Well the last post for 2010 — I think I’ve made some progress with this machine.
I made up some end plates to hold the scale in position. They can out pretty nice, except the rectangular pockets were undersize and had to be filed out to fit the scale. They had been modelled to +0.005 over the size of the scale. It was not an easy task to remove the support from the back — perhaps I need to lower the support tip Z a bit.
Using the FDM 2000 that has all the covers off, I trimed the length of the scale to fit in. I planned on locating to the front-right corner of the chamber of the machine. I could see the intial plan isn’t going to work as the Z Home Sensor ends up burried. I’ll have to space the DRO head closer to the table, and have a back bracket that is just behind the head (not extending past).
There weren’t any issues with the bottom side.
I ran a back bracket, and like my other small parts it came out very rough. A lot of filing & sanding ahead on this… The notch to slide over the Z Home Bracket was pretty well filled in even though it was 0.100″ on the model.
I got it mounted
Fitted loosely into the FDM 1650, it looks like it will work. I ran the Z Axis up/down and I got the same measurements on the DRO compared to a dial indicator. I’m going to remake the end plates to locate the scale correctly, and remake the back bracket so it will locate on the Z Home Bracket better.
I also need to figure how to get the cord from the inside to the outside — the covers will have to come off for this.
I tried running some small parts to judge how my attempt at alignment did. I simply scaled down the block I had run before & also re-orientated it so the bore was vertical, so I wouldn’t use too much material and support. The photo below shows ugly result:
Also, the head seemed to pick up some of the filaments knocked off the tip during the ‘wipes’. As the program was wiping every two build layers, the wipes were more frequent with the small part.
I noticed as the part was running, the feed seemed faster, and as I have the machine on a dolly (to move in-and-out of storage) it seemed to vibrate a lot.
The thickness from the top of the foam to the bottom of the tray (not to the dimples) was 1.537″ before starting.
I cut a 10″ x 10″ 0.030″ stainless shim for between the foam bottom & tray:
The shim fit easily into the bottom of the tray:
I rotated the foam 90 degrees so there was new material for the pins to stick into. The total height after adding the shim was 1.587″ so I gained 0.050″:
I ran a calibration part (cb12abs.sml). At the ‘Pause’ I raised the Z as far as it would go, and the tip entered the foam slightly. The part looked pretty good, and the bottom layer was imbedded in the foam:
A view from the top of the front-left corner, looked about the same as it has previously:
A view from the top of the left-rear corner shows a small gap between the start of the top layer of support, and the ending of the support run:
A view of the front shows a ruff bottom model layer (from the coarseness of the foam) but nice even layers built on top of it:
In my calibration piece, the support trace is not right in the center of the model traces. It is on the outside of all four sides of the model square. It almost apprears the ‘scaling’ factor for the support is not the same as the ‘scaling’ factor for the model material (or ‘Tip’ and ‘Alt-Tip’ in the sml file). I don’t see a way in the sscal.bat to adjust the scaling, only the X’s, Y’s, & Z’s. And the again, perhaps this is normal?
I think my next step, is to try a small model and see if this makes any difference at all.
In my first attempts at using my FDM 1650, I didn’t have a clue what I was doing, so of course I had loads of problems. Initially, I had no X or Y motion. I found the printhead had been shrink wrapped in position for shipping. Even after getting all that plastic off the head would not move. When I found the ‘HaveBlue‘ pages I found the Asymtek manuals that helped me get the X and Y moving again.
I knew I had to send the .SML file to the printer over a serial connection, and as I looked thru the QuickSlice files directory, I saw a Hyperterminal config link. I fired it up and sent a simple .SML to the printer. The head was zipping all over the place, except where it should be — panic stop was the only answer. Then I found ‘FDM Send’ in the Quickslice start menu folder. This produced reasonable FDM motions but I could see I needed to get the offsets set correctly to get good parts. This lead me to the cb12abs.sml calibration part, the ‘Tip Offset’ in the Quickslice start menu folder, and then the sscal.bat that calls the offset16.sml file. Opening each of these files in a text editor, I quickly saw that understanding what these files were doing was going to take some time and a lot of effort.
I’m more used to CNC controls where parameters can be viewed and changed with ease. This FDM 1650 and the QuickSlice software is like a black box with a green and red button. Just press the appropriate button — call Stratasys for a service call if it doesn’t happen.
I don’t have a huge chunk of time to try to do this all at once, so it will be on a time-available frame. I do want address the differences between the HyperTerminal & FDM Send, but I’m not going to start there. I am going to start with some .SML files, and disassemble them using the ACL language manuals. I’m going to try to figure out the registers used and what they are used for. Also, I’m going to try and figure how to get data back out of the control, and put it on a PC.
