The Sequel - Oh Blimey I bought a CNC Lathe (Beaver TC 20)

[awemawson]

It's been a bits an bobs day for the last couple of days, chasing up and investigating things. A few positives:

I've managed to find the logical word in the controller that represents the Touch Probes making contact, and for the simpler HPA arm mounted one proved that it works.

For the more complex Optical Probe I have a feeling that it isn't working - the actual contact switch on the probe is working, the Infra Red wake up signal is being issued by the cabinet mounted transmitter / receiver, but the electronics on the probe head isn't waking up and replying. I made a simple IR receiver up using a QSE158/9 IR ic that contains the detector and an amplifier / driver that turns on a visible  red LED when IR is found, and proved that the 'OMM' unit that is cabinet based is transmitting OK but being ignored. So far I've not managed to open the unit up for further investigation.

Then the Postman brought a mystery package - contents unknown  Turns out that a very kind MadModder is my benefactor, and sent me a compatible replacement catch for the one that was damaged. Same hole in the door but slightly more streamlined outer profile. Thank you very much Smiffy. However no further door progress as my 'next day delivery' from RS Components still hasn't arrived two days later.

I've been doing quite a bit of research how this machine is supposed to position it's main spindle rotation-ally for milling using the power tooling. There is an M code 'M19 S<angle>' that should stop the spindle at the specified degrees, but although it stops the spindle, the place is random  I've managed to find the bit of the controller that monitors the angular spindle position, and sure enough it displays correctly if you manually turn the chuck, or set it slowly turning - but the stopping when there bit doesn't seem to be functioning properly.

The another delivery of Tooling arrived from APT. I thought that I'd ordered a left and a right handed version of tooling for TNMG1604 (triangular) inserts, for SNMG1204 (square) inserts, and for DCMT11T3 (diamond) inserts, but it seems that I cocked the order up and the right handed TNMG one was infact for a 20 mm square shank  Never mind it'll go on the Colchester Master and I've re-ordered the correct item

Oh and another positive - I've braved the 'Gear Change Mechanism' and despite previous reports that it might have a fault, it seems to work OK 

[awemawson]

I've been having trouble with the tail stock. It is clamped to the bed by spring loaded clamps biased in the 'clamped' direction, and un-clamped by two small hydraulic cylinders. M51 comand clamps it, and M52 unclamps it. In addition there is a pressure switch on the hydraulics that reports the clamped state back to the controller.

The actual tail stock barrel is advanced hydraulically by an M11 command, and retracted with an M12. Despite me issuing an M51 command, the controller got all upset reporting that the tail stock was unclamped, and faulted out if I issued an M11. Fairly obviously something to do with the pressure switch, and eventually traced to intermittent contacts. Almost certainly due to the machine sitting unused for many years

Now it's a sealed unit, so no physical cleaning or squirting contact cleaner was possible. Remembering back to my days with relay logic, most are arranged that the contacts rub a little as they open and close to remove any oxide that forms, and the 'wetting' voltage and current are important to keep the contacts low resistance.

OK maybe I can increase the wetting current - I daren't increase the voltage much as I don't know the contacts ratings. A quick test with the AVO showed me the controller input circuit was passing a mere 15 mA when the contacts decided to close.

Now I had already written a little diddy program that clamped and unclamped the tail stock every second in an attempt to clean the pressure switch contacts, things had improved but about one in 35 closures was unsuccessful.

Time to roll out the big guns - I wired the contacts to my trusty current limited lab power supply, set the volts to 30 and the current limiting to 250 mA and ran the program. Initially still problems but after about 20 seconds - hey we have reliable contacts - this rather off the wall method actually works

Lab supply put to bed, contacts returned to the controller to play with, and let the little program run again, which it did faultlessly 

[Will_D]

I am so impressed with this thread Andrew. Yout rabnge of skills/knowledge is oustanding.

Guys: When Holywood casts the movie  who should play Andrew.

PS: How are the pigs?

Watering and cutting Rugby pitchers is keeping me out of the workshop and is so boring

Joke:

Was talking to a native American friend yesterday.

What is your wifes name is asked.

4 Horses is her name.

What a lovely name for the wife, What does it mean?

Nag, Nag, Nag, Nag!!

[awemawson]

Will-D those two Berkshires are off on a one way journey early Monday so don't tell them !

At last today the heavy duty piano hinge arrived for the Tail stock Door, so a suitable length was cut off, I picked up the four tapped M6 holes in he door frame and drilled matching 6 mm ones in one leaf of the hinge. When I'd checked that they aligned OK, I opened the hinge holes to 8.5 mm to get latitude for adjustment when the door is swung.

