Open JackiePease opened 2 years ago
Don't forget to alert @DoESLiverpool/laser-maintenance
I’ve done some preliminary checks and the laser isn’t firing at all.
I have checked the door interlock switch and flow switch and these appear to be behaving as they should. Bypassing the interlock also doesn’t enable the laser firing. I have also compared the signals to Sophia, that uses a similar setup to Gerald.
Currently I suspect the HT power supply for the laser tube is faulty, I plan to pop into DoES on Saturday with a HT probe for my meter to do some further testing.
After some further investigation, I worked out some more details on the interlocks which was confusing me.
There are 2 separate interlocks in use on the laser tube power supply, one labelled WG and one labelled TL (there is another labelled TH, but this is not used).
WG is used for the Water Flow switch, this pin must be low for the laser to fire.
TL must also be low for the laser to fire. This is connected to 2 devices, the lid interlock switch and the laser controller. This is normally at approx 5V, but changes to a low voltage when the laser is fired by the laser controller as well as the main lid being closed. This caused some confusion at first, as it doesn’t change state if you just open the lid, the laser must be firing for this pin to change state.
With this issue clear and the interlocks behaving as they should, I then went on to test the output of the laser with a test pattern running (basically cutting a large rectangle at 70% power). The output of the laser tube power supply was approx 5kV. From some investigation on t’internet, a laser tube needs approx 27kV to trigger and a normal running voltage of 17kV, so it very much looks like the laser tube power supply is not well.
The quickest and likely cheapest way to get a new PSU isn’t likely amazon. I’ve found one that looks to fit the bill (though we may have to drill some new mounting holes)
This is a 100W PSU which will hopefully give us a bit more reliability over an 80W PSU, given that the current tube is an 80/100W tube and we have measure over 120W at full power. Delivery is Friday/Saturday for this PSU. The cost is £95.99
There is also the option of another unit from Cloudray, which is a good brand, but this £179 but is available by Wednesday.
The PSU has arrived. It's in the box on @johnmckerrell's desk, as it was addressed to him.
@magman2112 @Sean-anotherone Any update on this?
You've tried the power supply and that wasn't the problem, right? And you've narrowed it down to the controller board?
I agree that if we're replacing the controller board it makes sense to take the opportunity to upgrade it to run Lightburn. Have you got a list of the stuff we'd need to buy for the (wider) upgrade/overhaul? If the controller is going to take a while to arrive then getting that ordered sooner (even before you've finished drawing up a full shopping list?) seems to make sense...
Sorry for being a bit slow to update this issue.
I installed the new laser power supply and this apparently showed a similar issue to the previous power supply. The next most logical issue was a possible fault in the laser tube. Luckily, we had an old tube that was down on power, but still working, from the visor project period. I re-installed this tube and found that we still had an issue with the laser not firing.
After a bit more investigation work, @Sean-anotherone and I found that the controller was not outputting a PWM signal to set the laser power - hence no output from the laser.
It was at this stage that we came to the consensus decision that it was time to upgrade Gerald to a new controller and undertake a major service at the same time.
I anticipate 3 main phases to the upgrade.
I am working on a shopping list for all of the parts needed at present, but I'm struggling to find some of the parts at the right dimensions to replace the existing mechanical parts.
To start the ball rolling though, we should order the new laser controller from Cohesion as follows:
Cohesion 3D LaserBoard (including power supply- $229 External Stepper Driver Cable (4 Pack) - $10
Also, the Laser tube supports are available from eBay for £43.19 at the following link
https://www.ebay.co.uk/itm/402329611438
I have also found a UK source for the RECI W2 laser tube at a reasonable price £395. This is significantly cheaper than the $912 that Cloudray want from their UK warehouse, but it is something we don't appear to need immediately if we want to save some short term expenditure. Another link to eBay for this tube is below:
https://www.ebay.co.uk/itm/125035017738
I will add further posts to this thread is i identify the remaining items to purchase in the coming days.
I just remembered something else.
At a later stage, we plan to upgrade the z-axis and install more belt tensioning adjusters to hopefully eliminate the bed slipping issue on Gerald. As the Cohesion controller already has the capability for Z-axis control, we could likely add this facility as well, but only after the major functionality has been implemented.