There will be files that I use each session, and I will make them available at the beginning of the post, and you can right-click & ‘save Target’ to computer, or just click and they will open in a new window. Here are the files I worked with today:
offset16-SML-101211- A text formatted version of offset16.sml located in C:QuickSliceV60-qsjdDefaults – This is called at the end of the sscal.bat which can be excuted from a terminal window, or activated by clicking the ‘FDM Offset’ Icon in the QuickSlice start menu folder.
cb12abs-sml-101211- A text formatted version of cb12abs.sml located in C:QuickSliceV60-qsjdStandard1650 (NOTE: I also viewed the cb12abs.pc file which I believe has be formatted for HyperTerminal, to see exactly which commands were being sent. I think the ‘FDM Send’ strips some things out of the .SML file being sent)
I first looked at offset16.sml to see how/where the the X, Y, and Z offsets are inputted into the control. They were quickly found. The command used was VS – Variable Store, and appears to be in absolute, inches, and in 0.001’s without a decimal point. The XYZ registers used were 121-123 respectively:
Well that was an easy one — of to the next one cb12abs.SML. Yikes! There is a lot of stuff right at the beginning that I couldn’t find docs for — it must be setting up the Automove Control to execute a program (?maybe) — I’m going to skip over that for now, as I see an initialize command (IN):
Next I want to find where the program seems to start (visually watching it run). There’s an FH for Find Home in X and Y:
Find Home (X and Y)
And then an FZ to Find Z Home:
Once it finds Z Home, It is store a ‘0’ in registers 100 and 101. This may be the Tip X register and the Tip Y register (?maybe):
Then there is a very long Quickslice Commented section. As it is all comments it is not used by the control (in the .pc version all the comments are striped out):
The a PS command to Pause:
This next line stalled me — I think it is time to stop for the day.
Let me start by saying that nothing in this post is really important. They are just a few things I had to try, now that I cut the last foam block too far. I had 10 pieces of the blue foam squares for the ICW (casting wax), so I mounted one up. Of course it melted right away at the ABS temperatures.
The foam when scored and snapped appart apears to have a finer cell structure than the blue foam sheets used in the construction industry:
When I removed the blue foam from the tray, on the back size, there were spots where the foam seperated from the pressure required to hammer the pins in (it was almost impossible to get the regular pins in — I had to hammer them in place — perhaps a different pin is used for the blue foam):
HaveBlue mentioned using using Blue Painters Tape on a plastic base and it worked okay. I thought I would try that on my foam. I covered the top with three layers to get about a 0.015 buildup. I ran the calibration program, and ran the Z up as far as it would go.
The tape didn’t stick too well to the foam, so I had a pretty good idea that this wasn’t going to work. Sure enough, the tape began lifting from the edges. It did raise one side of the calibration, but I was able to finish the program.
All, in all, not too bad. The bottom layers of model material were very flat. This indicates that problem is probably caused by the Z.
I’m going to put a 0.030 sheet metal plate under my foam to keep going for now (another HaveBlue suggestion). I also ordered 10 pcs of Dimension foam base (for $125) which will probably get here in a few weeks. I’m also going to try and size up how hard it would be to stick a DRO on the Z axis…..
While doing calibrations, I was worried how much variability there is in the FDM process. This exercise calmed me down. I ran six calibrations, at exactly the same settings as I used on 12/7.
The calibration parts looked consistent start to finsh. Looks like I need a better way to take these picture too….
I’m going to the hardware store and get some painter’s tape, and raise the Z up a bit for the next trial.
Here’s just the photos and details of the equipment I used is after that:
I have many cameras, but I have been using mainly two in my posts. I have a Sony CyberShot 7.2Mp that shot the pictures that follow. Great photos, but I have to resize the images to 1024 px wide for WordPress (using SnagIt Editor). I also like using my old Kodak DC4800 that I have several closeup lenses and several filters for:
Also, I have several microscopes and shop scopes. The scope shots at 100x magnification (I think) above were using a flea market Fisher Scientific Steromaster (I used some tape guides on the stage to locate the calibration squares repeatably):
I stick a DCM130 Scope Camera in one side and connect to an old Dell C600 Laptop with docking station. I use a program called Scope Photo to record the images to memory.
This is the DCM130 unit by itself:
This is the Shop Scope unit that came with the DCM130 when I bought it:
I also have a CnCam (optical measuring webcam) I can also use. I got to dig that out for this project too…
To get off to a good start, I stuck the foam tray in a Bridgeport Mill and took 0.100 In off the top with a flycutter. There’s still some irregularity but it is far better that the hacked up surface I started with. I held the tray in a vise, with reversed jaws, with the three dimples on the bottom hanging off the top surface of the vise. I need to check the dimples on the tray to see if they are all the same height — if they are not, the tray will sit an angle in the chamber — no time today for that, but I’ll put on the list for later.
I powered up the FDM 1650, and let it get up to temperature (ABS Materials, 270 Model, 265 Support, and 70 Envelope). I wanted to run the FDM Tip Calibration which is outlined in the ‘Utility Manual’ on pages 19 and 20.
From the QuickSlice Start Menu Folder, I clicked the ‘FDM Send’ Icon. In the file dialog I browsed to C:QuickSliceV60-qsjdStandard1650 and I selected the cb12abs.abs file to send. The ‘Pause’ button light on the FDM 1650 began blinking, so I clicked the pause button to begin the Initalization and Homing sequences. After homing, the ‘Pause’ button light again started blinking. I selected the ‘Z-Axis’ for jogging, and jogged down so the tip was just below the surface of the foam. I need a better way to see exactly where the tip is — I’ll add that to my list also. I unselected the ‘Z-Axis’ and I clicked the pause button to begin the calibration build.