After a test fitting of hinge and door balanced together in the frame, I then spot welded the un-drilled leaf of the hinge to the door, gave it all a good clean up with IPA, and sprayed the inside RAL9001 to match the machine.

The weather is distinctly cooler and it's raining today, so although the paint was 'touch dry' after an hour I left it a couple of hours more before inverting it to spray the outside.

The Postman brought the replacement lathe tool that I had wrongly ordered as a 20 mm shank rather than the 25 mm that I'd intended - so when I'm brave I can start tooling up.

Today I've spent quite a bit of time investigating how to position the main spindle to a known angle. There seem to be two 'M' codes involved. 'M20' that enables the spindle drive, and 'M19 S<angle>' that orients the spindle to the required angle. I have example programs showing me how it works but it doesn't 

I even called up the chap who used to use this lathe to confirm that the lathe has the capability, which he confirmed, there must be some parameter or setting needed to enable it I reckon 

[modeng200023]

If you go on improving the machine like this your friend will want it back!

[awemawson]

Well I'm STILL going round the houses trying to get 'spindle orientation' working. 

Today I've identified EVERY connection to the KTK Mentor spindle drive, and proved to my satisfaction that it is working properly when commanded. I have also been through every input line and output line that has anything remotely to do with the spindle or it's positioning, and proved it works. At a logical level it's pretty simple. The Mentor drive takes in an analogue value that represents speed (-10,0,+10) and has an input from a tacho generator that represents actual speed. There are various enable and inhibit signals all of which work. If I drive the spindle at 1000 rpm clockwise I get an analogue input of +3 volts and a tacho gen output of -60 (rounded figures). If I drive the spindle anti-clockwise at 1000rpm I get an analogue value of -3 volts and a tacho gen output of + 60 volts. If I manually turn the spindle I can display the output of the shaft encoder in degrees and they are sensible.

Conclusion: the Mentor drive system is working, as is the encoder feedback of position to the controller. The problem MUST lie in the controller itself or it's parameters.

I've then been through every parameter that I can find that has anything remotely to do with M19 spindle positioning, and again everything looks sensible. I've re-loaded the controller three times from the back ups that I sourced - even set the baud rate for loading down a couple of notches in case something was being miss-read.

. . . argh !

So to restore my sanity I've swung the Tail stock Door now it's paint is a bit harder. Paint is a bit lighter than the original, but will have to do.

Earth shifting tomorrow if the weather is kind, so maybe my spinning head will clear !

[russ57]

So, the rubber duck approach : explain to us ducks exactly what you are doing, what you expect to happen, and what does happen.
Sometimes the act of explaining it helps you to realise the issue.
Or, less likely, one of us may see the flaw.



Russ

[awemawson]

It also helps to vent my frustration Russ !

[hermetic]

I agree fully with your method Andrew, go away and do something else that keeps your mind busy, and let the unconcious computer have a go at the problem. You can only get so far in one session, and then you start to get frustrated and think yersen round in circles! Take a break. Earth shifting sounds fine, but always leaves me with a flasback just as I am falling asleep, I am back on the digger, and it has just gone beyond the point of no return, and is going over!
Phil

[cnc-it]

Mostly sorted Andrew..you are becoming a leading expert on the Beaver machines  I'm not far behind you having re furbished and repaired a few in my time.

[cnc-it]

On the spindle orientation..a very odd fault..as you would think the control would alarm if it wasn't in the correct position?

[awemawson]

Welcome to the forum, why don't you post a bit in the Introductions section and tell us a bit about your projects ?

[awemawson]

On the spindle orientation..a very odd fault..as you would think the control would alarm if it wasn't in the correct position?

It just 'hangs' with the running light lit, so I expect that it's waiting for an input from some point on the ladder logic. I just wish I could source a copy of 'Step5' software that talks to the controller and lets you examine the PLC ladder in real time. All the hooky copies I've seen so far are missing the authorisation code 

[awemawson]

I decided that at least I could easily solve the issue with the FXM-3 Field Coil DC Drive Card overheating by adding a fan - actually easier than making a larger, but remotely mounted heat sink, which was option 2. The heat sink is up at 415 volts so needs to be away from fingers!

240 volt 80 mm axial ball bearing fan ordered from RSComponents along with some M4 nylon stand off pillars arrived this morning while I was pushing earth about with the JCB803. This evening I drew up a simple mounting plate in AutoCAD which via SheetCAM was ported to the CNC Plasma table and cut out of 2 mm steel plate, edges bent up to form a flange for rigidity, and a test assembly performed.

