A new idea for keeping PCB material flat while milling artworks
The vacuum plate mentioned elsewhere on my blog serves me well when milling printed circuit boards on the Tormach PCNC440. It keeps the PCB material flat and makes the cut widths repeatable when using V cutters.
Idle hands and brain during social distancing has produced a possible solution that might be of interest and stimulation to others. It consists of a circular pressure ring that sits around the spindle chuck and tool. There is a second ring that sits on the spindle body connected to the lower ring with rods which have coaxial springs pushing down on the lower ring. The magic is to use mini ball transfer units on the lower ring to press down on the PCB and glide friction free around the PCB as the cutter does its stuff. The assembly is held in place on the spindle with 3 gripping screws. The downward pressure is adjusted by 3 screws that press against the spindle mounting frame.
The prototype was made using 3D printed rings. There is an image below. Apologies for the yellow PLA but finding any PLA at a decent price is very difficult in the present circumstances.
The idea seems to work and has produced some good consistent quality PCB prints. It does have disadvantages in that you need to have a larger PCB blank to allow for the larger footprint of the pressure ring. It is probably only of practical use for PCB milling but then the problem of flatness is less critical in drilling the board and routing the profile.
It arrived today after nearly a month in transit due to the current lock down restrictions. On opening the package I was impressed with the quality of the engineering. It is a nice device. It uses the usual 3 pronged contact mechanism. Supplied with the probe is a tube of grease that helps protect the contact reliability. The interface cable has a 5 pin DIN that plugs into the Tormach expansion socket and the shank is a standard TTS compatible size.
I ran through the initial preparatory procedure and then loaded it into the Tormach 440 spindle. Pathpilot has a number of excellent set up routines to adjust the probe and make measurements. One of these, the Effective Tip Diameter is quite critical. All this went to plan and very quickly. Some initial probing gave repeatable and accurate results so first impressions are good.
I’ll give some updates as the probe gets pressed into service but my first impressions are good with repeatable accurate readings.
In the course of checking out the ITTP probe I needed a reference cross check on the various setup measurements. My Haimer Taster seemed a bit erratic and on inspection I discovered the axial shank holding bolt had worked lose. This meant a re-calibration of the eccentricity of the probe point would be needed.
The alignment process involves adjustment of four grub screws in the shank body. These tweak the ’tilt’ of the shank to get a concentric rotation of the probe ball point. As there are four screws I use two hex Allen keys to make the adjustments to each in line pair. This is quicker than with a single hex key being swapped from side to side. It is a bit like the process I use when centring a 4 jaw chuck. The adjustment is done against a dial gauge riding against the probe ball point. Once you get the knack this process doesn’t usually take too long using the two key method.
The frustration is that the Allen keys provided with the Haimer are a bit chocolate based and the ends chew up easily. The result is you tighten a grub screw and the hex key end twists and gets jammed into the hex socket in the grub screw. While trying to waggle the jammed key you mess up your carefully made adjustment. Aaaargh !
I ground back the worn end of the Allen keys to clean up the hex profile but they quickly degraded. In the end I took the grub screws out completely and replaced them with some M4 cap head bolts. Joyful !
Yes I know it doesn’t look pretty but it is now a real pleasure to make the adjustments with a couple of larger T wrenches. It is probably a criminal thing to do to such a lovely instrument but life is too short.
I & J Arc Code Calculator (with updated spreadsheet)
I had a need to hard code a circular PCB cut out CNC code that would cut four arcs around a milled PCB and leave four breakout tabs to retain the board in place in the blank until the job was finished.
To create I & J codes you need to know the start point, end point and radius of the arc. The end point becomes the X and Y. The delta X and Y location relative to the radius centre point X and Y becomes the I and J values. You can also add a depth of cut value for Z as part of the block. Note that the Arc is assumed to run anticlockwise when using a G3 code running from start point to finish point. Use G2 if you want a clockwise motion. The principle is the same with both rotations.
You end up with a block code of the format G3 (G2) Xa Yb Zc Id Je where a,b,c,d,e are the coordinate values. I found that working with positive and negative values when trying to find the I and J values relative to the centre was hurting my brain. A spreadsheet was needed …..
You can download the sheet as a ZIP file from the link below.
I have my Tormach PCNC440 wired into the workshop network and as a result if a new version of PathPilot is issued my PathPilot controller warns me. This is quite nice as there is no formal emailing warnings of new issues by Tormach. Anyone whose machine is not Internet connected would need to check periodically with the Tormach site to see if an update was needed.
To continue the story … last week I got a warning of a new version of PathPilot (2.4.0) was available and I duly downloaded. One of the immediately obvious changes in the new firmware was a G37 tool measurement routine which works in conjunction with a simple Normally Closed tool setter. From my previous ramblings you will see that I had done a combiner box to allow both probing and toolsetting to share a common input to the Tormach. In theory I was therefore ready to go ….
