BEWARE!!! Version 3.8.30

[PNast]

The pause resume demons are back AGAIN in Version 3.8.30. I paused to nudge the Z a little and on resume it went down and then tried to shoot across the table at 100 MPH! This has been a long running problem with the software. It was definitely fixed in the previous version. I was having problems with the pause and resume acting funny all day so I had my hand on the red stop button, luckily I was able to stop it before it did some real damage to my very expensive phenolic plastic vacuum table. I have gone back to version 3.8.26, no problems with this version.

Be very careful when pausing and resuming, this can be EXTREMELY dangerous when it shoots across the table so fast!

Paul

[khaos]

I am still on 3.6.x. No issues. No reason to upgrade.

[PNast]

I love the location presets, don't think I could live without them! Ver. 3.8.26 runs flawless.

[jTr]

I must say - previous version was the first time the restart/crash issue showed up in my machine. Had high hopes this one would have gotten rid of it, but thank you for sharing your experience - now I know, so I see no need to update since this is still an issue.

Interesting that we have different/inverted reactions with machine/software version combo's...

Feelin' your pain - wish there was a solution. Just need to continue reporting to SB tech in hopes a solution can be found.

The million dollar question(s): why does it seem only some of us experience this while a majority seem to have no issues like this at all (or silently endure)? Is it inevitable since we are all using different control computers? Are we not scrubbing the cpu's well enough?

Would love to hear from someone that stumbled upon a solution.

jeff

[Brady Watson]

Be sure to send a note about your problem to SB support. They do not monitor this forum for support issues on any reliable basis.

-B

[PNast]

I think there are so many different variables at play in each shop. Probably the only way to solve this would be for Shopbot to include a control computer with each machine. In the beginning I had a cheap laptop and I saw lots of notes about problems with slower cpu's, I was having a bunch of problems and bought the most powerful cpu I could get. Now I see discussions about too fast and multi-core cpu's having problems. It turned out to be buggy control software.

[David Iannone]

I think there are so many different variables at play in each shop. Probably the only way to solve this would be for Shopbot to include a control computer with each machine. In the beginning I had a cheap laptop and I saw lots of notes about problems with slower cpu's, I was having a bunch of problems and bought the most powerful cpu I could get. Now I see discussions about too fast and multi-core cpu's having problems. It turned out to be buggy control software.


Yep, if shopbot was to offer a control CPU for sale I would be first in line.

Dave

[PNast]

The million dollar question(s): why does it seem only some of us experience this while a majority seem to have no issues like this at all (or silently endure)? Is it inevitable since we are all using different control computers? Are we not scrubbing the cpu's well enough?

Would love to hear from someone that stumbled upon a solution.

jeff

I do mostly pocketed text signs with 3D carvings so I have a lot of bit changes. I use a 1/4" downcut Three Flute - Solid Carbide Low Helix Finisher (60-242) for the pocket cleanout and a 1/16" Single Flute - Solid Carbide Downcut Spiral O Flute (64-000) to go in and get all the material the 1/4" couldn't get. I have NEVER been able to get the 1/16" bit to zero on the money so I always have to pause/nudge/resume. This is most likely why I see this problem more than others. I used to hold down my boards with a bunch of Rockler clamps and figured the Z-zero was off because the middle of the board moved. I now have a gasket vacuum hold down system with a powerful Becker pump and still have the same Z-zero problems, even when using HDPE plastic. I know there is No way the plastic is moving.

If anyone has figured out a sure fire way to check that the next bit is zeroed correctly I would love to hear it!!!


2509425095

[scottp55]

Paul,
Forgive if out of line newbie, BUT had same problem with the Northern machine with Zeroing bits and he couldn't seem to get the same results I was:(
I swapped all my files to reference from the same exact spot(Home) on the spoilboard and then changed files to reference from spoilboard. His finished product may now vary from mine by .01-.005" in overall thickness, BUT at least his bits are now on the same Z-Zero page WITHOUT material variances.
Think you have a phenolic top? But also a full size machine. Any chance of setting up Z-Zero location outside spoilboard (or ON, but an area you don't ever use) and then referencing files from spoilboard and getting the material totally out of the equation?
Have to make different files for different material thickness, and may have to raise SafeZ a hair to allow for variance in material.
Just wondering.
scott

[dlcw]

I am still on 3.6.x. No issues. No reason to upgrade.

