parlour.nz

Felder Format-4 Profit H08 CNC

Format-4 Profit H08 CNC

—

by

Jason
in

Having picked up this machine 2nd hand earlier this year, within Aotearoa NZ, from a joinery firm that had gone into liquidation (so I couldn’t ask them for information). And then also having multiple parts lost and damaged in transit to me, I’ve spent many days trying to make sense of what was going wrong, and how to fix it.

Feedback from Felder Europe was automatically redirected to the local distributors and official local technical support of ‘Jacks’ NZ, which would typically take weeks to reply to my various queries in the original configuration setup and fault finding. And typically whilst waiting for Jacks, I’d eventually find the answer through either trial and error, or getting feedback from an increasing network of local and independent professionals’ support.

So the following is what I’ve been learning so far. This information is mainly for my own reference, but also for everyone else as there is very little online support or guides for either tpaCAD, WoodFlash software, or really anything to do with the Format-4 CNCs.

Some links to YouTube and other CNC software content and information:

Digital copies of user manuals:

Power Requirements:

The machine lists a 3 Phase, 50A fused/rated supply requirement.

Using a clamp ammeter on any one of the positive phase wires (as most of the high draw hardware is three phase, but the PC just runs on one of the phases, etc.): when the router bit itself is cutting, and the twin gantries are moving around, it’s only around 2A to 5A (3 phase) total power draw, depending on the depth of cut, etc.

But then each of the high power vacuum pumps (which are kind of optional) are about 21 to 22A per vacuum pump (with the three phase supply).

So it could be up to 5A with no vacuum table operation.
Up to about 27A with one vacuum pump during regular use.
Or up to about 48A when both vacuum pumps are running during cutting.

Slightly more advanced admin password access to WoodFlash and tpaCAD:

  • To do anything other than the basic running of programs, you need to click on the yellow key icon in the Windows taskbar (base right of the screen), whilst any of the WoodFlash or tpaCAD software applications are running.  The slightly more advanced user password is ‘test’.
  • This should enable you to manually operator the pneumatic pins, and disable the vacuum pumps faster at the end of the program cycle.
  • If you need a higher level of access password for more advanced configurations, unfortunately you will need the daily passwords from an official service engineer, who will use their phone app to find out the four digit numeric password for today’s date.

Material Thickness from VCarve to WoodFlash

  • The easier option when in VCarve if you set the Z zero position to ‘material surface’ only when starting (or editing) a new file. After exporting the cutting path to .tcn, when you import that into WoodFlash just leave the thickness as imported within the line program (ie T = 18mm); keep the spoil board thickness (Lf) as it is, and leave the TS (material thickness on the machine bed) as blank or zero. 
  • Otherwise, if you set the the Z zero position to ‘machine bed’ within VCarve. When exporting to the .tcn file format and opening it up in WoodFlash, the ‘T’ (in the program line thickness) needs to be changed from its default imported thickness (ie 18mm) to 0.005mm, keep Lf (spoil board thickness) as is. And then TS (as part of WoodFlash) needs to be changed to your actual material thickness loaded onto the spoil board (the same as the thickness you’d programmed in VCarve, ie 18mm).

Adding & changing router tools to the software

  • After very carefully adding a new router bit and matching collet to one of the tool holders (keeping it all clean and dust free).  Use vernier callipers or similar to measure and double check the tool cutting diameter, maximum cutting depth (sticking out of the collet) and from the tip to the back of the wide top part of the tool holder (the widest part, just under the tapered protrusion that get picked up by the spindle).
  • Then open up the ‘cog’ program, go to the router bits section, choose your preferred tool bit and double click on the specific variables to be able to change any of it.
  • You can only choose an particular icon once from the selection (so you may need to do some pixel art to generate a new icon to match your own tool).  Check spindle speed, feed rate, plunge rate, etc. with the bit manufacturer.  tool depth (as shown with the picture) must be accurate to less than 5mm, otherwise can break the bit when automatically checking tool depth on the pressure plate.
  • Tool IDs seem to be auto-arranged numerically, but 301 – 308 (or 1301 to 1308) seem to be the defaults.
  • Hit save.  
  • Then open up the other tab within that cog software, to then drag and drop the new icon into the tool magazine.  Save again and exit.
  • Then return to WoodFlash and use the ‘preset tool’ option with ’test’ admin privileges, to get the machine to double check the tool depth (ie new popup window with a picture of the spindle and a router bit over the pressure plate).  Then enter the new tool number (ie 1304), and it should blast air against the pressure pad, then retrieve the new tool in the tool holder from the magazine, take it to the pressure pad and slowly lower it until it makes contact. Then it’ll move off to the side and you can ask it to check tool length in the other tools too.
  • Once clicking ‘end’, this should save the specifics itself into the ‘cog’ software, but you won’t see those changes until restarting the software.  So add all new tools within the software first, then close the cog program, then open Woodflash and check the tool lengths automatically, then close everything and possibly restart to ensure all the new settings are saved.
  • It’s also advisable to make a backup of the tool magazine settings at this stage tool (check the user manual for how).