All looks good, so it got a coat of Satin Black from a rattle can, and I'll probably fit it tomorrow if time permits.

I praise the day I rescued that CNC Plasma Table - oh boy does it make this sort of thing easy 

[cnc-it]

Thanks Andrew, yes I will do an introduction had forgotten about that.. 

On the spindle orientation..a very odd fault..as you would think the control would alarm if it wasn't in the correct position?

It just 'hangs' with the running light lit, so I expect that it's waiting for an input from some point on the ladder logic. I just wish I could source a copy of 'Step5' software that talks to the controller and lets you examine the PLC ladder in real time. All the hooky copies I've seen so far are missing the authorisation code 

I have had this happen on a few different machines...it always turned out to be the post processor..either me using the wrong one for the machine or when I was testing a new modified post.  If it's a PLC error or a mechanical error I always get alarms..but I have never used Siemens..only Hurco, Fanuc and Heidenhain.

John

[awemawson]

John, I'm putting in code copied directly from the Beaver hardware manual, also from a set of Beaver programming course notes that I have. I'm doing it under MDI, and also running it from a small program of the same code loaded into the machine, all with the same result.

On the Generation 2 TC20 Lathes there is a 'Setting bit' that turns this feature on. I have the details for it. I can identify no such bit on this Generation 1 machine, but I'm sure that there must be some mechanism to enable or disable it, as the necessary spindle encoder was a optional extra.

I'm sure the key is going to be getting hold of a working copy of the Step5 software, though no doubt using and interpreting it will be a long and involved process !

[cnc-it]

Yes I agree..with a name like Step 5 it sounds like only the tech who wrote the software knows how it works  but well worth a try!

[awemawson]

So having picked up a load of grass seed from the merchant early this morning, I was able to get on and fit the fan.

First I completed the assembly, incorporating a fused terminal block, then I re-set the field current back up to 3 amps. This involves intercepting the FXM-3's output and inserting an Avo. Pot duly tweaked. Then I replaced the original card mounting screws with the nylon pillars and wired it up.

All very straightforward. I can't now point my I/R thermometer at the Thyristor heat sink, but from the flow of air over it I'm sure that it's fine 

Last night in a flash of inspiration, I remembered that in the machine door pocket had been a floppy disk with three files on it - copies of the machine parameters and PLC program. Just suppose the PLC program was different . . . . could that have the key to the axial positioning issue? So I ploughed through page after page of data and sure enough I found a difference in one line . . . . . whoopee !

So having fitted the fan, once more I re-loaded the controller data just hoping . . . . no, no such luck . . .no difference . . .and no idea what the different data represents. Another oddity to tuck away and remember in the future.

[cnc-it]

It's funny I was going to mention if there could be a difference in the PLC software or the parameter set...my Partsmaster parameters  from new where changed at the college to add/remove options over the years but no records were saved..

[awemawson]

Yes I agree..with a name like Step 5 it sounds like only the tech who wrote the software knows how it works  but well worth a try!

STEP5 is actually a PLC programming language developed by Siemens but has been superseded by STEP7. I did persuade Siemens to let me download a trial version of STEP7 (1.8 GigaBytes!) which took 7.5 hours to unpack then produced nothing useful Anyway apparently STEP7 won't talk to this generation of controller !

(STEP5 runs under DOS or the earlier versions of Windows before NT)

I do actually have STEP5 but no authorisation code for it, and Siemens can't supply as it's obsolete !

[David Jupp]

http://forums.mrplc.com/index.php?/topic/24709-alternatives-to-step-5/

[awemawson]

Thanks for that David, but the available one of those is £600 

I suppose I'm going to have to apply the 'infinite number of monkeys' approach, and just keep slogging on. Now I'm beginning to realise how they felt at Bletchley Park solving their enigma !

[cnc-it]

So Andrew is the cooling fan on the FXM-3 a part fix or can you run the motor at full current now..? I've never seen one of Mikes boards but looks very similar to the standard issue FXM-3..are there any advantages using Mikes newer board?

[awemawson]

Mike told me it was designed for a maximum of 3.5 amps, so it is on the limit with the size of motor in this beast. But he also told me that of the boards he made, he only had transformer failures, never the thyristor. So having re-set it at 3 amps it's still on the conservative side of the spec, but I far prefer electronic to run cooler than hotter!

The redesigned board is simplified, has only one transformer, and uses a relay to indicate 'OK' rather than semiconductors. I've had the comment from someone else who repaired Beavers that the new cards were more reliable than the original.

[cnc-it]

Thanks Andrew that's good to know.. I'm not sure what amps the original board was rated at?