out the rivets and re-drilled the holes lower down to create a bigger air gap between the inverter and other cables to help reduce potential noise problems.

# Hoovered out dust and rubbish from both cabinets

# Tried to replace the rj45 breakout boards with direct connections to shielded rj45 connectors but discovered the shielded usdigital cable is not the same thickness as shielded cat5/6 and physically wouldnt fit into the rj45 connector.

# Properly terminated the shielded encoder cables, earthing them at one end and sealing the ends with heatshrink. This more or less solved my y-axis tripping problem, also upping the mach 3 pulse length from 2 microseconds to 3 microseconds seems to have made a difference.

# Happy that the VFD was not tripping anything as a result of noise in its current state (power & motor connections only) I proceeded to wire up the control side; forward, reverse and 0-10v analouge for speed control. I wired this all through some spare shielded encoder cable, earthing it at the VFD end.

# Figured I'd test it out with just switching first and no variable speed. Typed in M3 to the MDI and the motor came on and ramped up to the default max speed of 50hz. Not wanting to run the motor at full speed for long during testing I typed M5 to stop it. Now I could smell burning and fearing the worst I dived for every big red switch I could find. Happy that all power had been removed from anything that could burn I sniffed about and found that the pulley on the spindle motor had come loose causing the belt to rub against the frame of the machine.

At this point I had to pack up and go to work (i.e the other job which isnt much fun but actually pays where as this one is much fun but so far is still just costing me money.) So I still dont know if the speed control bit works, or reverse, or if these new cables have created additional underseble interference.

It looks like tomorrow I will have to remove the motor from the machine and take it back to soapers to see if they can make it more robust.

Took the motor out today and returned it back to alan to make some modifications to the pulley. I think now the plan is to cut a keyway into the pulley and make use of the keyway that is already in the motor shaft.

Before removing the spindle motor I had it up and running via mach 3 including speed control this time. Everything seemed to work ok and it seemed the additional wires hadnt created any extra noticable interference.

Spent the best part of today stripping down/cleaning the suds pump and motor. After taking appart and putting back together it is still loud but nothing is jammed or rusted so I think i'll just run with it as is. As the old saying goes if it aint broke dont fix it. Its 23 years old, theres going to be some slop in the bearings, but right now I cant justify the financial outlay to do anything about it, especially as it still does what its meant to.

I did however replace an o-ring seal which had become stretched so hopefully that'll fix the problem I saw of water being sprayed out from the intake end. Certainly when I tested it in a bucket of water today everything looked good.

Alan didnt hang about with my spindle motor. Went down there first thing this morning and him and his grandson were already hard at work; by the end of the day it was all done and ready for me to collect and they are certain this time it wont come loose.

Last mission of the day was a trip over to MS engineering. Martin very kindly leant me one of his old dial indicators with a magnetic base that I should be able to use to calibrate my spindle. The accuracy isnt amazing but he suggested I use this to gain an initial level of accuracy and then recalibrate with an even more sensitive dial indicator to get it perfect.


Looks like tomorrow will be quite a busy day. Got to get the spindle motor back in, wired up with belt tensioned, got to get the suds pump back in and wired up and then have a go at calibrating each axis. In theory if this all goes well I could actually mill out my first part tomorrow or at least have a go at cutting something using the mdi.

Collected the spindle motor from alan this morning and went up to the workshop. Put the newly modified motor back in, wired it back up, tensioned the belt etc. This time checked that M3 was clockwise and M4 was CCW. Turned out it wasnt so I had to swap two of the phases.

Put the freshly refurbished suds pump back in, seemed a lot harder to put back in than it was to take out so I ended up removing a few side panels and the pneumatic to hydraulic converter in order to gain better access to the hole in the side of the casting.

Discovered I'm getting quite a lot of 'driver watchdog errors'. I'm not entirely sure of the cause but I think its interference related, possibly to do with the VFD and I'm thinking mayby the noise is being induced upon the mains and getting into the pc that way though this is just a gut feeling rather than based upon sound evidence.

Had a play about with the dial indicator for axis calibration but I needed to fix something to the table to measure off which I didnt have, nor did I have any T-nuts to fix something down with.

Just plonk the base down on the table and put the gauge on
something not on the axis[head, column, knee].
You should be able to calibrate using a ruler or tape measure.

Shouldn't any interference be taken care of by the pc's powersupply?

You cant really get more than 0.5mm accuracy with a steel rule though. Tape measure would be even worse.

The guage I have is good for 0.01 but when I've got as close as I can with that I'll use another which can measure to 0.001

Electrical noise is a strange beast as you are entering the realm of radio; and everyone knows that true understanding of radio is not a science, but a black art.

