Getting Rough Surface when scanning smooth item

[Burl Source]

I tried to find any posted info before asking this question. I am sure it is probably happening because of operator error.

When I scan a smooth item (pistol grip) the scan ends up with a rough surface texture between that of rough concrete and asphalt. I am using a new machine (fall 2010) from CarveWright that came equipped with the scanning probe.

I have tried with each quality setting for the scan and all end up with the rough surface.
When I attached the item to be scanned to the scanning sled first time was with double stick tape. When that did not work out smooth I thought the tape might be allowing the item to move. Next scan I glued the item to the sled with CA glue so there was no way the item could move.

The image from the scan shows the rough surface and when carved it looks just like the image, rough. I would post the pattern image if someone is will willing to tell me how. I know this machine can do great stuff. Just need a little help getting the operator to function correctly.
Thank You in advance for your good help.

[mtylerfl]

Hello,

If the surface is VERY "textured" it is almost certainly due to movement of the item being scanned. Double-stick tape might not be holding the item firmly enough to prevent movement that you would be not be able to observe, but that's probably what's happening. Strategically placed hot glue might be better, but even some flavors of that is too "rubbery" to prevent movement. You may have to experiment.

If the surface is rather LIGHTLY textured overall, this is normal for any scanning operation done with a "touching probe". Only way to avoid that is with a laser scanner, unfortunately. Usually a light texture can be smoothed somewhat in the Pattern Editor, but use smoothing SPARINGLY...I see a lot of patterns where smoothing was overdone and the result looks a little like melted chocolate and the finer details are totally lost. I prefer to leave a little texture if necessary, then just realize that I can sand that out at finishing time after the carve.

Hope that helps. Holler if you have more questions. Also, you can post the MPW of your scan if you want us to take a peek at it to help you 'judge' whether the texture is normal or not.

[atauer]

One thing that I didn't read in either of the two posts so far concerned the flexshaft. When doing a scan, if you leave the flexshaft inserted into the truck, you can get some vibrations which will result in a lower quality scan. Also, sometimes when closing the clear cover, the motor tends to spin up for a split second, and if you have the flexshaft inserted, its going to spin the chuck, which in turn spins your probe, which could potentially rip the cables out of the probe.

[mtylerfl]

One thing that I didn't read in either of the two posts so far concerned the flexshaft. When doing a scan, if you leave the flexshaft inserted into the truck, you can get some vibrations which will result in a lower quality scan. Also, sometimes when closing the clear cover, the motor tends to spin up for a split second, and if you have the flexshaft inserted, its going to spin the chuck, which in turn spins your probe, which could potentially rip the cables out of the probe.

Good point, Alex! I forgot to mention that. Thanks!

[atauer]

You would be surprised at how many people are unaware of that. It may be a small tip, but it can make a big difference. The user just needs to make sure that when the flexshaft is removed, that the cable and spring do not fall out on the floor and get lost or stepped on. I have seen that happen a few times. Whenever I remove the shaft for a long period of time, I take a small sandwich bag and a twist tie, and place it over the end of the shaft. Seems kinda crazy, but I have yet to lose a core. We even ship them this way to prevent loss of the core. It works really well.

[cnsranch]

FWIW -

This isn't "technically" correct, but my take on scans it this -

In order to get an accurate scan, the probe has to be barely touching the item - the sensitivity has to be set lightly or we wouldn't get a good image. So, imagine as that the point of the probe is moving along the surface of the item to be carved, it picks up any variation in the item. That "jackhammer" sound we hear is the movement of the probe - it's almost like it's being dragged along the item. So, the resulting scan has minute bumps all over it - that's what the probe is picking up.

Bringing the scan into PE, and smoothing the scan ,very lightly, eliminates most, if not all of the bumpiness without losing the integrity of the scan. Be careful with smoothing, though - it doesn't take much to be too much.

A few years ago, MT did a write up on the probe here in the Forum - it's here somewhere, I just can't find it right now. It was an excellent explanation of what to do with the scan and how to refine it once it's been uploaded into Designer. Some of the points he made then don't apply now (one example is being concerned with the scan losing it's original size - the software now takes care of that), but it was a goodie.

Frankly, I don't know if that write-up was put into a Tips and Tricks edition or not, but if you can find it via searching for it, it's worth your time.

Final tip (learned the hard way) - immediately save the original scan as an MPW before making any changes in PE. That way, you can always get the original scan back if you're changes are too radical (that came from MT's many warnings about MPW's).

