How Much Post-Proccessing Do You Do?

[jemelby]

<preface>
Many fo the forums regulars have responded to my niggling questions over the past week or so, I greatly appreciate their contributions. I have found answers about the mechanics of SB, software, construction and more. I have also recieved my SB Promotional CD which has some really great video. Thanks all.
</preface>

As I continue to research CNC and the Shopbot, I am also formulating the basis for a business plan. Production costs, material, marketing, finacing... ACK! All neccessary steps. One question that keeps troubling me is production time. As a woodworker, I may spend 10-15 hours building and assembling a piece of furnature, and another 10-15 hours sanding and finishing. Of course, this would never do in a production shop. Time and turn-around are what is drawing me to the SB.

For today's example, I will use an order for a sign as the example. Once I get an order, I would like to deliver the product as quickly as possible. I am quite familiar with the pre-production times (design, material selection, etc...) and I have an fair understaning of production times. What I would like to know is once the sign is cut, how much time is spent cleaning it up? For starters, how clean is the cut? If you use a V-bit exclusivly, I would imagine that you'de get a bottom similar to "A". Do you switch bits, or use a second spindle for different parts of the cut like "B?" If you had to switch bits, that would basicaly mean running the job twice. Plus, you'de have to re-zero everthing (I think). Or do you just go with "A", and spend the additional time cleaning it up with chisels and planes? What other time-consuming processes have I not considered?


7592

As aways, the insite provided fro this forum will be invaluable as I contemplate my post-military career options.

[jemelby]

After watching the final .ppt on the SB CD, I sort of know part of the answer, now. There was an example given where the "job was done in three passes", 1/4", 1/8" and finally 1/16". So it would seem that bit swaps are standard practice.

This presents an additional question. How do you keep everything calibrated to the original zero between passes? A rubber stamp comes to mind. Try as you might, you can never re-stamp the same spot twice. There will always be a little bit of under/over lap.

I presume that experience, skill have a fair amount to do with it.

[andrewm]

The zero plate for the Shopbot is a must have not realy an option. It is a metal plate with a wire that you set on you table under the bit, or on the surface of your workpeice, depending on where you want zero. You then run a routine which moves the tool down till it contacts the metal plate, which acts as a sensor, and stops. The bot records this point and calculates the zero point. If done right, it is VERY accurate. All in all, swapping a bit and zeroing only takes a few minutes.

[jemelby]

Thank You Andrew. I understand, now.

To expaciate on the concept of multiple passes:

If you were to use a V-bit for the shaping, and a 1/4" spiral for the clear-out, would you require two seperate cut files? A review of the documentation tells me that you can specify the type of bit you are using. The software cuts a path dependant on this. But I wonder. Does it re-compute for width and cut an appropriate path, and is it reliable/practical? It seems that every other word spoken about the SB is either "accurate" or "fast", but I am thinking about the automatic depth index on my surface planer. I never use it because it's idea of 3/4" is different from the rest of the world's (I guess I am ever the sceptic)

p.s. Andrew, since your so close to Atlanta, perhaps you know something about IWF '05. The web page isn't up yet, and I was hoping to get to it be for the next deployment cycle. I was going to ask my wife's uncle in Roswell, but he is a school principal who golfs, and wouldn't know where to start.

[henrik_o]

quote:All in all, swapping a bit and zeroing only takes a few minutes.

What does that mean, exactly? And, more importantly, how does this work out in the production flow?

For low volume/high value production, it isn't that big a deal, obviously. But I could see it becoming somewhat of a problem for high volume/low value runs: you'd need one man to tend to the machine pretty much all the time if a two minute bit change procedure is needed every ten minutes.

How do you 'botters that run a hivol/loval operation deal with this?

[richards]

Bit changes are a minor nuisance. Most of the cabinets that I cut have 5mm shelf-pin holes, 8mm system holes, and perimeter cuts. The 5mm holes are drilled with a 5mm bit. The 8mm holes are sometimes cut as circular pockets with the 5mm bit and sometimes drilled with a 5/16-inch bit. The perimeter cuts are almost always cut with a 3/8-inch bit.

The procedure that I use is to make a separate cut file in PartWizard for each cutter. The cut files are slightly modified with a header and footer section to turn the router on/off, to ramp into cuts, and to move the router to -18,5 at the end of each cut for easy bit swapping. (I use my 120-60 Alpha as a 96-60 machine which leaves plenty of room to move the router away from the cutting area.) Swapping and re-zeroing the Z-axis takes about 2-3 minutes.