Spoilboard milling & drilling

  • The spoilboard program needs the word ‘spoilboard’ as file name, to enable vacuum override pressure warning (under ‘spoil board milling’ as the override option.
  • When spoilboard drilling, change parametric from 0 to 1.  Also change material thickness in line item, to 0.005, and spoilboard thickness to 0.005.
  • But when milling the surface, set spoilboard thickness to intended size (and material thickness as 0.005mm)

Air constantly blowing when power to the machine 

  • This is designed as a ‘positive pressure’ through the main spindle to help try and keep the dust out.  It can be temporarily disabled with a custom air regulator add-on, added to the main gantry (not standard with these machines).. About 3-4 turns upwards stops air.  Remember to switch back on again before running the machine (hang a flag or something so you don’t forget).

Grease not getting to the tool heads / bearings / etc.

  • The grease within the machine should be a very expensive, but also quite thin/fluid grease.
  • The small pipes get easily blocked, and you’re not likely to know, until parts start to fail / make noise.
  • But you can check each of the 4mm mini-hoses by using a grease gun with a grease nipple – 4mm  hose adaptor (you may need to make one yourself).  Try squeezing grease down the line, if it’s unusually hard it’s blocked, so squeeze harder to try and clear the blockage, or trace the line all the way through (most likely blocked at tube exit) to clear the blockage if manual pressure doesn’t do it.

Thermal Breaker Circuit Warning

  • Particularly if semi-intermittent as an issue, and from a ‘cold startup’
  • If you get a ‘Thermal breaker circuit warning’, most likely this will be dust in the two breaker switches for the vacuum pumps, within the control tower (blue cables should read around 24v constantly).  If any of the blue cable lines (2,3,4) read less than around 24v, it’s most likely dust within that breaker.   Solution: with main isolator off, loosen the cable retaining screws, and then blast with compressed air whilst flicking the switch on and off a bunch of times.  Then retighten all those screws, restore main power, and test voltage again. 
  • If those four cables all ready around 24v.. but pin 12 input to the TPA board reads less than 24v, the problem switch will be further down the electrical line.. ie breaker for the tower cooling fan, etc.

Machine machines angled grooves into the spoil board

  • If when running the spoil board program, you get slightly angled grooves in the MDF around 60mm apart, it’s likely the main spindle has been kicked out of vertical alignment by bashing into something whilst machining.
  • Will likely need a technician’s help (even experienced professional help can easily take 4 hours to do)
  • Remove dust covers from around the spindle (whilst spindle is at the end of the machine, so you can drop them out the base by removing the end panels underneath where the ‘auto offloading conveyor’ would be).
  • Then after loosening the hex key bolts within the front aluminium block, loosen the lock nuts around the adjustment grub screws towards the top of the circular block, and fine tune in and out (with opposing actions).
  • Then retest with a micrometer within the spindle, rotating over the freshly machines spoilboard (whilst the table vacuum is also on to pull the table in).
  • Then replace all covers, etc.

Auto-tool changer dropping tool holders

  • Particularly if the machine favours tools at one end of the magazine rack, but drops tools at the other end, most likely the tool changer has been knocked out of alignment (when measuring parallel to the end of the matrix table).
  • Loosen the retaining bolts, and carefully reposition parallel to the matrix table (trying not to move it left or right).
  • Then try the ‘preset tool’ option, but keep the speed dial low, so you can carefully watch the spindle aim to pick up the tool changer.  Check visually in both X and Y axis directions when viewing, to ensure it’s centrally placed.
  • You can change the machine pickup profile within WoodFlash, but this requires the daily changing password from Felder (a four digit numeric code, based on the current date).  So if you don’t have that, try to manually realign the tool changer with the spindle as it lowers very slowly, at the two ends of the magazine, plus one of the central to double check.
  • If it’s about 1mm out in just one holder, it’s likely the retaining plastic fork mechanism is sitting slightly skew.