Thats the same resolution, if you go 10mm at .01 or 500mm at .5mm.
I think you could resolve .25mm though and you have760mm x axis travel don't you. my tenth gauges have even less travel, I don't have a .001mm gauge.


I have a centrescope, I could use that with a 10" inch standard. The centrescope is good for a tenth of a thou. You could use a magnifier and a steel rule, I checked a steel rule andit was pretty acurate, though the markings are pretty wide when viewed thru a centrescope .

I would have thought once you have the figure for pulses that's the figure you use and any deviation is screw error?

What about the filtering, the matsuura had several big metal boxes labeled as filters, I think the power to the inverter went thru one. The inverter was also living in a seperate enclosure. The inverters I looked at on ebay all said they had filters built in.

Filtering is an option I'll look at if I have to though i think it is quite complicated and expensive. I'll try and solve the problem with earthing and shielding first.

Might also try using some of those ferrite beads to run the noisy cables through as they are supposed to cut down noise.

Whats a centrescope? Can't say it is something I have encountered before.

Going to have to buy myself a pair of large machine vices soon. Any reccomendations ? A mate of mine showed me some really nice hardened ground ones but theyre about £400 each so I think to get started with something a little cheaper would be preferable.

Didn't make amazing progress today - took far too long making some square nuts by hand so i could bolt down the vice to the table to give me a solid non moving object to calibrate off.

Calibrated the X axis quite easily, the Y didnt seem so good though I cant help thinking the sticking dial indicator isnt doing me any favours. The calibrated mach 3 steps per unit were only very slighly out compared to my calculations. I had caluclated 800 for x and y and the measured was about 803.

Will have another bash tomorrow at calibrating the Y and then do Z as well, then I might have a go at milling a part from modelboard or LDPE. Will see how it goes. All getting quite exciting now. I've had enough of messing about with rebuilding the mill - I just want to machine stuff now!!

It's Alive!!!



Very close to cracking open a bottle of champagne as victory is indeed near.

Today I calibrated the axes and got them all to within 0.01 which is all my gauge could measure. Then went on to mess about with and learn the basics of Lazy Cam (2D CAM package). Imported an old dxf I had milled on my previous cnc into lazycam and after a bit of trial and error managed to get it to produce a toolpath which resembled the drawing.

Imported this to mach 3 hit go and minus a few teething problems it was doing as it was told and making chips!

First go it started off ok but then the X axis drive tripped out messing up the rest of the cut, This happened a few more times so I tried halving the max speed in mach 3 and this made things better. Now it would run a whole program without tripping a drive but there is still one more problem to solve.

When a program starts and the table is moving to position at the same time as the spindle starting sometimes the table spazzes out (yes that is a technical term) and judders violently for a few seconds before calming down and finishing the rest of the program without complaint.

I'm not sure what this is but my guess is one of two things - interference from the VFD or voltage drop caused by insuffucient supply current.

It could be that whilst ramping up the motor the VFD is even more noisy than when running at a constant speed and thus this is causing the problem at the start of a job.

However I'm more inclined to belive that at the start of the program the combination of starting a 4hp spindle + rapids of 2-3 axes + two computers is too much for my 100metres of 1.5mm cable on a 13amp plug. As a result the axes start ok for the first second or two as they drain the charge in the PSU capacitors but then the mains cant refil them quick enough so the servos judder as they try to achieve the commanded position. Once the spindle reaches full speed the current it is drawing drops and about the same time the cut sequence also drives the servos slower this drops the total demand for power to something more sensible and thus the machine carrys on ok after this.

This is all theory and I could be totally wrong. I guess I could test this by writing some g-code to start stuff in order to minimise current draw. Really I could do with hooking the thing up to a big generator and seeing if the problem still exists. I am of course secretly hoping it is a power problem cause although expensive 100 metres of 6mm cable isnt too difficult to install. Curing a VFD related noise issue could be a lot more complex.


Annnyway. Fact of the matter is - it is actually working!!! Despite the teething problems mentioned above, I still can draw something in CAD and my cnc mill will cut out something that very closely resembles what i've drawn!

Had another long play with the machine and made more progress as well as cutting more bits.

I've started to dislike lazycam as a cam package as it seems to very limited (Even for 2d) and i've recently read that its likley to stay in a permenant beta state cause the developers cant be bothered with it as there are already so many better simple 2d cam packages out there. With that in mind I guess it's time to start looking for one.

