I don't recall seeing a thread dedicated to problems arising before, during, and after construction of a DIY CNC router, and thought it might be of interest to others, as well as open up discussion on certain topics:
Overplanning and Over-Designing
I'm sure there are many here who enjoy the process as well as the fruits of their labor. I'm as guilty as any in this matter. I do wonder how many of those in this category actually produce something for profit or hobby, or whether the act of building a machine is not a means to an end but the end itself. Obviously, constant modifications and similations can save a lot of time down the road, especially when expensive components are used. But there has to be a breaking point where too much time is spent in the design stage, and the machine remains as "vaporware."
Cost of Materials Not Fully Considered
Scrap metal may be inexpensive, but having them fabricated into what is needed may not be. Some processes are very difficult if not impossible without a fully equipped shop, and consideration for the cost of "sweat equity" needs to be taken. Bonding and leveling supplies like epoxy are very expensive. The number of even common nuts and bolts can be astronomical, and if can't buy them in bulk, so is their cost. This is one thing where a little over-planning can be a good thing, as a bill-of-materials can be generated from most CAD. Tools, abrasives, blades, drill bits, taps, solvents, finishing supplies, even electricity and water all add to the cost; though I'd bet a lot of jobs don't have these costs factored in.
Inexpensive Electronics
There have been many threads here concerning out favorite imported CNC drive chip, and some good ones. Usually, the cheapest of these imported drives cannot handle the voltages needed to efficiently drive the steppers they are normally sold with. Here's one area where not skimping can pay big dividends down the road, as I go over briefly later on here....
Moving Rail Versus Moving Bearing Blocks in Z Axis Carriage
I personally feel that the moving rail design can be stiffer than the moving bearing design. I would postulate that the moving bearing block designs work best with relatively smaller Z axis travels, while the moving rail design would work best for larger Z axis travels. I suppose linear components and construction of carriage and saddle play a big role. I also think the moving rail can cause some design issues that would make it more difficult to build than the moving bearing design. That said both methods have been made to work, though for large Z travels I still think moving rails are the way to go.
Software Not Considered in Machine Cost
I see this quite a bit. A killer machine is made with top notch components and great workmanship. High end electronics and mechanicals are used throughout. Yet these machines are run using two-decade-old software for CAD design and workarounds in the controller software; or the machines are run using somewhat limiting "freeware." I honestly don't get this. If simpler parts are what is needed, simpler mechanicals and electronics could possibly get the job done as well, and the money saved can be used for software that is likely easier to use, has more features, and can handle more intricate work. I guess I'm not as up-to-date with freeware packages, and they're probably a lot easier to use than even a few years ago. But my experience has been that the learning curve is a lot higher, and the navigation may not be as intuitive.
Mismatching of Leadscrews with Stepper Drives for a Given Application
This is another area where potential problems occur. I've been bitten by this bug early on as well. Mainly, why do my powerful 425in-oz steppers not have enough oomph to move my relatively light gantry any faster than 50ipm? And if I try to, it stalls? Maybe the stepper was a mismatch to my 10tpi leadscrews that I bought real cheap. Thank God I did because I ended up buying screws again - multi-start leadscrews. Using a bigger power supply and better drive literally more than doubled my performance. I'm definitely not an expert on sizing steppers to a particular situation but have tried to learn as much I could, and have made better decisions with later builds. There is great information by Mariss Freimanis on the GeckoDrive website, along with some equations that would get you closer than any guesswork would.
The Fruitless Pursuit of Chasing Unattainable Accuracy
One thing that I've had to wrap my brain around is that everything in the CNC world is built to a specific tolerance. It is the exception, not the rule, that zero-tolerance parts exist. And they may only exist under certain conditions or variables. From the steppers or servos, to the screws, to the linear rails, to the support bearings... even the spindle bearings and the endmills used, all contribute some form of "error." Even if you did buy the most expensive components, it would have to me mated to parts made to the same tolerance, and kept in a controlled climate. And even then, the parts may have some flex or play in them. It's not a trivial thing for a commercial machine to make a part to within .0005" or even .001", and even then you'd be at the mercy of the measuring device used to check.
The Damn Absolute I,J/Incremental I,J Setting in Mach3
This damn setting has burnt me a few times, and I don't seem to learn. But to me it appears when a fresh install of Mach3 is done it defaults to Absolute, whereas the post-processor of some common CAM uses Incremental. I do have a CAM that has a configurable post-processor and can change this if needed...