PC Machine can’t see the CNC

  • On booting up the PC, all the XYZ axis read 12345.0 – the PC can’t see the main CNC, so take out (and carefully plug back into the same socket) the network cables from the PC to the TPA control hardware. Restarting the PC (after a 30 second pause of the main power switch off) may also resolve it.
  • However, if in addition to this non of the yellow pneumatic lights are on at the CNC itself, the 24v transformer fuse may have blown within the transformer.  Test the voltage output of the transformer itself (should be around 24v).  If reading zero Volts at all outputs (but 230v going in), replace internally with a T6.3A 250V slow ceramic fuse, for 24V transformer.  These fuses are around 20cents each.  A new transformer is closer to $800.

Adding an MDF spoil board

  • Weirdly to me at least, MDF is air permeable when used on the CNC.  It’s about as porous to the machine as a kitchen paper towel is to a human.
  • Add a fresh sheet of 18mm MDF onto the clean matrix table, placed close to your normal full sheet placement (ie 1mm away from the pneumatic pins, to save interfering with them).
  • Add a coat of paint or varnish to the outside edge, all the way around the MDF sheet, to help channel the suction through the main table (like a giant wide stumpy straw).
  • Then mill both sides of the board (starting a test pass wider than you think the board is, and then slowly working downwards around around 0.25mm increments until you’ve milled a tiny bit of the entire surface). Using the ‘spoilboard’ default program and referencing the two appropriate tool IDs for an 8mm bit and the 80mm or so milling bit.
  • Flip the board over, and repeat milling off just a little over the whole surface.
  • Then machine the retaining bolt holes (running the spoil board program to use an 8mm router bit). 
  • Secure the spoil board to the matrix table with plastic bolts (that need to have a screwdriver groove in the top).
  • Re-mill the board after it gets carved into too much (ie 0.5mm inset from the main board), up to 0.5mm at a time.  Continuing to no lower than 10mm for the total remaining spoil board thickness (to avoid cutting into your plastic retaining bolts).

‘Emergency reset activated’

  • If no matter how many times you press the blue reset button, it continues to tell you an emergency button has been pressed, and it’s not the mushroom emergency button on the main tower by the keyboard, if you don’t also have the optional remote, nor the optional conveyor, it’s likely to be the emergency button towards the middle of the machine next to the vacuum table zone selection levers.

Cascade of multiple machine errors that don’t make sense.

  • If you’re getting loads of on-screen errors that don’t seem to make any sense, check the USB connections of the power transformer things in the front section of the machine.

Other CNC error codes

  • ‘Test ESP2.De20p: OFF’ – ‘Standby Electric Spindle 2 Suction Head Closed’.  If the machine starts to run as expected, and then seems to pause at the edge of the matrix table whilst the spindle is spinning, this error indicates the dust shroud has not dropped (it hangs off two pneumatic pistons), possibly because of you taking off the covers when trying to realign the main spindle to the spoil board.  If slowing down the machine with the dial (and not breaking the light barrier) so it doesn’t suddenly run into or away from you.. and then if pushing or pulling on the pneumatic piston slightly dislodges it, you may need to make some fine adjustments to that piston alignment with a 32mm spanner, or levering the piston just a little to manipulate the metal fixing bracket location.

First configuring the laser beams light barrier

  • You likely won’t be able to see the laser (and don’t try to look into it).
  • Moving a suitable magnet around the surface, of the base area of the light transmitter (attached to the side of the tower) should switch the laser light from just a hidden laser, to also a visual laser.
  • There should have been some paper cross-hair guides for the mirrors, for where the visual laser light should reach each mirror (it’ll be the upper end of the upper mirror, or lower end of the lower mirror. As the main laser reflects towards the centres of each mirror post).
  • Position and secure the mirror towers around the CNC, and then fine adjust each mirror with the adjustment screws around each mirror, in sequence until it’s positioned on the next in sequence mirror too.
  • Once making it all the way around to the laser receiver, that light should change from red to green.
  • You may also be able to bypass that warning circuit by shorting two of the blue cables in the back of the tower (refer to the wiring diagrams for which).
  • Alternatively, repositioning the transmitter and receiver on the side of the tower to bypass the mirrors might also be an option.