The model material only barely adhered to the foam base. The model was a parallelogram shape – not square. When the support tip moved into position, it knocked the model traces off the foam base. It appears the support tip is too low in Z. This is what the left-front corner of the calibration looked like under the scope. There’s no support layer like it should have as the support just squirted onto the foam:
I repeated the above, with the only difference being I tried to bury the tip a little deeper into the foam before starting. This did stick a little better, was square in shape, but still was knocked off the foam when the support tip moved into position.
The scope picture of the left-front corner:
I needed to get the support tip ‘Z’ correct. The manual said ‘Enter the new offset values by using the FDM Offset icon’. So I clicked the FDM Tip Offset icon:
This is scary! I’m more used to a CNC Machine offset table, where I can see the numbers in the control, etc…. What numbers do I input? What is this really doing? Right-Clicking the icon in start menu, and selecting ‘properties’, I can see that is launching a DOS batch file ; C:QuickSliceV60-qsjdbinsscal.bat . There are docs on sscal in the ‘Utilities Manual’ on pages 7-9. I decide I don’t know what’s in the control, so I had better select the ‘-d’ choice to set all values to default (what ever they are) — also I add to my list to check the Asymtek docs for a way to get parameters out of the unit to a PC.
I send the cb12abs.sml file again. This time the support ‘Z’ does not knock the model off the foam.
And the scope picture:
A definate improvement. I feel uncomfortable with the adjusting using the icon, so I start a terminal window. I enter the following commands to move X +.005 and Y +.010 using COM2:
cd C:QuickSliceV60-qsjdBin [Enter]
sscal -a .005 .01 0 b [Enter]
I send the cb12abs.sml file again. The result does not look a lot different:
And the scope picture:
I wonder if the values I sent changed the values in the Asymtek? I decide to send them again, so I start a terminal window. I enter the following commands to move X +.005 and Y +.010 using COM2 again:
cd C:QuickSliceV60-qsjdBin [Enter]
sscal -a .005 .01 0 b [Enter]
I send the cb12abs.sml file again. The result looks slightly different (or maybe that’s what I wanted to see):
And the scope picture:
I also tipped the front up to look at the layers:
I print out the sscal.bat, and it calls a ‘offset16.sml’ file — another item for my list of things to do.
Not being sure if anything had changed, I ran the cb12abs.sml again:
And the scope picture:
And the front view:
I need to be able to see what numbers are in the Asymtek control, this trial-and-error stuff seems like the hard way to do things…. I’ll pick up on this another day.
I was watching a FDM 2000 on eBay (#370463793935) and I was curious what the tag on the printhead said. I emailed the seller but didn’t hear back. Visiting the page again, I clicked the image having the printhead, a got a pretty good image of it. Zooming in the tag clearly shows it is stamped ‘ICW':
Looking at the powered on temp settings, you can see they are way lower than the ABS heads:
As I don’t have any interest in a casting machine, I’m not going to bid on this one. Below are the pics the seller posted (click the small image to get the big photos):
I shot some more photos of the X-Y Stage (the bad one) this morning, and there was so many rather than insert into the other post I stuck them in a new one.
In the two months or so since I took it appart, it seems my memory wasn’t as good as it should have been. Here’s a shot of the cable tensioning screws (when mounted in the chamber, it will be the left-rear corner):
Looking from the back-side, it is the side with the Z connectors (or just opposite side of X-Y connectors if you don’t have Z):
So, it does look like you can remove the 12 screws and back ‘U’ shaped cover plate to get at the cable tensioning adjustment. You have to be a contortionist to work around all the stuff that comes thru the top of the chamber though. I’ll open the back of the FDM 1650, snap a picture, and insert it here showing what you may have to deal with to get at the screws. Also, once you break the plate loose, you may not be able to move it too far due to the inside X and Y cables (they are clamped down).
Here’s the rest of the photo’s, around the inside, then how the covers were situated:
Bottom model layers are very uneven, and it’s taking 15 – 20 layers until it gets straightened out.
I tried some ‘scope pictures but they don’t seem to show me much (I think I need a better ‘scope cam).
Sometimes it is just ruff support layers that are getting telegraphed to the model layer.
Even good support roughness seems to allow rough first model layers.
It’s like the part is growing out of a tangled mess (artistic in a way but it will never sell).
Here’s what the support looked like under the ‘4’.
The material and support I have been using is what was loaded in the machines. It’s at least 8 years old and maybe 10. I have two sealed spools of P400 ABS & one small sealed spool of P400 Release left, in addition to what’s left in the machine. I’ve been sticking to the temperatures marked on the Stratasys spools. I’m going to stick an oven thermometer in the next run I do to see what the chamber temp is running compared to the control.
I need to try some smaller parts to get short run time & less consumable use, and make up a designed experiment to see if I can get a handle on what’s going on.