Feedrates Too Slow/Spindle Speed Too High
This seems to be pretty common, especially in materials like Aluminum or Plastics. Commercial machines have much lower spindle speeds than routers, though they do have much more horsepower. This problem however may not be avoidable for some, with fixed-speed routers, or lower horsepower routers that do not have enough "guts" to power through materials at lower speed. A couple solutions would be to use bits or endmills with lower flute counts, or increase the feedrate. If the machine is not ridgid enough, smaller depths of cut may be needed along with the higher feedrates.
Wrong Bits Used For The Job
I started a thread on router bits and endmills, which is stickied to the top of this forum. While it can get expensive, it's always best to use the right tool for the job. The work will be cleaner, the tool will last longer, and your machine will run better. Of course, one could get carried away and have hoards of unused bits. The best thing to do would be to purchase bits as needed, instead of buying prepackaged sets. On the same topic, I am a firm believer in buying the best quality bits possible. There are many cheap imported bits on eBay, and my experiences have not been good. Believe it or not, high quality bits from reputable makers can be had for not much more than the cheap bits, and it just takes a little digging to find them on eBay. The aforementioned thread list brand names for searching.
LookAhead My Be Set Too Low in Mach3
I find I get better performance by increasing this number in the General Config screen. Apparantly it is set default at 20. Even 15 year old computers have enough processing power to set this number a bit higher. Some testing may need to be done. I have mine set at 192, though the max is 1000. Too high will cause problems. If you do a lot of 3D work this may save you a lot of time, especially if you can set you CV distance very low and have relatively high acceleration settings. There have been some reports of "smoothing" of certain details, which may be related to the CV settings. I have my CV distance set to 20 and my acceleration at 30, and do not experience any corner-rounding. A slightly oversized bit may also cause slight corner rounding as well...
Climb Versus Conventional Cutting, and Rough and Finish Passes
I find I get the best results, and less chatter, climb cutting as opposed to conventional cutting. I also find that the bit or endmill is less likely to "catch" the workpiece, since the tendancy is for the bit to push away from the cut line. Knowing this, I always leave a small amount for a finish pass, which gives me less bit deflection and tighter tolerances. This is especially true with materials like aluminum, and less ridgid machinesl though I find it also holds true for wood as well.
Overplanning and Over-Designing
I'm sure there are many here who enjoy the process as well as the fruits of their labor. I'm as guilty as any in this matter. I do wonder how many of those in this category actually produce something for profit or hobby, or whether the act of building a machine is not a means to an end but the end itself. Obviously, constant modifications and similations can save a lot of time down the road, especially when expensive components are used. But there has to be a breaking point where too much time is spent in the design stage, and the machine remains as "vaporware."
Cost of Materials Not Fully Considered
Scrap metal may be inexpensive, but having them fabricated into what is needed may not be. Some processes are very difficult if not impossible without a fully equipped shop, and consideration for the cost of "sweat equity" needs to be taken. Bonding and leveling supplies like epoxy are very expensive. The number of even common nuts and bolts can be astronomical, and if can't buy them in bulk, so is their cost. This is one thing where a little over-planning can be a good thing, as a bill-of-materials can be generated from most CAD. Tools, abrasives, blades, drill bits, taps, solvents, finishing supplies, even electricity and water all add to the cost; though I'd bet a lot of jobs don't have these costs factored in.
Inexpensive Electronics
There have been many threads here concerning out favorite imported CNC drive chip, and some good ones. Usually, the cheapest of these imported drives cannot handle the voltages needed to efficiently drive the steppers they are normally sold with. Here's one area where not skimping can pay big dividends down the road, as I go over briefly later on here....
Moving Rail Versus Moving Bearing Blocks in Z Axis Carriage
I personally feel that the moving rail design can be stiffer than the moving bearing design. I would postulate that the moving bearing block designs work best with relatively smaller Z axis travels, while the moving rail design would work best for larger Z axis travels. I suppose linear components and construction of carriage and saddle play a big role. I also think the moving rail can cause some design issues that would make it more difficult to build than the moving bearing design. That said both methods have been made to work, though for large Z travels I still think moving rails are the way to go.