I had a brainwave that the jerky axes could have something to do with pulse frequency, maybe the pc being able to send out enough pulses or whatever so I tried reducing the mach 3 pulse frequency to 75khz and retuned my motors to run at max speed at the new pulse frequency. This seemed to make things a lot better and solved the jerking of x/y axis at the start of a cut cycle.

Of course lowering the pulse frequency has affected the voltage output of my step/dir to analouge pcb which now even after adjusting the trim pot only puts out about 8 or 9v. I will contact arturo to see if theres a work around for this, if not I might have to look at 'tricking it' by raising the max VFD output to 60hz to 9v will cause the vfd to output 50hz. Bit crude but and easy fix to get full speed out of the spindle.

I'm pretty sure the VFD is the cause of a lot of the little unexplained problems i've been having. Everything just seems to become less stable the faster the spindle is turning , Somtimes after telling the computer to turn on the spindle at full speed it will actually ramp up, then ramp down and ocsillate like that for a few times before either running ok or causing mach 3 to trip with a 'driver watchdog error' so it almost seems like RF interference from the VFD is actually effecting the pc's own internal circuits which is a little bit disturbing.

Another little bug i've discovered is that since calibrating the axes, mach 3 doesnt do as it's told. If I type in G0 y50 for example the dro is showing +49.999. This must be a mach rounding error as the dro has no encoder feedback and should display exactly what is typed.

I've also got a bit stuck with some simple VB scripting. I want to turn on my oil pump when the spindle is running but adding the code to the M3/M4 macros stopped them from working because those macros contain functions which are called from other parts of the program which get confused if they see other code there as well.

I have the same machine as Dom and have also retrofitted it with Gecko G100, stepper drives and drivers and Mach3. The machine works well! However, no spindle. I know my controller works. It was working before I did the retrofit. I DO have an analogue output from the analogue port on the Gecko G100 and this is going into pins 4 and 12. I have bridged pins 19 and 32 (Spindle stop) and pins 20 and 9 (spindle drive off). Some LEDs light on the controller. Power on the FRU card and ready on the CPU card. But no rotation of the spindle. The fan works.

I would appreciate any help on this.

Cheers.

Mike.

Mike, You're probably better off asking this on cnczone, we dont really have a huge number of active members here. For the most part it's just me rambling on to myself ... but I dont care

I've been researching CAM programs as Lazy CAM which comes with Mach3 really doesnt cut the mustard. Something that became immediately apparent was how expensive it's all going to be. I have a few options in mind now though and after doing a day of research on the subject I feel much better educated on the topic even if i'm not actually much closer to buying anything.

I did also find a few free/beta packages which I might try out for the time being. Right now I dont need anything too comlex. Just to get my encoder covers milled up so I can machine metals and run coolant without worrying about destroying the encoders would be a great start.

Also got an email back from Tom regarding rutex fault codes which are as follows:


The LED blinking error codes for the Rutex 20x0 drives is:

No VM supply: 01010101
Lim/Switch Input: 01110100
Im (overcurrent): 01110101
Over Voltage: 00000001
Following Error: 00000101
Encoder Error: 00010101

All OK: 11111111 (full on)


Hopefully this info will allow me to trace the occasional trips I'm getting.

Congratulations..



It's a microscope on a morse taper[comes in other fittings]. some have cross hairs mine has two parallel lines which are close enough together that the instructions say you should easily be able to locate to a tenth of a thou.



I have a selection of pretty horrible vices that came out of the skip when a local firm moved. I fancied a pair of matching vices but they are just way to expensive.

So how is your retrofit going graffian? You anywhere near as close as me to finishing ? What gear are you using ? You should create a new topic on the forum and post some photos.

No where near as close, I don't seem to make any headway.
Sent my motherboard back to MSI and the one they replaced it with was
ATX rather than miniATX[it was the only one they had with parallel port], so I had to alter my mother board tray and mounts, then found all my nice ribbon cables for panel switches didn't plug onto the new mother board.. Not touched it since march, will get back to it in a few weeks.

I know that feeling well. I've certainly had more than my fair share of problems along the way with this project but I finally feel like i'm getting close to light at the end of the tunnel.

Only problem I'm having now is actually using the machine. My experience of using cnc is really all from school where we were sheltered from the harsh realitys of cad/cam and post processing through a highly oversimplified combined cad/cam/nc package which basically let you draw stuff, colour it in, assign depths to the colours and then you could just go file/plot/ok and it would start cutting.

Trying to get my head round real world CAM pacakges is a nightmare, especially cause right now i'm not in the mood for learning CAM and what I really want to do is just see my machine cut stuff!