Yep. 3.6.x works great. I rely on my machine to make a living. I use eCabinets and Shopbot Link A LOT. If the control software were to cause me headaches like I'm seeing here with ver 3.8.x, I would have to outsource all the work my CNC currently does, and at that point I might as well shut my shop down.

It ain't broke, so I'm not going to fix it by "upgrading" control software.

If a REALLY stable ver of 3.8 came out (already loaded and configured on a manufacturer supplied computer), I might consider looking at it (plug and play, not plug and pray). But I would need a very quick and simple fall back to ver 3.6.x to be back in production (unplug the manufacturer supplied control computer and plug my 3.6 computer back in).

Just some thoughts from an old guy.... :D

[PNast]

I have tried it all! Zeroing to the same spot on the table, flat head bolt mounted in the table. I have noticed that the wire connected to the zero plate is very thin, maybe this has something to do with the inaccuracy? With my old setup I had the zero plate screwed down to the table so I could zero to the same spot over and over. I removed the wire from the zero plate and cutoff the little circle end. I then soldered an alligator clip on. This made the accuracy better but didn't completely solve the problem.

[Simops]

Yep. 3.6.x works great. I rely on my machine to make a living. I use eCabinets and Shopbot Link A LOT. If the control software were to cause me headaches like I'm seeing here with ver 3.8.x, I would have to outsource all the work my CNC currently does, and at that point I might as well shut my shop down.

It ain't broke, so I'm not going to fix it by "upgrading" control software.

If a REALLY stable ver of 3.8 came out (already loaded and configured on a manufacturer supplied computer), I might consider looking at it (plug and play, not plug and pray). But I would need a very quick and simple fall back to ver 3.6.x to be back in production (unplug the manufacturer supplied control computer and plug my 3.6 computer back in).

Just some thoughts from an old guy.... :D

Yep I agree. Im still on 3.6.44 and never had an issue arise (touch wood). Just not convinced yet with 3.8.30. Last thing I want is to have to muck around sorting out control issues.

I know a lot of you wNt a SB provided control computer but I also have Vectric installed and another CAD program for on the spot use and would not like to lose that ability.

Cheers

[PNast]

I really don't see SB offering a computer, the hardware and operating systems are always changing. Being able to load and run design software would be a must. For me, 3.8.26 still has an occasional hiccup when pause/resuming like losing position and crashing the software. It has never gone haywire and shot across the table like previous versions and 3.8.30. I reported the issue a few versions back and Ted had emailed me for more info on it so I am sure they are trying to fix it. I just wanted to make sure everyone who installed the new version is aware that this can happen. It can be dangerous and it has done damage to my machine in the past. I have broke many bits, crashed my dust shroud into the clamps and mangled it, damaged my table. I worry about my spindle when it crashes so hard. I am so paranoid now that I stay very close to the emergency stop button every time I pause/resume or start a new file. It wouldn't be so bad if I didn't have to keep nudging the 1/16" bit to get it to cut the same depth (or close) to the 1/4" bit. I still truly believe the Z-zero inaccuracy is because the wire to the zero plate is like 5 thin strands, it was a nightmare to get them soldered to the alligator clip. I think sometimes it doesn't see the connection fast enough, this is why the zero plate gets so chewed up over time. I put the clip to the bit to zero when I can but I have to clip to the collet with the 1/16" and it always zero's deeper.

[adrianm]

Don't see why they couldn't offer a completed dedicated control unit. The work they're doing with the Fabmo Engine seems to be pointing towards de-coupling the control program from being purely Windows based. With a completely dedicated solution the updating of OS and hardware becomes irrelevant.

