richards
10-08-2004, 02:42 AM
This is an awfully long post, but it might help a few fellow newbies understand the process of finding the correct rpm/feed-speed combination. . .
When I was learning to play the trumpet many years ago, my dad asked my mother, "When will the noise start sounding like music?"
Starting a few days ago, my Shopbot started playing sweet music.
I've been fighting feeds and speeds since the Alpha arrived in July. The machine has worked great, but I never felt like I was getting it to work as good as it could. Sometimes the router bogged down too much. Sometimes I fried bits when I ran up the RPMs or slowed down the speed. Sometimes the cut was not as smooth as expected. Sometimes I wondered if I would ever catch on.
However, when I stepped back a bit and looked back at experiments that I've run in a totally different field - photography, suddenly, everything began to make sense. For those of you who are photographers, you know that there are three basic things that determine whether an exposure is technically correct; film speed (ISO/ASA), shutter speed, and f/stop. You also know that if you know the correct f/stop and shutter speed for a certain exposure, that you can quickly determine other f/stops and shutter speeds. For example shooting a gray card with a tripod mounted camera using a film with an ISO/ASA rating of 125 at high noon on a cloudless day would require an exposure of 1/125 at f/16, 1/250 at f/11, 1/500 at f/8, or 1/1000 at f/5.6. Any one of those combinations would produce negatives/slides with the same graycard density. A good photographer would know when to use any one of the shutter speed f/stop combinations. A good photographer would also know when to rate the film at a higher or lower ISO/ASA value due to lighting conditions or processing variations.
With the Shopbot, there are also three factors that determine proper operation: chip load, router/spindle speed and feed rate. For example, if I wanted a chip load of 0.030 when using a single flute bit, my Porter-Cable 7518 has five different rpm/feed-rate combinations to achieve that particular chip load: 10,000 rpm at 5.0 ips, 13,000 rpm at 6.5 ips, 16,000 rpm at 8.0 ips, 19,000 rpm at 9.5 ips, or 21,000 rpm at 10.5 ips. Any one of those combinations would produce a chip load of 0.030 with that single-flute cutter. (For the sake of discussion, we'll totally ignore the router's power curve and assume that we're using a router that produces equal power at each speed setting.) The hard part for those of us who are novices at CNC routing, is to know which of those combinations is correct for the cut that we want to make. There is a simple way to find out. All that it requires is to run a series of tests at each speed. For example, starting at 10,000 rpm and 5.0 inches per second, I could make up to seven cuts in some 3/4-inch material; one cut at 0.10 inch depth, another fresh pass at 0.20 inch depth, another fresh pass at 0.30 inch, a fourth at 0.40 inch, and so on until the router started to slow down or until the maximum depth was reached. Next I could run a test at 13,000 rpm at 6.5 ips at those same depth settings. Again running the test to completion, or stopping if the router started to bog down. In the same manner, I could run a series of tests at the other rpm/feed-speed settings. At the end of the tests, I could pick any feed-speed/rpm combination that didn't bog down the router at the depth that I needed to cut.
However, if all tests passed without bogging down the router, which feed-speed/rpm combination would I choose? That would depend on the part that I needed to cut. For example, if I needed to cut small squares, 3-inches per side, where ramping would be a big factor, I would need to run an actual test at each feed-speed to see what the actual feed-speed with ramping included actually was. For that type of test, I would create a file without any z-axis moves or any jogs and without a speed setting (no MS,x-axis,z-axis command), manually enter the five ideal feed-speeds at the console, one for each test, air cut the file and then look at the logs to get the running times, and then divide the times by 12 (3+3+3+3) to get the actual inch-per-second produced. If, because of ramping, the fastest ips setting obtained was 6.5, even though I had asked for 10.5 ips, I would know that 13,000 rpm would probably be the correct router speed with a feed-speed of 10.5 (derated by the controller to 6.5 ips due to ramping) to get the desired chip load. Making a few test cuts at 13,000 rpm and 10.5 ips (derated by the controller to 6.5 ips because of ramping) would verify the test.
If the part to be cut was a cabinet side with long runs where ramping would only be a minor part of each cut, the fastest rpm/feed-speed combination could probably be used.
Of course, as experience was gained, I could use a different chip load setting, thus modifying the router's rpm or feed-speed to get an even better result, just as a professional photographer would alter the ISO/ASA rating of a film to tweak things based on his experience.
The last few days have been an awakening as the Alpha has cut thousands of 2x6 inch and 2x10 inch pockets 0.40 inch deep in popular and baltic birch. The first cut up the center of each pocket, using a 3/8-inch spiral, 2-flute bit, at 0.40 inch depth required a slow 1.25-ips to keep the router from bogging down too much in most of the popular, but the rest of the passes in each pocket were cut at 12-ips. The pockets were cut in half the time that it required before testing and tweaking. The quality of the cut was much better, and the same 3/8-inch bit was used without frying or premature dulling.