All the belts on the z-axis will have to be removed and measured, as these are continuous belts, before ordering replacements.
Payment issues trying to order these - will try again tomorrow
I have ordered the board and stepper cables from Cohesion (warning of potential delivery delays) and the laser tube supports from eBay (should arrive by 21 Feb). It seems best to leave ordering the tube until we need it.
The Cohesion board and leads have arrived - they're on @DoESsean desk
Good stuff !
There are now programmes of electrical, mechanical and optical work to think about.
We can make a start on the electricals right away, I'm gonna suggest just gutting the present harness and doing it all again with new cable, as it's a bit of a birds nest.
The laser tube supports have also arrived - also on @DoESsean desk
All of the laser parts are now stored in the controller bay of Gerald, in case anyone goes looking for them.
We plan to start re-cabling Gerald next, including re-running the cable for the safety switch through the cable channel at the back of the machine.
Here's a first stab at the "coarse" level of jobs-to-do list. Controller: Remove the fibreglass carrier board from the electronics bay.Remove controller board Install Cohesion board on some standoffsRecable the stepper drives and all the suppliesRecable the stepper motors for all 3 axes (some decent flexible cable might be needed for this)Reinstall the carrier board and connect to the HV side Remove the small control panel from the machine and make/install a blanking plate (suggest selling the control panel on eBay, it's fairly new)
HV side Run multicore cable from the electronics bay to the HV bayRewire the HV control connectorRewire for the red-dot laser supplyRewire for the coolant and lid interlocks (I suggest running these in series as they are both n/o switches that close when the safe condition is met)
Laser Drill for and install new tube mounts3d or resin print some mount adapters to allow the current tube to be installed (will be replaced with a W2 at next opportunity, which needs the larger diameter mounts)Install tube and eyeball the alignment, pending the rest of the machine being finishedService / modify / improve the mirror mounts to make alignment easier in futureReplace the mirrors and focus lensReplace air assist tubing
Mechanicals Replace the shaft on the x axis reduction ( I made a new one already, its in the cupboard)Spec, order and install both Y axis rails and slidesMake the hole through which the water hoses , HV cable and red-dot supply enter the tube bay, bigger (right now they fit but it's a squeeze and they get in the way of the red dot mount plate)Replace X and Y belts ( if possible, install a linear spring on both, so tension won't need manual adjusting ever again)Make & install brackets as necessary to fit hall effect homing and end stops to the machineGrease, exercise and properly level the 4-off Z jacks, using dial gauge mounted on x axis sled to ensure focus true right across the bedRedesign, spec and buy parts for and install new Z adjustment drive
Other Gonna need more RO water, it comes from L9 7AJ - Mersey Aquatics, off long lane (take the 25L container there and they refill it for under a tenner, water is excellent quality <5uScm, better then cheap DI)A bigger/better air assist would improve the cut - Geralds is a fraction of the size of the one Tony came with, or swap Geralds air assist blower with Sheila?
That's probably enough to be getting on with for now... :)
Sean
On Saturday, 12 February 2022, 17:53:56 GMT, Mike Gorman ***@***.***> wrote:
All of the laser parts are now stored in the controller bay of Gerald, in case anyone goes looking for them.
We plan to start re-cabling Gerald next, including re-running the cable for the safety switch through the cable channel at the back of the machine.
— Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. You are receiving this because you were mentioned.Message ID: @.***>
Thanks for the write-up @Sean-anotherone. Presumably an extra body or two to volunteer to help out with the jobs wouldn't go amiss? If anyone fancies learning more about looking after the laser-cutters...
On 17 Feb I started the rework of the instrument bay, cohesion board is mounted and first tranche of cabling for that end of the machine is done.
Still needs the appropriate connections being made from cohesion board into the stepper drives, then the support board can be reinstalled.
I've updated the title so it's a bit clearer that this is the ongoing Gerald upgrade issue
Progress on the electronics today, with the main instrumentation panel being reinstalled. and the wiring for the magnetic lid interlock being worked out and readied for the main hookup session.