Obviously, having a tool changer or even multiple heads would be better, but a tool changer is pricey (I was quoted $18,000), and having two heads is great - unless you need three or four or five. For me, changing bits is the best compromise.

Returning to the same 0,0 point each time a new file is run has not been a problem. Assuming that move speeds are appropriate so that the steppers don't lose steps, repeatabily is excellent.

[Brady Watson]

James & Henrik,
If you anticipate changing bits that often, then you might want to consider 1 of 2 things. 1. Get a SB with 2 Z-axes and spindles. Each one can hold a different bit. 2. Buy a SB with a Colombo toolchanger on it.

Both of these options costs money...BUT if you are doing the kind of production that you speak of, then cost is pretty much a moot point.

In the real world, (void of number crunching and optimistic engineers) not many people scramble around trying to change bits as fast as possible. Who wants to work that way??? To me, changing out bits isn't a big operation. The time you spend loading, unloading and cleaning off the spoilboard is most likely longer than the time it takes to change bits. If you can get more efficient at load/unload, then you just saved time...Even an automatic toolchanger takes time to power down the spindle, position over to drop current tool, reposition to new tool, load, lock and bring spindle back up to speed...then move over to the cutting surface.

-Brady

[mikejohn]

Mike
A 5/16-inch bit is a fraction over 7.9mm.
Which gives you the more accurate hole, pocketing the 5mm, or drilling with the 5/16-inch?
......Mike

[bill.young]

Sometimes you can split the job up, so that you cut a whole stack of parts (or sheets) with one bit and then put them back on the table and cut the same stack with the second bit. It doesn't work for all jobs and you need a good way to index the parts so that they go back on the table in the same place, but it can save a lot of bit changing.

Bill

[gerald_d]

Any particular reason that this thread title has been made bold?

[henrik_o]

Mike R,

quote:Bit changes are a minor nuisance. Most of the cabinets that I cut have 5mm shelf-pin holes, 8mm system holes, and perimeter cuts. The 5mm holes are drilled with a 5mm bit. The 8mm holes are sometimes cut as circular pockets with the 5mm bit and sometimes drilled with a 5/16-inch bit. The perimeter cuts are almost always cut with a 3/8-inch bit.
Ah, that sounds rather like system 32 frameless cabinets, which just happens to be the main reason why I'm looking into cnc. I have so many questions, but I'll limit myself to a few, here and now;

With up to three toolchanges per sheet, in the cycle time for one sheet (let's assume standard size kitchen components) how much does the stop program, home, change tool, zero, start program (is that the approximate layout of the operation?)procedure account for? Roughly speaking?

Do you use dowel assembly? If so, how do you drill the end holes on the parts that will line up horizontally in assembly? This is one of the main things I have yet to learn, cnc machining on-sheet joinery techniques.

Is the perimeter cut clean enough to take immediately to edge-banding, or does it need post-processing? Does it depend on material used? When you have edges not suitable for automatic edgebanding, is it possible/reasonable to do on-sheet "guide cuts" (that correspond to a lengthwise tenon on the edgebanding material)?

quote:The procedure that I use is to make a separate cut file in PartWizard for each cutter.

That sounds reasonable, but is it by choice or forced? Is it possible to program several bit changes into one cut file? Does your approach involve duplication of effort in the design stage, or is it more or less straightforward anyway?

quote:Obviously, having a tool changer or even multiple heads would be better, but a tool changer is pricey (I was quoted $18,000), and having two heads is great - unless you need three or four or five. For me, changing bits is the best compromise.
Yeah, it depends on the application I guess. I wonder where the threshold is, though, what with the alpha moving into speed territory that (to my, admittedly, limited understanding) makes them at least possible to consider for industrial volumes.

I can't see a current application where the cost of a toolchanger would be justified for my operation, with the possible exception of making basic kitchen cabinet doors (ack, that's probably not the right word in English, educate me) where I'd be looking at quite high volumes, but a very pitiful return per part. In such a scenario, multi-head or toolchanger might be the only way to do it economically.

Anyway, I think I've been enough of a nuisance (
) and would just like to say I'm terribly thankful you took the time to reply and give me a lot of things to think about.

Also, I'm very sorry that I've basically barged in here and tried to hijack your thread, James. Much obliged for you starting it off on such a poignant note.