[Burl Source]

Here is an attempt to post one of the scans with the rough surface.
The piece scanned was smooth (sanded to 600 grit). The item scanned was superglued to the sled so no chance of movement or vibration. I am doing the scans with the flexshaft detached.

EDIT
Uploaded wrong file but this shows an accurate image of what the scan surface is like.

[AskBud]

Here is an attempt to post one of the scans with the rough surface.
The piece scanned was smooth (sanded to 600 grit). The item scanned was superglued to the sled so no chance of movement or vibration. I am doing the scans with the flexshaft detached.

EDIT
Uploaded wrong file but this shows an accurate image of what the scan surface is like.
Let's see if we can prove, or locate your scan problems.

Mount one of the pistol grips on the sled, but turn it 90 degrees.
Also, mount a quarter or larger coin/medal, that has highly defined features, nearby.
Program your scan to do both items at once. I use hot glue to mount my models.

My theory is that you will find the the marking on the grip will be about the same, and the coin/medal will be very true.

I assisted another member quite a while ago with a similar situation where he carved the praying hands. He cut away all but the hands, sanded, and re-scanned his carving. The resulting scan revealed all the grain of the wood.
I think you may find the same result.
AskBud

[cnsranch]

Couple things, Mark -

Pics below are your grips, and a scan I did a couple of weeks ago with no smoothing, etc. - right out of the scanner.

Mine (pic on the right) needs to be cleaned up in Pattern Editor, and with a little smoothing, I'll be good to go.

Yours (pic on the left) is a different story - you can see lots of definition on the scan, even looks like you can see just where the probe did and did not touch the surface.\


Make sure you check the set screws on the probe - mine had worked loose, causing the probe to be sloppy in the adapter.

Also, you may want (if you haven't already) to re-scan using optimum settings, that is, if the one you did was set to "good".

[Burl Source]

Ask Bud and CNS Ranch,
My apologies. I got tied up with other stuff and was unable to check the forum before now. Thank you for taking the time to give your good advice.
The set screws on the probe are tight. I was unable to loosen them. When I toggle around the probe point it does not always return to the same position. I am not sure if that is an issue or not.

Right now I am running a new scan with 3 different items glued in place. It will take about 3 hours but I will post the results when I see how it turns out.

Once again, my apologies. You were probably thinking "Darned new guy, asks a question and never checks back for the answers. "
Thank you for your help. It is very much appreciated.

[Burl Source]

I finished the scan with a grip positioned differently, a knife blade and a quarter all glued securely in place. Still seems pretty rough to me.

In the previous post I mentioned the probe tip does not return to the same spot when toggled around and released. When the scanner probe is going on a downhill slope of an object (knife Blade) the sound changes to sound a bit like the rattle of a loose piece of metal. Only when scanning in the downhill direction, not uphill.

This scan was done on best setting with the flex shaft detached.

[AskBud]

OK,
Now, do you know anyone close by that has a machine?
If so, you could take your sled and probe to their place and run the test again. The coin should have been very clean (even if it is a light image). No dimples!
This could show if you have a Probe, or machine issue.
Let us know.
AskBud

[fwharris]

Mark,

Based on your ptn and your statement about the probe tip not returning to the center spot I would say there might be something wrong with the guts of the probe. It could be a spring out of position or broken.

If you/probe are still under warranty I would give them a call and also email them the mpw to look at.

[cnsranch]

Mark

As I recall, there are only three screws that hold the probe together. Remove those and take a look at the guts of the probe, if there's anything amiss, you'll see it right away.

You'll be surprised just how simple the thing is.

[Burl Source]

OK,
Now, do you know anyone close by that has a machine?
If so, you could take your sled and probe to their place and run the test again. The coin should have been very clean (even if it is a light image). No dimples!
This could show if you have a Probe, or machine issue.
Let us know.
AskBud

I don't know of anyone nearby with a machine.
I think 1st thing I will do is take apart the probe and see what's up there. Or......maybe I'll call headquarters 1st so I don't void the warranty or anything.
I will let you guys know what happens.
Thanks for all the help and good advice.
Mark

[Burl Source]

I talked to the guys at HQ and it sounds like there is issues with the probe I have. 2 of the 3 screws are too tight to be able to loosen so I will be sending the probe to have it looked at. They were very friendly and helpful. Figured I would try one more scan. The positioning on the sled helps some. I will see what I can do in the way of cleanup on this scan.