[dlcw]

I really don't see SB offering a computer, the hardware and operating systems are always changing. Being able to load and run design software would be a must. For me, 3.8.26 still has an occasional hiccup when pause/resuming like losing position and crashing the software. It has never gone haywire and shot across the table like previous versions and 3.8.30. I reported the issue a few versions back and Ted had emailed me for more info on it so I am sure they are trying to fix it. I just wanted to make sure everyone who installed the new version is aware that this can happen. It can be dangerous and it has done damage to my machine in the past. I have broke many bits, crashed my dust shroud into the clamps and mangled it, damaged my table. I worry about my spindle when it crashes so hard. I am so paranoid now that I stay very close to the emergency stop button every time I pause/resume or start a new file. It wouldn't be so bad if I didn't have to keep nudging the 1/16" bit to get it to cut the same depth (or close) to the 1/4" bit. I still truly believe the Z-zero inaccuracy is because the wire to the zero plate is like 5 thin strands, it was a nightmare to get them soldered to the alligator clip. I think sometimes it doesn't see the connection fast enough, this is why the zero plate gets so chewed up over time. I put the clip to the bit to zero when I can but I have to clip to the collet with the 1/16" and it always zero's deeper.

You would think that Shopbot would be willing to offer financial help when your machine is damaged or bits broken because of bugs in their control software. Just a thought....

[Mark Farris]

I was fortunate enough to be invited to a small gathering of botters in St. Augustine Fl a couple months ago while Ted and Bill Young were in town. The final discussion topic was Ted revealing what they have been working on. He was very excited to tell us that they were going to be offering an upgraded card, or the addition of another card (wasn't quite sure of which), that would bring all the file processing and machine control "on board". Files would be loaded via USB connection, wifi, or network from your computer, tablet, or even smartphone. this will eliminate the control computer and and remove all the variables caused by different hardware and connection methods. He hopes to launch the option in the next year or so at an upgrade cost of a few hundred dollars. They are very excited about this upgrade and hope it will solve a lot of the problems most of us have experienced at one time or another. I am still at 3.6.xx because I have also had problems with the newest control software and am looking forward to the new system.


Thanks,
Mark

[PNast]

You would think that Shopbot would be willing to offer financial help when your machine is damaged or bits broken because of bugs in their control software. Just a thought....

The most important thing to me was getting the issue resolved and Shopbot's customer service rates Excellent in my book! They will work with you no matter what. I wish other companies would give half the customer service that SB gives. I would have never finished my vacuum table without all the help from Matt and Ryan. They did replace the dust shroud and the other stuff I wrote off as the cost of doing business.

The on board control card sounds great! I will buy as soon as it is available.

[Ajcoholic]

I was fortunate enough to be invited to a small gathering of botters in St. Augustine Fl a couple months ago while Ted and Bill Young were in town. The final discussion topic was Ted revealing what they have been working on. He was very excited to tell us that they were going to be offering an upgraded card, or the addition of another card (wasn't quite sure of which), that would bring all the file processing and machine control "on board". Files would be loaded via USB connection, wifi, or network from your computer, tablet, or even smartphone. this will eliminate the control computer and and remove all the variables caused by different hardware and connection methods. He hopes to launch the option in the next year or so at an upgrade cost of a few hundred dollars. They are very excited about this upgrade and hope it will solve a lot of the problems most of us have experienced at one time or another. I am still at 3.6.xx because I have also had problems with the newest control software and am looking forward to the new system.


Thanks,
Mark

Mark,
Would you still require the computer to set things up, move/center z zero etc? Im not understanding this that well, but from the sound of what you say you would transfer the cut file right into the "shopot" and forgo the computer transferring the file as it cuts, right? (or wrong?)

I seem to be very lucky. I have not ever had any issues with crashing, etc and i have been upgrading to the latest control soft/firm ware.

I still have no idea why some guys have so many problems, and others like myself have absolutely zero issues (and I mean zero). Im still using the same inexpensive laptop I bought when I got my machine. I know very little about computers, and why these issues affect some but not others. I sympathise with those having issues.. that would really cause frustration to say the least.