[jemelby]

Hendrik;

You've not barged in at all. I will anxiously await someone's answer.

I'm especially curious for an answer to your question about "banding ready."

[elcruisr]

Henrik,
edgebanding straight from the machine is entirely possible with the right cutting tools. We do it all the time!. We are currently running an RTA job that requires edgebanding and several different hole sizes per sheet. We can manage this with one or two bit changes and doing some holes as a drilling routine and some as an area clearance. A manual bit change is really not so bad as long as the operator dosn't mess about at the job.

As for industrial volumes we will have cut close to 80,000 parts this year, is that enough to qualify? We just bid a job several weeks ago for 400,000 parts. That one would have required us purchasing a second Shopbot that would have run for just under 4 months straight to fill the order but would have still been profitable with the purchase. Looks like the plastics distributor lost the bid at his end though....

What we are getting interested in is an air drill to handle our most common hole drilling tasks.

Eric

[ckurak]

James,

This is a little off topic, but since you asked...

IWF (International Woodworking Machinery & Furniture Supply Fair) is held every two years in Atlanta. Since there was one in August 2004, the next one should be in 2006.

For a list of woodworking trade shows, go to http://www.iswonline.com/
Then, click on "Coming Events" where you will find a listing of LOTS of industry trade shows.

The next large industrial woodworking show that I know of is in February in Greensboro, NC. Next year's BIG show is in Las Vegas in July. Check out the iswonline website for more details.

[richards]

Mike J.

5mm vs 5/16-inch bits - The 8mm holes are for 8mm dowels. The dowels that I buy are usually a little under-sized and even then they vary enough in diameter that using a 5/16-inch bit works great. With a 5/16 bit, I drill down about 1/4-inch, retract to clear debris and then drill some more. The problem with my 5mm bits is that they have a very small cut lenght, meaning that the holes have to be 0.4 inch or less in depth. The 5mm bits also take longer to machine the holes since they are drilling as well as pocketing. The choice of which to use is based on the depth of the hole and the number of 8mm holes. When I need the "perfect" 8mm hole, I usually use a 1/4-inch spiral bit and pocket the hole.

Henrik,

Bit change time - The Alpha jogs at 30 inches per second, so moving the carriage on/off the table only takes 2-6 seconds. In any case, I usually figure 3 minutes maximum per bit change, including jog time, z-zero time, program load, etc. If I'm really pressed for time, I can cut the changeover to less than 2 minutes.

Dowel assembly - I do use dowel assembly. The first tool that I bought after assembling the Shopbot was a Delta horizontal boring machine (approx. $2,000). Boring the holes in the edge of the material takes longer than cutting out the parts with the Shopbot. If you have/buy a manual horizontal boring machine, you'll need to devise a table to support large panels and jigs/stops to position the panels. Marking each panel for drilling will take too much time.

Edge banding - Usually the final pass on all panels is a very light 0.02 - 0.04 inch cut, which leaves a good clean edge in particle board coated melamine. With MDF and plywood, I give the edge a light sanding using 100-120 grit paper. My edge bander is the real high-tech Freud ($200) bench top machine. Don't laugh, it works quite well. (As a rule, I don't edge band or finish the parts that I cut. I leave that for the customer to do. My focus is designing and cutting parts.)

PartWizard - Is it by choice or force? PartWizard is the only tool path program than I have. (I use AutoCAD Lite to design the parts.) PartWizard works great for me, but I've programmed computers since 1975. So, tweaking the code to do what I want done isn't terribly hard. To keep things simple, I like my cut files to be realatively small. Sometimes I include the FP command to chain several small programs into one larger file. HOWEVER, I NEVER include pauses for bit changes in a file. I don't want the router to start up because I bumped the keyboard - fingers take forever to grow back.

Speed and efficiency - Except for the first sheet of the day, I rarely have to wait for the Shopbot. Most of the time, I'm still fiddling with the previous sheet in some way when the Shopbot finishes the current sheet of material.

As a final note. If you're going to design, cut and assemble cabinets, the Shopbot will probably be the least expensive piece of equipment in your shop. Your production quality edge-bander will cost much more than the Shopbot. An automated horizontal boring machine will cost much more, and so forth. In fact, a spindle and a good vacuum hold-down system will cost at least as much as the basic Shopbot.