This shows the new scan but I noticed another problem.
On the flash card it shows the scan as 4.94 x 4.07, when I download the scan into pattern editor it changes in size to 3.406 x 2.906
If I attempt to change the 1st number to 4.94 the second changes to 4.215 when the proportions lock is left on.

[mtylerfl]

Hi Mark,

The scan quality you are getting is certainly not normal, as long as the items are secure (no movement) during the scan. From what you shared before, it sounds like you indeed DO have them secure.

As far as the size change, verify with us that you are using the most current version of the Designer software...one of the "fixes" awhile ago was the resizing issue of scans. The size should remain the same when you bring the scan into the Pattern Editor, as long as you are using the most current version of the software. Please let us know.

I'm sorry you'll be without your scanning probe during the round-trip to HQ. Based on what you've already told us, I tend to agree that something was amiss with the probe itself, and I'm sure the folks at CarveWright will be able to get you fixed up shortly.

While you are waiting for the probe to be returned, you may find the following Tips & Tricks interesting reading (maybe you've already read it?). It's an article with some odds and ends tips regarding the use of the probe, etc.

ISSUE 9 June 2008 – Scanning Probe Techniques (http://www.carvewright.com/downloads/tips/CarveWrightTips&Tricks_Jun08.pdf)

All the Best...

[Fletcher]

You'll be surprised just how simple the thing is.

Yep, you can probably fix it yourself as it sounds like what fwharris said, a spring out of position or broken. The probe itself is just a mechanical switch and since your machine is "generally" scanning the parts, it is still receiving signals from the probe and converting those signals to points in space - therefore probably not a software or electronic issue. If you look at the attached image of my scan(on Normal mode) beside yours, you can see that there must be a mechanical issue and the slop you mention in the tip makes sense to give you this output. Good luck!

[Burl Source]

As far as the size change, verify with us that you are using the most current version of the Designer software...one of the "fixes" awhile ago was the resizing issue of scans. The size should remain the same when you bring the scan into the Pattern Editor, as long as you are using the most current version of the software. Please let us know.



OK, first of all, time for a big DUHHHHHHH.
I had read in other threads about the update for the resizing but hadn't bothered to see if I was running the current software. I had 1.175. I have updated to the current one now. Now I will try one more time to take it apart before sending the probe in to be looked at. Fletcher, thanks for the comparison. Big difference even without considering yours was done at normal and mine at best. We'll get this all figured out and then I can get started with some of my wild ideas for projects. Thanks for the time and effort you guys.


Update
Took the probe apart and didn't see anything obvious. Reassembled making sure everything is lined up. Now I am running a new scan to see if there is any difference.

[Burl Source]

This is the new scan of a pair of grips.
A little cleaner than before but not quite right compared to what Fletcher posted.

What do you guys think? Good enough? or send it in to be looked at? I can make this quality of scan work for some projects but if I settle for this it will limit some of my other ideas. Maybe I am just being too optimistic. I will do what you guys advise me to do.
Thanks, Mark

[fwharris]

Mark,

How much if any clean up ( smooth/de noise) did you do? It does look a lot cleaner than what you were getting.

And what did you find when you took the probe apart?

[Burl Source]

Mark,

How much if any clean up ( smooth/de noise) did you do? It does look a lot cleaner than what you were getting.

And what did you find when you took the probe apart?

In the scan I had not done any cleanup. Just fill in of the holes. The probe looked ok when taken apart. I was careful everything was in place when reassembling. This is a definite improvement, just not sure if this is as good as it gets.

[Burl Source]

Just did a test carve and it looked good. At best setting a lot of the bumps that showed on the scan did not carry over to the wood. I will try with optimal setting and a harder wood and see how that looks.

[mtylerfl]

In the scan I had not done any cleanup. Just fill in of the holes. The probe looked ok when taken apart. I was careful everything was in place when reassembling. This is a definite improvement, just not sure if this is as good as it gets.

Hi Mark,

That's about right - always going to have a bit of a texture with a touch-probe. That looks about normal to me. Sometimes you can get a little smoother, but again, yours looks about par now.

[Burl Source]

Hi Mark,

That's about right - always going to have a bit of a texture with a touch-probe. That looks about normal to me. Sometimes you can get a little smoother, but again, yours looks about par now.

Thank You Michael,
In the test scans I have seen how the positioning of the item to be scanned can affect the final outcome (carving too). Some of what looks to be flaws in the scan do not even show up on the carving. I guess next thing for me is to get more proficient with the software. Thanks everyone for all the good help.