The next job is to remove the pulleys for the Y axis (front to back) and order new bearings for all of them. They are held in with crosshead screws backed by nuts that are accessible from the side-bays of the machine. The back right one has completely failed bearings, the others less so but they should all be replaced. If anyone wants to get stuck into this job that would be appreciated.
The main Y axis slides are confirmed good and do not require anything other than lubrication. The X axis slide rail was replaced last summer and is confirmed good.
The two main cabinet fans (lower bays left and right) also need testing, with replacements being ordered as necessary. They are 125mm (I think) 240Vac fans and should be fitted such that they draw fresh air into the machine (to prevent smoke from entering and gumming up the electronics) The old wiring has been cut, but enough is left to fit a plug and test the fans.
As ever, these are big works and are progressing, but any help would mean a quicker end to the overhaul.
Bearings required for refurbishment of y axis are…
4 off 8/16. X 4 mm
4 off 21/12. X 5 mm
Suggest using 2ss bearings for resistance to smoke
The bearings have been ordered from Simply Bearings. They should hopefully be with me by Tuesday. Total cost was £12.02, including postage, for name branded bearings, these were only about 10% dearer than no-name bearings so was an easy choice.
Request from @Sean-anotherone (who hasn't got his password-manager to hand to be able to log into his account :grin:):
Can we get 2x 110mm mains fans to replace the dead units in Gerald.
I've ordered two of these fans.
Fans have arrived and I've left them on Gerald.
@Sean-anotherone would like ~1m of drag track (or crocodile track) sized 15mm x 30mm, so we've ordered this one
The drag chain has arrived. I've left it in Gerald.
It has been a while since we updated this issue and there is a lot of progress to report.
@Sean-anotherone has completed all of the wiring for the new controller. This included the following.
Sean and I then re-installed the laser tube and worked through aligning it using the new laser mounts. We are now getting good registration across virtually all of the bed. Further tweaks will be needed to optimise these settings for a 100% coverage on the bed.
Control of the Z axis (Bed up and Down) is now controlled via Lightburn, the original bed control buttons have been removed. We still need to connect the limit switches for the bed at the moment.
There is an outstanding issue where the automatic homing on starting the machine is not working at present. This means that the home position is wherever the laser head is when the machine is switched on. This means that you have to manually move the laser head to the bottom left corner before switching the machine on. We are still investigating why this doesn’t work on the machine at present.
The main issue outstanding is the controller PC. The original PC was only 32 bit, so is not suitable to run Lightburn. It is possible to control Gerald using the PC for Tony, but this then only allows one machine to run at time. @MatthewCroughan and @Sean-anotherone did some work to configure one of the DoES shared Linux PC’s with Lightburn last Thursday Maker Night, but this is not fully optimised at present and the screen resolution doesn’t work well with Lightburn. It is possible for people to use their own copies of Lightburn and their own laptops to control Gerald.
Accuracy wise, X and Y precision is within 0.5mm, which we will improve over time. There is an issue with the Z axis at present. It still works but moves approx. 8mm for each commanded 10mm movement. I believe tuning of the stepper driver is needed to improve this (setting the configuration to the correct steps in Lightburn causes the motor to slip).
Overall though, Gerald is now mostly working with just a few snags to work through. Gerald is also cutting very well. I have changed some of the speed settings in the configuration to make cutting work at acceptable speeds, but there is still further tuning to be done with these parameters as well.
Once the PC issue has been resolved, it is time for people to start using Gerald in anger, so that we can find and fix any issues that arise in more normal operations. Things like automatic homing and z-axis controls are things we can work around at present until a full resolution can be found.
The resolution of the spare purple laptop we were thinking of using to control Gerald is 1366x768, which is too small for the QT5 Based GUI of Lightburn to utilize, the official developers of Lightburn recognise this too https://forum.lightburnsoftware.com/t/workable-screen-resolution-for-lb/14495/2
Mike, I should have a small Acer PC that can control LIghtburn. It's the one I used when you helped me set up in the garage. Currently in Iceland but back next week if you are interested. Let me know. Cheers. On Monday, 14 November 2022 at 13:08:00 GMT, Mike Gorman @.***> wrote:
It has been a while since we updated this issue and there is a lot of progress to report.