Software Not Considered in Machine Cost
I see this quite a bit. A killer machine is made with top notch components and great workmanship. High end electronics and mechanicals are used throughout. Yet these machines are run using two-decade-old software for CAD design and workarounds in the controller software; or the machines are run using somewhat limiting "freeware." I honestly don't get this. If simpler parts are what is needed, simpler mechanicals and electronics could possibly get the job done as well, and the money saved can be used for software that is likely easier to use, has more features, and can handle more intricate work. I guess I'm not as up-to-date with freeware packages, and they're probably a lot easier to use than even a few years ago. But my experience has been that the learning curve is a lot higher, and the navigation may not be as intuitive.
Mismatching of Leadscrews with Stepper Drives for a Given Application
This is another area where potential problems occur. I've been bitten by this bug early on as well. Mainly, why do my powerful 425in-oz steppers not have enough oomph to move my relatively light gantry any faster than 50ipm? And if I try to, it stalls? Maybe the stepper was a mismatch to my 10tpi leadscrews that I bought real cheap. Thank God I did because I ended up buying screws again - multi-start leadscrews. Using a bigger power supply and better drive literally more than doubled my performance. I'm definitely not an expert on sizing steppers to a particular situation but have tried to learn as much I could, and have made better decisions with later builds. There is great information by Mariss Freimanis on the GeckoDrive website, along with some equations that would get you closer than any guesswork would.
The Fruitless Pursuit of Chasing Unattainable Accuracy
One thing that I've had to wrap my brain around is that everything in the CNC world is built to a specific tolerance. It is the exception, not the rule, that zero-tolerance parts exist. And they may only exist under certain conditions or variables. From the steppers or servos, to the screws, to the linear rails, to the support bearings... even the spindle bearings and the endmills used, all contribute some form of "error." Even if you did buy the most expensive components, it would have to me mated to parts made to the same tolerance, and kept in a controlled climate. And even then, the parts may have some flex or play in them. It's not a trivial thing for a commercial machine to make a part to within .0005" or even .001", and even then you'd be at the mercy of the measuring device used to check.
The Damn Absolute I,J/Incremental I,J Setting in Mach3
This damn setting has burnt me a few times, and I don't seem to learn. But to me it appears when a fresh install of Mach3 is done it defaults to Absolute, whereas the post-processor of some common CAM uses Incremental. I do have a CAM that has a configurable post-processor and can change this if needed...
Feedrates Too Slow/Spindle Speed Too High
This seems to be pretty common, especially in materials like Aluminum or Plastics. Commercial machines have much lower spindle speeds than routers, though they do have much more horsepower. This problem however may not be avoidable for some, with fixed-speed routers, or lower horsepower routers that do not have enough "guts" to power through materials at lower speed. A couple solutions would be to use bits or endmills with lower flute counts, or increase the feedrate. If the machine is not ridgid enough, smaller depths of cut may be needed along with the higher feedrates.
Wrong Bits Used For The Job
I started a thread on router bits and endmills, which is stickied to the top of this forum. While it can get expensive, it's always best to use the right tool for the job. The work will be cleaner, the tool will last longer, and your machine will run better. Of course, one could get carried away and have hoards of unused bits. The best thing to do would be to purchase bits as needed, instead of buying prepackaged sets. On the same topic, I am a firm believer in buying the best quality bits possible. There are many cheap imported bits on eBay, and my experiences have not been good. Believe it or not, high quality bits from reputable makers can be had for not much more than the cheap bits, and it just takes a little digging to find them on eBay. The aforementioned thread list brand names for searching.
LookAhead My Be Set Too Low in Mach3
I find I get better performance by increasing this number in the General Config screen. Apparantly it is set default at 20. Even 15 year old computers have enough processing power to set this number a bit higher. Some testing may need to be done. I have mine set at 192, though the max is 1000. Too high will cause problems. If you do a lot of 3D work this may save you a lot of time, especially if you can set you CV distance very low and have relatively high acceleration settings. There have been some reports of "smoothing" of certain details, which may be related to the CV settings. I have my CV distance set to 20 and my acceleration at 30, and do not experience any corner-rounding. A slightly oversized bit may also cause slight corner rounding as well...
Climb Versus Conventional Cutting, and Rough and Finish Passes
I find I get the best results, and less chatter, climb cutting as opposed to conventional cutting. I also find that the bit or endmill is less likely to "catch" the workpiece, since the tendancy is for the bit to push away from the cut line. Knowing this, I always leave a small amount for a finish pass, which gives me less bit deflection and tighter tolerances. This is especially true with materials like aluminum, and less ridgid machinesl though I find it also holds true for wood as well.