[henrik_o]

Brady W,

Thanks for the input. Yes, I have looked into the 2nd z-axis option. I guess I'm still trying to get an idea of what operating the machine actually entails, practically, and how cost-efficient certain alternatives are etc.

quote:In the real world, (void of number crunching and optimistic engineers) not many people scramble around trying to change bits as fast as possible. Who wants to work that way??? To me, changing out bits isn't a big operation. The time you spend loading, unloading and cleaning off the spoilboard is most likely longer than the time it takes to change bits. If you can get more efficient at load/unload, then you just saved time... [etc]

Excellent points. Have you veteran guys had a thread where you describe your operation flow in any detail? I mean, the kind of hands-on info, like cleaning the spoilboard (which I hadn't even thought about) and so on and so forth? How interacting with the machine actually works out?

Just so you don't get me wrong, I'm not here to hunt for seconds here, a minute there. Rather, what I'm interested in is learning the capabilities of the machine and how I will interact with it, practically, once I get one (fingers crossed). And as for the time aspect (which is more about learning about it than having demands, or anything) it is not about a tool change taking two minutes, or two and a half minutes -- it's more about how it impacts work flow. If direct interaction with the machine during continuous production will take, say, fifteen minutes per hour, then ideally (fat chance, I know, but ideally) I'd want that to be one fifteen minute block every fortyfive minutes -- not five three minute blocks, one every nine minutes of the hour.

[henrik_o]

Bill Y,

quote:Sometimes you can split the job up, so that you cut a whole stack of parts (or sheets) with one bit and then put them back on the table and cut the same stack with the second bit. It doesn't work for all jobs and you need a good way to index the parts so that they go back on the table in the same place, but it can save a lot of bit changing.
Interesting. What are the limitations of this strategy, in your opinion?

Another related question this raises is how to deal with parts where both sides (up & down) need to be machined, is simply flipping the workpiece fairly straightforward, or are there catches to this approach? (I'd assume hold-down, and accuracy over a whole sheet would be issues, for example.)

[henrik_o]

Eric L and Mike R, you've given me quite a lot to think about, and I thank you for that, but I'm pressed for time right now so I'll defer replying for a little while.

[bill.young]

Henrik,

If the final pieces are basically the same size and shape then you can cut the parts out with your first bit and then build a jig to keep them held in place and indexed to do the cuts with the rest of the bits. If the machining time for each bit is short, though, then you may not have time to do anything else before you need to change blanks, which can be inefficient if you work by yourself.

If there are lots of small and odd-sized parts then you'll probably need to keep them in full sheets until the last bit cuts them out. That when this methods loses it's appeal for me...just moving the sheets around and getting them back in the same place can be a hassle. The less you handle material the better...some mdf sheets weigh 100+ lbs.

I use the first method when I make sets of puzzle letters. You can see a few pictures of the process the bottom of http://seaside5592.blogspot.com/

Bill

[henrik_o]

Bill Y,

Thanks for the info, it sounds interesting. I hope to be able to bring this up again when I can be a bit more specific about what it is that I want to be able to do.

The puzzles look great (and the boxes too)!

I have bookmarked your blog, looks (and reads) fabulous.

[henrik_o]

Eric L,

quote:edgebanding straight from the machine is entirely possible with the right cutting tools. [...] A manual bit change is really not so bad as long as the operator dosn't mess about at the job.
That's what I want to hear!


quote:As for industrial volumes we will have cut close to 80,000 parts this year, is that enough to qualify?
Certainly. That sounds like quite a lot on one 'bot.

Do you have your own product line, or do you mainly produce parts? What is your idea of profitability regarding different production scenarios? I realize that's a terribly vague question to ask, but if you could give me some hint it would be much appreciated.

[gerald_d]

Eric, mentioning those volumes, have you worn out your rails, racks or pinions yet? We can see noticeable wear/deformation of the rails, but not so much that we are losing sleep over it. (the hard v-rollers are "smearing" the softer angle-iron rails - the alpha's bent plate rails would be even softer)

[henrik_o]

Mike R,

quote:Bit change time - The Alpha jogs at 30 inches per second, so moving the carriage on/off the table only takes 2-6 seconds. In any case, I usually figure 3 minutes maximum per bit change, including jog time, z-zero time, program load, etc. If I'm really pressed for time, I can cut the changeover to less than 2 minutes.
Ok, so it's nailed to the two to three minute range. Good to know. Two things, though.