@Sean-anotherone has completed all of the wiring for the new controller. This included the following.
Sean and I then re-installed the laser tube and worked through aligning it using the new laser mounts. We are now getting good registration across virtually all of the bed. Further tweaks will be needed to optimise these settings for a 100% coverage on the bed.
Control of the Z axis (Bed up and Down) is now controlled via Lightburn, the original bed control buttons have been removed. We still need to connect the limit switches for the bed at the moment.
There is an outstanding issue where the automatic homing on starting the machine is not working at present. This means that the home position is wherever the laser head is when the machine is switched on. This means that you have to manually move the laser head to the bottom left corner before switching the machine on. We are still investigating why this doesn’t work on the machine at present.
The main issue outstanding is the controller PC. The original PC was only 32 bit, so is not suitable to run Lightburn. It is possible to control Gerald using the PC for Tony, but this then only allows one machine to run at time. @mattwilsondotuk and @Sean-anotherone did some work to configure one of the DoES shared Linux PC’s with Lightburn last Thursday Maker Night, but this is not fully optimised at present and the screen resolution doesn’t work well with Lightburn. It is possible for people to use their own copies of Lightburn and their own laptops to control Gerald.
Accuracy wise, X and Y precision is within 0.5mm, which we will improve over time. There is an issue with the Z axis at present. It still works but moves approx. 8mm for each commanded 10mm movement. I believe tuning of the stepper driver is needed to improve this (setting the configuration to the correct steps in Lightburn causes the motor to slip).
Overall though, Gerald is now mostly working with just a few snags to work through. Gerald is also cutting very well. I have changed some of the speed settings in the configuration to make cutting work at acceptable speeds, but there is still further tuning to be done with these parameters as well.
Once the PC issue has been resolved, it is time for people to start using Gerald in anger, so that we can find and fix any issues that arise in more normal operations. Things like automatic homing and z-axis controls are things we can work around at present until a full resolution can be found.
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are on a team that was mentioned.Message ID: @.***>
Thanks for the update @magman2112 (though I think it was @MatthewCroughan that you wanted to tag in :-)
The original PC might not have been a 32-bit machine, although the version of Windows on it will have been. That's because LaserCut only worked on 32-bit Windows (which took @ajlennon a lot of grief to track down, IIRC). So it might still be an option for use if the screen size is better for it (and it's a good enough spec)?
Mike, Adrian, I should have a small Acer PC that can control LIghtburn. It's the one I used when MikeG helped me set up in the garage. Currently in Iceland but back next week if you are interested. Let me know. Cheers.
On Monday, 14 November 2022 at 16:37:30 GMT, Adrian McEwen ***@***.***> wrote:
Thanks for the update @magman2112 (though I think it was @MatthewCroughan that you wanted to tag in :-)
The original PC might not have been a 32-bit machine, although the version of Windows on it will have been. That's because LaserCut only worked on 32-bit Windows (which took @ajlennon a lot of grief to track down, IIRC). So it might still be an option for use if the screen size is better for it (and it's a good enough spec)?
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are on a team that was mentioned.Message ID: @.***>
We made some significant progress on Gerald on the last Maker Night.
@Sean-anotherone brought in an install memory stick for Windows 10, with which he installed Windows 10 on one of the DoES Laptops. This was to eliminate some issues we have had in trying to get Lightburn to work under Linux. This PC now works well as the controller for Gerald, though we still need to install the License Key to fully complete the installation of LIghtburn on this PC.
The other main outstanding issue was homing on startup for Gerald. With a bit of detective work, we looked at trying different versions of the Smoothieware firmware. We found that it was the latest version of Smoothieware which affected the homing, as the latest version also stops the homing on Tony.
So with the right version of firmware and Windows 10 installed for Lightburn to work, the last 2 major issues have been cleared on Gerald, so it is ready for people to try now.
There are a few outstanding issues that we will look to resolve in the coming weeks.
First, we need to install a limit switch for the rear limit on the Y axis (there is only a limit switch on the front of this axis at present). This will make it safer for people to retract the laser head to clean the bed.