From my many years in industry I should know that all that glitters etc … the new routine did not work. I thought it must be me but in the end I logged a support call with Tormach and sent them my log file. I also logged the problem on the NYC CNC forum to see if anyone else was having the same issue. I did get one response saying that he was not having an issue. The plot thickened and nothing back from Tormach.
A couple of days later the same responder said there was a fix update available from the Tormach site. It seems the software worked well in G20 Imperial mode but not in G21 Metric mode. He was running Imperial and I was trying to run in Metric Software update downloaded and all is well. It is rather nice. You tell PathPilot where the tool setter is located and then to run the auto tool measurement you put a G37 command in the GCode after a tool change. Away the spindle goes to the tool setter location, dunks the tool and updates the Tool Table. Magic.
Still waiting for Tormach to close off my enquiry and let me know they had fixed the problem and as a result there was a new firmware available. But it is Christmas and maybe they had other pressing matters.
Back from after a few weeks in France and back in the workshop.
Every now and then there is a project that is on the go and you can’t sit down and focus on it. It is a sort of mechanical procrastination. A reluctance to put the first pencil mark on the paper. You then suddenly find all sorts of other things that you kid yourself are more important / higher priority and you get distracted. You know that job will still be there but maybe tomorrow, not today. You suddenly develop a clear conscience about doing something else while you do some background thinking ….
This particular day started off by cleaning down the accumulated swarf (chips) in the 440 tray. Really important job. This led to a check behind the various 440 slideway bellows to see that all was well with the oiling mechanism on the slideways and the ballscrews. X and Y were fine but Z was dry. Not good news.
The 440 is supplied with a manual oiler as standard. This is a reservoir of oil and a pump/plunger which you pull out and release to initiate a slow pressure to the oil distribution pipes.
I checked the plunger and it didn’t feel like it was applying much pressure. This is not the first time I have experienced this problem. If I pumped a few times it felt better so something should have been happening at the oiling points on the Z. I disconnected each of these where I could and sure enough if I pumped hard enough some oil dribbled out but not with much pressure. Something probably not right with the plunger ?
Squeezing round the back of the mill I removed the top of the reservoir (4 retaining screws), disconnected the oil pipe union and lifted the plunger clear. The reservoir can be left in place sitting on the mounting bracket.
There is a large end cap at the union end of the plunger cylinder which I removed and sure enough I could see a mangled O ring. To get the plunger out you have to be a bit brutal. You pull the T handle plunger back out of its housing against its spring using the handle as shown above and then grip the shaft with pliers so you can then twist the handle off. What you don’t do then is suddenly release the pliers grip or the plunger will go into low Earth orbit under the pressure of its spring …
Having disassembled the plunger it was obvious that the O ring had failed quite badly. Tormach support do not offer spares as the oiler is a third party item. They do not know what size the O ring should be. Checking in my box of miscellaneous O rings it looked like a 9mm ID, 3mm thickness part would do the trick. Smearing the O ring with DC4 silicon grease allowed easy re-assembly into the piston bore and then back onto the 440. I now had lots of pressure and oil was apparent trickling down the Z slideways and ball screws. Job done. No pumping needed, just one pull out of the piston handle was generating a slow release of oil to the key areas.
The job I should have been doing was still sat on the bench glaring at me but psychologically I was doing something more important.
Next problem was the Fogbuster air activation valve. Under CNC control this reliably switched on but sometimes would not switch off when commanded to. There are various forum discussions on this problem and many contributors just replace the solenoid valve with a different version. Forum chat also recommended that electrical transient snubbers are fitted across various inductive loads in the Tormach control unit. I had some of these in my stock box (Tormach offer a kit for this). They are simply a series resistor and capacitor in an epoxy block. They are fitted across any inductive device to suppress switching transients. I dived into the control box and fitted one across the controller relay coil that switches the Fogbuster ON and OFF and another one across the outlet from the control box feeding the Fogbuster solenoid coil. See picture below.
The problem seemed to be improved but still occasionally the solenoid did not switch OFF.
The Fogbuster solenoid has a clear housing over the activation coil connections and there is a LED inside this that comes on when the Fogbuster is switched on.
This connection housing plugs into the coil and the mechanics of the solenoid body and is released with a screw in the end. Toggling the coolant ON and OFF via the PathPilot user interface I could see the LED responding correctly to the ON and OFF commands but occasionally the solenoid was not closing. It was therefore not an electrical problem but mechanical.
On top of the solenoid housing is a single large nut which when released allows the solenoid coil to be lifted off. This leaves two countersink screws which hold the mechanical plunger housing in place and if these are removed the plunger can be gently removed. Inside the valve is very simple. A central hole allows the air to pass through and when the solenoid is de-energised a spring forces the solenoid plunger to seal this hole. I gave everything a thorough blast with compressed air and re-assembled it. Care is needed re-assembling as there is a tiny O ring seal on the plunger cover. The solenoid now responds correctly to the PathPilot commands.
Another tick. Job done. Warm glow.
That other job is still sat on the bench glaring at me ….