[Mark Farris]

Andrew,
It was my understanding that you will still use SB3 either on your computer or via an app on a tablet or smartphone for moving it around, zeroing, etc. The difference will be that the files will be onboard when you run them in a more controlled environment.

I must admit I havent tried later versions of the new control software since I had problems, but I just didn't have time to have problems right now. After working with Ted on my problem I suspect it had something to do with my specific control computer, but never figured out what. They were not able to duplicate my problems at Shopbot even after running my files on a variety of different control computers. It is very frustrating when it happens and I hope this new system will take care of everyones issues on an otherwise great machine.

It was good to see you again @ McGrew's.

Mark

[ted]

Hello Paul and others!

Problem with Stop/Resume in 3.8.30

We've been working on this one for the last week and I believe that we've made some progress. For those interested in the technical side, the issue is this; there are about 16 possible states in which the execution of a file can be at the time a Stop button is hit. Each one requires slightly different handling in order for the tool to stop in a measured manner, to pull out of the material, to allow actions like nudge, then to manage getting back into the material and continue at exactly the same place that it left off. It has been a pretty good challenge ... but we've always liked the idea of being able to resume after a Stop while providing options, rather than just leaving a file and starting over.

We have been able to replicate a problem with Stop/Resume that may be the one that Paul is experiencing. It happens when the Stop button is hit in a state such that: the tool is at the end of a jog, the Z height is below Safe-Z (defined in [VC]), and there are Moves (rather than another Jog) in the section immediately following the jog-Stop. With this issue, as with most, once we are able to identify and repeat some sort of misbehavior, it is usually pretty straightforward to fix it. I believe we have this one resolved.

There is new version of the software with the fix, Sb3.8.32, that we will make public on the website on Wednesday -- assuming it passes several days of testing here. *Interestingly, we get same failure in Sb3.8.26. Thus it is, apparently, an old problem (it just does not invariably show up because it is specific to a particular set of conditions at the time of the Stop).

** And, help please. After you download and install the new software. Take a few minutes, run some of your typical files in air-cutting mode above the material. Hit the Stop button in different parts of the file and see how this version works on your system. Let us hear from you. We think we've gotten this one addressed ... but it is very difficult to test the thousands of scenarios that come up with everyday use of tools in a variety of shops with different applications and using a wide range of computers.

That Issue with setting the Z-Depth

Paul, until I saw that first picture, I had not realized how far off your Z-heights were. Something does not look right there. So let me review a few things:

1. Let's assume that all issues covered by Scott's suggestion are taken care of. That is, that you are zeroing at exactly the same location on exactly the same height target. And, that you have the ground clip on the cutter shank. (Note that it is very difficult to assure a bolt-head is flat unless you have surfaced it in place -- different diameter cutters may touch it differently.)

2. My first question is whether it is similarly offset every time? In your picture, it looks like the finish-pass cutter is going deeper. Is this invariably the case? One possibility, is that the two cutters are contacting your zeroing target at different heights because they have different shapes and areas (e.g. the 1/4" may give a quicker electrical contact).

So let's look at the situation right after you have zeroed the second cutter and without moving the X or Y. If you move it down (slowly) to 0.0, using a thin shim to test the closing gap, does the Z move exactly to the surface of the target? If it does, then let's move it out in the middle of a cleared area and move the cutter down (slowly as above) to the depth of the pocket. Where is it now ... exactly at the material, or in it? If it's in it, then we know that the problematic offset has happened before cutting starts. If not, then we know something is happening after you have started the finish pass. If this latter case, how far do you have to go in the file before the offset becomes apparent?

3. For the moment, I'm going to assume the most interesting outcome. That, when you checked, it looked like the second cutter was exactly zeroed at the target, but that it was deep when you went out to the middle of the pocketed area? I'm goint to elaborate a little more on this possibility, even though it may not be the case ... just to give us some conversation points. The thing that is interesting about this case is that you've zeroed perfectly, but the depths still don't match. I think there are at least two kinds of explanation:

a. The first is related to the pressure from the 1/4" down-spiral during cutting. It is forcing the material down and the spindle up. If either material or spindle shows any give, movement, or mobility, then the cutter will essentially 'surf' over the material at a slightly higher pocketing level than you intended. Because the 1/16" cutter does not generate the same amount of force, it will not surf as much and end up cutting (what looks like) deeper.