Second, we need to connect the limit switches for the Z-axis and confirm that they work. We also need to tune this axis, possibly on the stepper drive unit, as this axis doesn’t have the correct movement - it is currently moving approx 7mm for a requested 10mm movement. Tuning this in firmware reaches the limits that the stepper motor can drive at present.
We are also going to look at setting up the autofocus facility that the new controller gives us. We will need to identify and source a suitable limit switch (Which I have seen on Rob’s laser cutter) and make a bracket to install it, but this will make Gerald much easier to use going forwards.
Last, but no means least, we still need to improve the belt tensioning for the Z-axis, to reduce the slipping that still happens on this axis over time. This will become much more of an issue now that we have full control of the z-axis in Lightburn.
Brilliant stuff! Just back in the country now and playing catchup. Out of interest whatis the problem with Lightburn on Linux? Ta!
After a few odd days of work on Gerald, it’s time to catch up on the changes that have been made by @Sean-anotherone and I.
First, we have now connected the Z axis limit switches and they do work, but not as we expected. We need to work through the combination of controller and Lightburn settings to get this axis working in the right direction and homing correctly. At present, the bed is on its upper limit and Lightburn gives an error when you try to move the z-axis. We believe we have a way forward with this aspect, when we can next get into DoES.
Next, the IEC mains input socket broke (whilst I was trying to fix a laser alignment issue) which delayed work on Gerald for a while. A new combined switch/fuse/socket unit was installed last Maker Night which has now resolved this issue.
Finally, we knew we had some laser alignment issues as in #1741, so last Maker Night we finally made time to do a complete re-alignment on the laser. Prior to this fix, I had found that one of the laser mounting brackets was loose and also one of the spacer blocks was missing on one side of the tube - these are things we missed during the initial re-assembly of Gerald and all of these have now been fixed. After a few hours of work though, we believe we now have a good alignment on the laser and we’re getting good alignment across the whole bed.
So Gerald is again ready for use, with 2 current caveats.
Latest update.
Gerald is available for general use, is calibrated within about 0.1mm on 30mm test squares and cuts well right across the bed. The bed is now level within <1mm.
The motorised bed continues to defy logic and has caused no end of head scratching, there does not seem to be a combination of stop-switch or travel settings/inversions that allows it to Z-home in the right direction and be adjustable.
To focus Gerald, for simple jobs just slide the air-assist head ( which houses the focus lens ) up and down it it's housing exactly the way it's done on Tony. For thicker materials, the bed can be lowered using the lightburn controls in software, but it's tricky to focus using this approach, so once lowered, just use the slide method to fine focus.
Please raise the bed back up to the soft-zero position when finished, because if the bed is ground into the jack-stops it will skip the belts and take significant time to reset - this used to happen a lot with the old controls and is slightly easier now. Once the bed sensor logic is sorted out it won't be possible to do this.
As said, machine is now available for use, noting the new approach to fine focus, and any volunteers to try and get to the bottom of the z issue would be welcome!
I did an induction on Gerald this morning using the "Permission to Make" file. The line engraved text just looked like a scribble. I remember that it wasn't too bad before Lightburn was installed so I think the problem is likely to be related to Lightburn settings rather than hardware.
We have experimented a bit with the pulse-width and step frequency on Gerald, but this sounds like there is more to be done.
The stepper drives on Gerald are still the original ones and while they're quite old, they work fine. Almost everything else ( apart from the y axis bearings) has been replaced or overhauled, the belts are tight and the alignment good.
I recall that the manual is still available online, so perhaps a bit of time checking the pulse width and current limit settings against the optimal ones would be a good start?
An alternative would be to leverage that the new PSU in Gerald strikes the tube at lower powers than before, so just drawing the lettering more slowly and at lower power would achieve the same result, albeit a few seconds more slowly.
Gerald was also pressed back into service early when the HV failed on Tony, so this tuning can be wrapped up in the work to sort out the Z axis, which still misbehaves because of strange logic in the limit settings.
Sam reported that Gerald wasn't working when she was in this morning, but thought it might have been something she'd done. However it's still not working this evening.