The sound of good quality work being produced efficiently is sweet music indeed.
Mike
When I was learning to play the trumpet many years ago, my dad asked my mother, "When will the noise start sounding like music?"
Starting a few days ago, my Shopbot started playing sweet music.
I've been fighting feeds and speeds since the Alpha arrived in July. The machine has worked great, but I never felt like I was getting it to work as good as it could. Sometimes the router bogged down too much. Sometimes I fried bits when I ran up the RPMs or slowed down the speed. Sometimes the cut was not as smooth as expected. Sometimes I wondered if I would ever catch on.
However, when I stepped back a bit and looked back at experiments that I've run in a totally different field - photography, suddenly, everything began to make sense. For those of you who are photographers, you know that there are three basic things that determine whether an exposure is technically correct; film speed (ISO/ASA), shutter speed, and f/stop. You also know that if you know the correct f/stop and shutter speed for a certain exposure, that you can quickly determine other f/stops and shutter speeds. For example shooting a gray card with a tripod mounted camera using a film with an ISO/ASA rating of 125 at high noon on a cloudless day would require an exposure of 1/125 at f/16, 1/250 at f/11, 1/500 at f/8, or 1/1000 at f/5.6. Any one of those combinations would produce negatives/slides with the same graycard density. A good photographer would know when to use any one of the shutter speed f/stop combinations. A good photographer would also know when to rate the film at a higher or lower ISO/ASA value due to lighting conditions or processing variations.
With the Shopbot, there are also three factors that determine proper operation: chip load, router/spindle speed and feed rate. For example, if I wanted a chip load of 0.030 when using a single flute bit, my Porter-Cable 7518 has five different rpm/feed-rate combinations to achieve that particular chip load: 10,000 rpm at 5.0 ips, 13,000 rpm at 6.5 ips, 16,000 rpm at 8.0 ips, 19,000 rpm at 9.5 ips, or 21,000 rpm at 10.5 ips. Any one of those combinations would produce a chip load of 0.030 with that single-flute cutter. (For the sake of discussion, we'll totally ignore the router's power curve and assume that we're using a router that produces equal power at each speed setting.) The hard part for those of us who are novices at CNC routing, is to know which of those combinations is correct for the cut that we want to make. There is a simple way to find out. All that it requires is to run a series of tests at each speed. For example, starting at 10,000 rpm and 5.0 inches per second, I could make up to seven cuts in some 3/4-inch material; one cut at 0.10 inch depth, another fresh pass at 0.20 inch depth, another fresh pass at 0.30 inch, a fourth at 0.40 inch, and so on until the router started to slow down or until the maximum depth was reached. Next I could run a test at 13,000 rpm at 6.5 ips at those same depth settings. Again running the test to completion, or stopping if the router started to bog down. In the same manner, I could run a series of tests at the other rpm/feed-speed settings. At the end of the tests, I could pick any feed-speed/rpm combination that didn't bog down the router at the depth that I needed to cut.
However, if all tests passed without bogging down the router, which feed-speed/rpm combination would I choose? That would depend on the part that I needed to cut. For example, if I needed to cut small squares, 3-inches per side, where ramping would be a big factor, I would need to run an actual test at each feed-speed to see what the actual feed-speed with ramping included actually was. For that type of test, I would create a file without any z-axis moves or any jogs and without a speed setting (no MS,x-axis,z-axis command), manually enter the five ideal feed-speeds at the console, one for each test, air cut the file and then look at the logs to get the running times, and then divide the times by 12 (3+3+3+3) to get the actual inch-per-second produced. If, because of ramping, the fastest ips setting obtained was 6.5, even though I had asked for 10.5 ips, I would know that 13,000 rpm would probably be the correct router speed with a feed-speed of 10.5 (derated by the controller to 6.5 ips due to ramping) to get the desired chip load. Making a few test cuts at 13,000 rpm and 10.5 ips (derated by the controller to 6.5 ips because of ramping) would verify the test.
If the part to be cut was a cabinet side with long runs where ramping would only be a minor part of each cut, the fastest rpm/feed-speed combination could probably be used.
Of course, as experience was gained, I could use a different chip load setting, thus modifying the router's rpm or feed-speed to get an even better result, just as a professional photographer would alter the ISO/ASA rating of a film to tweak things based on his experience.
The last few days have been an awakening as the Alpha has cut thousands of 2x6 inch and 2x10 inch pockets 0.40 inch deep in popular and baltic birch. The first cut up the center of each pocket, using a 3/8-inch spiral, 2-flute bit, at 0.40 inch depth required a slow 1.25-ips to keep the router from bogging down too much in most of the popular, but the rest of the passes in each pocket were cut at 12-ips. The pockets were cut in half the time that it required before testing and tweaking. The quality of the cut was much better, and the same 3/8-inch bit was used without frying or premature dulling.
The sound of good quality work being produced efficiently is sweet music indeed.
Mike