How could the spindle move? There is a small amount of backlash in the gearboxes and potentially more in a loose or poorly engaged pinion. One or both could contribute to a problem like this. The good news here is that the Z axis can be 'preloaded' up or down by changing the spring tension (more or less spring). This adjustment can eliminate most backlash (especially when the tensions on the bearings are tweaked to make sure the Z does not bind). Here's how to check for backlash: put a dial indicator (facing up) directly under some part of the Z that will hit it when moving down; drive the Z down to a position midway in the travel of the dial idicator; zero the indicator and the tool; move the Z up to .5 and return to 0 noting dial indicator value; move the Z down to -.5 and return to 0 noting dial indicator value; the difference between the two values is the amount of backlash you have in the Z. This is because you have approached the zero from two different directions that take up the opposite component of the backlash.


How could the material be moving with the great vacuum system you have? If there is any tendency for the boards to move with pressure from above, then during cutting the force from the spindle could be pushing the material down. A down-cut tool produces an amazing amount of force as it moves along.



b. The second explanation is related to the effect of machining away material. Assume there is some amount of flow of air through the materal; if you remove 1/3'd of the material in a large area of pocketing, you've reduced by 1/3'd the downward pull on the board. I don't see any gasketing in your image, but this problem can be particularly severe if there is a gasket that has springiness and holds the material slightly above the plenum and the vacuum pulling it down.

***I'll finish this section by indicating that the problem Paul is describing is a tough one and has periodically arisen for signmakers who are pocketing large areas over time and trying to maintain a smooth finish with cutters that return to re-work of finish areas that have been previously machined, even in cases where there is no tool change. Because only a few thousandths of an inch of error will create a very noticable height difference, this can be very challenging. I know of shops that gave up vacuuming through MDF simply because of the change in thickness of the MDF from moisture removal over the course of many hours of machining would not allow this kind of pocketing. It is also the reason for the appeal of percision sign-board that can be mechanically clamped flat and has little tendency to wander.

Throwing Away the PC (well not totally)

As Mark indicated above, we are in the process of developing an entirely new control platform for digital fabrication. We call it 'FabMo' -- short for digital fabrication and motion platform. This new system is designed to get us away from the need to do the vulnerable USB streaming from PC to tool, and away from the dependence on interacting with a Windows PC to run your tools.

Until relatively recently, there has been little option for running CNC tools other with a PC, which in most systems does all the heavy lifting for motion control, right down to timing the stepping of the motors. Now, however, new microcontrollers have capabilities to handle much of the time-critical, low-level, motion planning and execution. As well, surprisingly powerful new single board computers (SBC's) have processors capable of management, communication, and interface functions. FabMo puts these both in the control box of a tool. They provide for efficient operation of fabrication tools in a number of different ways suited to the demands of the situation and from a number of different devices. This simplifies and standardizes tool operations and avoids the idiosyncracies of individual PC's. Most of us will still use a Windows PC for designing and toolpathing ... but we'll run the tool with our phone ... a tablet, a Mac, or a PC at the tool (wireless, USB, or ethernet) ... your choice.

We are introducing FabMo first for Handibots, probably at the end of July (we did a sneak preview at the Maker Faire in San Francisco earlier this month), but our intention is that this system will eventually serve ShopBots and digital fabrication tools more broadly. While developed by ShopBot and others, it is fully open-source and un-branded. You can read about it here (http://gofabmo.org) ... and I expect you'll be seeing it in video previews and on Handibots before long. Our plan is that any ShopBot can be upgraded to this new system for $200-$300, essentially plugging the new modules into the existing Interface Board in your Control Box.


Ted Hall, ShopBot Tools