Once you have the X-Carve stiffened, tweaked, and perfectly dialed-in, there is one more step you can do to take the X-Carve into the realm of precision CNC routing.
I have recently discovered ultra high precision aftermarket collets manufactured in the US by Think & Tinker, Ltd. under the PreciseBits brand. I have seen one of these installed on a Bosch 1617 router which was used as a CNC spindle. It reduced the runout of the collet/bit to a point that it couldn’t be measured by a dial indicator. They also have one available for the DeWalt 611 that a lot of us have chosen to use with the X-Carve. I am seriously considering this upgrade based on what I witnessed with the Bosch 1617 results. There is a video on the Tube 'O You posted by Brian Oltrogge demonstrating the upgrade on his Bosch 1617 routers.
I am not affiliated with Think & Tinker, Ltd. I am just extremely impressed with their products and wished to share with fellow X-Carvers.
I’ve had mine on the Hitachi for 5 years now and it sill works like a dream. Newbies might have to practice with the spanner and changing out the inserts for different size bits but it quickly becomes 2nd nature.
That was the first thing I did when I upgraded to the Dewalt 611. little on the pricey side at $80… but I got them right away with both email update before shipping and a follow up afterward to make sure I got it and that I was happy with it. got two collets for 1/8 and 1/4 shanks.
I don’t have the specs for Elaire collets, but the PreciseBits runout is 0.0004" for the standard “Precision Grade” and as tight as 0.0001" for their “Extreme Precision” line. They are capable of higher precision than Elaire, but… in the real world we are also limited by the router we are using and its built-in runout. I’m not sure you would notice increased resolution between PreciseBits and Elaire in the DeWalt 611 in machining the materials that the X Carve is normally being fed. For instance, the difference in real world use would not really be noticeable in wood or the other soft materials. Maybe only in PCB machining or aluminum milling. PreciseBits does have the superior design between the two. It is capable of keeping the bit straighter in the collet, therefore allowing less deflection during operation.
The stock collet and the Elaire collets are very good. In my opinion the lower runout of the Precision collect does not really make much difference since the tolerances of the rest of the machine will swamp the gains you get from the Precision collet. Save the money on more expensive collet and invest it in stiffing up the machine. You will see much better gains from a more rigid frame.
Obviously I’m going to be quite bias here so I’ll limit my response to general information on routers and runout.
A common misconception seems to be that routers of all types have an ER collet that will fit them. This doesn’t work as none of the routers on the market that I’m aware of use the taper angle for ER (8°). So even if you could mount it you would end up pinching the collet at one point and probably cocking it over to one side. That would give you VERY poor clamping and runout.
I see a lot of talk about reducing the runout on router not being worth it because it’s not going to increase the overall accuracy of the system due to other limitations. It’s true that for some (or possibly most) applications the accuracy bottleneck is not the runout. However, what is usually overlooked is that runout comes right off the top of your total chip load. An easy way to think of this is that as the tool wobbles due to runout it’s taking a bigger bite (basically it’s moving in more than one direction while it cuts). This means your cuts have to be made at slower feeds overall than if you were able to reduce the runout. This also effects forces that influence deflection, tool life, and cut quality.
For example let’s say we have a 1/8" cutter with a 5% chip load and one collet with 0.003" runout (not at all uncommon) and one collet with 0.001" runout and we will run at 16k RPM. If we do the calculation for both (0.05 x diameter - runout x RPM) we get 52 IPM for the 0.003" collet and 84 IPM for the 0.001" collet.
Hope that’s at least of some help. And again, not trying to push our stuff. I’m for ANYTHING that gets your runout down. And I also realize that everything is usually a compromise. Sometimes we have to work with what we have or funds limit us. Just trying to give out information so everyone can make as informed a decision as possible.
I’ve already done the steel bar gantry mod and over the weekend did my Y. I went ahead and purchased the precision collet that Inventables sells for the 1/8" so I don’t have to use a reducer and we’ll see how that goes. It ought to help since a reducer can’t be that good and I use 1" cut length 1/8" bits, so that’s a lot of potential for runout- at least I would think.
Sorry, I don’t want to break any rules here. Obviously I’m bias toward our own products. What I will say more generally is look for products that actually list tested runout. Many places with do a “typical” number (much more common in ER collets). This is usually where someone will take a certain number and measure that sample. Then use the average of that number for entire batches. This can result in a wide range of quality in the shipped product.