F0500 Invertor Error

  • This is a semi-generic error code of the invertor itself, that in my case would briefly flash on the invertor screen, and then relay a message to the main system and stop the program.
  • In my case, the error would happen pretty consistently in the same part of a new cutting program, but a different part of the matrix table with a different program. It didn’t matter how many times the same program would be restarted, the error would keep happening, most often within 20 seconds of the tool starting to cut.
  • The machine itself was not being overloaded (ie run speeds, spindle speed and power, etc. were all good). I took off protective covers from all over the machine and found piles of MDF dust across many of the connections, which were then cleaned. But the problem would keep repeating at a new spot on the board with each new cutting program. The only consistent thing was that it was inconsistent across programs, but consistent with each one. It didn’t matter how slowly the XYZ axist movement was decreased down to, the F0500 error would keep returning.
  • Thankfully one of the CNC Engineers I spoke to had some ideas, so asked me to pause the XYZ movement at the start of a program, and just leave the machine running with the spindle at full speed for a few minutes : this ran fine the whole time, but then would stop again soon after starting to cut.
  • So further isolating the issue with a potential 3 phase cable fracture (most likely on the X-axis, and within the main loom of cables running the length of the machine), we took the CNC to a position of a known program, close to where it would normally stop by itself with the overload error, but then go and shake a bunch of the cables running to the gantry. Although this didn’t trigger the fault code, it did enable the CNC to run for up to a couple of minutes longer.
  • After trying to feel my way along as much of the orange 3 phase cable as possible, and also ruling out a loose power connection at the main spindle itself (re-enforced with some more cable ties). The fault was finally traced to the ‘snake pit’ of cables at the top of the main gantry, just as all the cables come out of the main channel and then split to the various panels. The main orange cable had been made just a little too long, and so had been slowly rubbing against the base metal of the ‘snake pit’ and a sharp edge from a panel at the base of that box. This had then worn through the outer insulation and so was causing a low level short with the earthing sleeve around the 3 phase, and that was triggering the F0500 error.
  • Fix: lift the orange 3 phase cable slightly further out of the ‘snake pit’ and cable tie it where it leaves the cable track. Similarly secure it the other end of the pit just as it goes into some conduit to go to the spindle, but still leave enough slack for full height movement of the spindle, etc.
  • Re-insulate the damaged insulation of the 3 phase cable, and check for any other fracturing (ideally replace the whole cable, or cut out and re-double-insulate the damaged section).
  • Secure some low friction foam at the base of the snake pit (over the hard metal plate edge) to help prevent further friction in the future for any cable.

Misalignment of the X-axis error

  • Due to human error, one time I forgot to re-open all the table vacuum valves again after switching back from a partial sheet to a full sheet of wood. As a result, when it released a newly cut panel from the sheet where there was no vacuum, that panel ended up jamming against the gantry as the spindle moved, thus shifting out the gantry slightly from square.
  • So using a precision straight edge, and manually bringing the whole gantry to near the X=0 position of the matrix table, I was able to trace the straight edge down the gantry Y-axis guide bars, and check for the gantry being completely parrallel to the matrix table. In my case it was out by around 1mm. So by looseing a few nuts and bolts, and some subsequent gentle percussive encouragement to the gantry with a rubber mallet, I was able to get the gantry much more parrallel again to the matrix table.
  • However, on re-zeroing (Set Point) the machine to start over, it warned me of the same misalignment of the X-axis. So even through I knew it was now very much square (and so not out of alignment), it must have been out of square before, but overridden as an error. That semed pretty common for the previous users of this particular machine.
  • Without the technician daily password, I don’t know how to override that error again. So after first pre-marking the metal bracket on the vacuum valve side of the machine with a sharpie, then loosening the lock bolts a little, after some very careful tapping with a soft mallet to tease the metal bracket used by the sensors, to a slightly forward or backward position (less than a mm difference from the original), I was able to overcome the error. But this did require running the Set-Point again multiple times, repositioning slightly, test & reset, etc.
  • The CNC is advertised as accurate to 1/10mm, and although both sensor brackets for the X-axis are in diiferent positions (around 80mm out of physical position relative to the end of the machine), you really should need to re-adjust the bracket much at all after the error to get it working again. Most likely it’ll be less than a mm of movement (which is why it’s so important to pre-mark its position before you loosen any bolts)
  • If you have the full technician password, there’s probably an easier way to tell the machine to just accept the new X axis sensor positions relative to the main gantry. Unfortunately I don’t yet know how to do that, but I will add to this blog post if and when I find out.