Great topics!
One thing to know is that the Z-axis is just a solenoid with a spring return. A solenoid's force (-Z in this example) is proportional the the electrical power applied and how clost the soft iron slug (armature) is to the end of travel. This last I believe varies as the square of the penetration distance so the further you can get the slug to penetrate the solenoid, the higher the force. I measured between 4 and 7 volts on the solenoid, depending on "Thickness" selected. If you raise the inserted holder (be it for the swivel blade, pen, Dremel extension handpiece, etc), the downward force will increase until the solenoid bottoms out - with a metallic 'click'.
Not sure what's hanging up in your machine to cause it to skew. Obviously something mechanica I'd guessl. In my hopefully educated opinion, I would NOT remove the plastic rollers as rubber/elastomeric ones will not exert enough pressure on the PCB. Also, of the two metal rods that pinch the PCB, ONLY the bottom one is driven so you will get little/no torque/force out of the top one or anything you mount to it. The downward spring tension on the top metal rod and rollers, pushing the PCB into the bottom metal rod's knurling (grooves), is plenty to grab the material, IMHO. The purpose of this assembly is to force the material against the knurling on the bottom metal rod, the only thing transfering force to the material. Chosing "Enhanced Cut" will roll the PCB in and out (Y direction) wo hopefully wear a grooved path in the bottom of the material for additional friction. I have not needed this in the past. The copperclad against the grooves in the metal rollers gives you the best non-slip friction possible, in my opinion. I believe rubber will encourage slipping.
I think on my web site, I recommend 0.5 oz copperclad or thinner. I have had difficulty cutting through 1.0 Oz. I think Digikey sells this. Try Think and Tinker as a source for high speed carbide router, mill and drill bits. For the milling bits, I've had good results with the 30 deg ones. If you're trying the swivel knife approach, by all means go the 3rd party carbide blade route as the standard steel ones will wear out quickly. I have also found a source for paper backed copperclad instead of fiberglass, possibly extending the life of the blade. More later.
I have tried converting the Gerber 274x files to svg and purchased the Designer Pro addition of the software but I'm not sure this is required as I'm getting great results with their raster-to-vector TRACE app included. I'd like to hear more if you get better results. Also, I've learned from Silhouette Support that very closely spaced vector objects are drawn as one, possibly compromising small traces. This normally doesn't affect paper cutting other than being able to cut a lot faster through highly detailed, closely spaced vectors. They have offered to make a firmware change for "us". What a great support team!
Mike
>I just received my Sillhoute Cameo and have a few first impressions.
>Mike, if you could chime in here I'd appreciate it because I can't
>seem to achieve your results.
>
>1 - I cannot get a pcb to track through the machine accurately. The
>cutter seems to hang the board up so that it feeds irregularly,
>skewing sideways back and forth as it goes in and out of the machine.
>This seems to be true whatever speed I select.
>
>2 - Most of the tracks I get are not completely through the copper. I
>am using the highest setting on the cutter.
>
>I am using Gerbv to convert gerber files to svg, then importing the
>svg files using the "Designer Edition" of the Sillhoute software. It
>seems to work well, but the feed problems make it hard to tell if
>everything is accurate. I ordered a pen holder for the machine, so I
>should be able to use it as a plotter to see exactly what is happening.
>
>The machine looks like it would be very easy to modify to feed copper
>clad reliably. All it should take is putting a few o rings on the
>shaft for friction and removing the plastic rollers. Hopefully I'll
>have time to take it apart over the weekend.
>
>Mark
[Non-text portions of this message have been removed]