First, in your typical operation, how much of the cycle time per sheet does tool change account for? A couple of three minute changes for a sheet that cycles in half an hour or more isn't much to quarrel about, I guess, but if the cycle time is down to fifteen minutes, we're suddenly looking at a substantial part of that time. I don't yet have the understanding of what cycle times would look like on the 'bot, but once I do get an appreciation, then this would be crucial to evaluating the return on getting a second z-axis or a tool changer. What's your general impression?

Second, do you feel that during production runs there are sufficient periods during which the 'bot cuts merrily away, or do you or a partner/employee tend to be kept busy by it all the time? I have one line of cabinets that I'd hope to expand by quite a margin with cnc equipment, but experience has shown that this line requires two people to handle for several steps during post-processing. I'd like to work out a work flow in which tending the 'bot could be done by one guy during the stages where only one man is needed in post-proc. If one person is needed to tend to the bot every seven or eight minutes, then that will create bumps in the flow, and I will "lose" productivity not just for the handler tending the 'bot, but also for the second guy waiting for the first to return.

Of course, this is all <i>very</i> theoretical, and I may be chasing the wrong ghosts, but hey, I have to start somewhere.

Any input you'd have on this would be very welcome.

quote:Dowel assembly - I do use dowel assembly. The first tool that I bought after assembling the Shopbot was a Delta horizontal boring machine (approx. $2,000). Boring the holes in the edge of the material takes longer than cutting out the parts with the Shopbot. If you have/buy a manual horizontal boring machine, you'll need to devise a table to support large panels and jigs/stops to position the panels. Marking each panel for drilling will take too much time.
I have my eyes on a reasonably priced multi-head horizontal boring machine, but I'd really, really like it if it was possible to do the joinery machining on-sheet. Would it be possible to cut horizontally for size-0 or size-10 (that's what we use, lamello, maybe it's different on the other side of the pond) biscuits using the 'bot? I don't know if there are bits available that can do this, but that would be totally sweet.

Where can I learn about cnc on-sheet joinery techniques? Anyone have a good idea or link handy?

quote:Edge banding - Usually the final pass on all panels is a very light 0.02 - 0.04 inch cut, which leaves a good clean edge in particle board coated melamine. With MDF and plywood, I give the edge a light sanding using 100-120 grit paper. My edge bander is the real high-tech Freud ($200) bench top machine. Don't laugh, it works quite well. (As a rule, I don't edge band or finish the parts that I cut. I leave that for the customer to do. My focus is designing and cutting parts.)
Hey, if it works well for your operation, then all's good. I fully admit that looking at Holz-Her edgebanding centers makes me quite pale (and my wallet suicidal), damn, those things cost an arm and a leg. Anyway, a light sanding sounds quite ok -- as long as we don't have to actually remove measurable material to get a good clean edge, then it's allright.

quote:PartWizard - Is it by choice or force? PartWizard is the only tool path program than I have. (I use AutoCAD Lite to design the parts.) PartWizard works great for me, but I've programmed computers since 1975. So, tweaking the code to do what I want done isn't terribly hard. To keep things simple, I like my cut files to be realatively small. Sometimes I include the FP command to chain several small programs into one larger file. HOWEVER, I NEVER include pauses for bit changes in a file. I don't want the router to start up because I bumped the keyboard - fingers take forever to grow back.
They have that annoying tendency, ya. Good thinking. I have some programming experience, so I don't feel too daunted by that, though a new environment is always a new environment and all that. I expect to cut a lot of air in the beginning, heh. I assume tool change in one file is not possible/reasonable in partwizard, then?

quote:Speed and efficiency - Except for the first sheet of the day, I rarely have to wait for the Shopbot. Most of the time, I'm still fiddling with the previous sheet in some way when the Shopbot finishes the current sheet of material.
Hm, yeah, that goes some way to explaining what I was wondering about cycle times and work flow.

Do you run a one-man operation? Two-man?

quote:As a final note. If you're going to design, cut and assemble cabinets, the Shopbot will probably be the least expensive piece of equipment in your shop. Your production quality edge-bander will cost much more than the Shopbot. An automated horizontal boring machine will cost much more, and so forth.
Yes, good observation. However, it does seem to me that there's a pretty big second hand market for these machines over here, and good quality can be had for a reasonable price (I'm currently looking at a used but top notch $10k Holz-Her edgebander, for example, having ruled out the new offerings). The second hand market for cnc machines is pretty much nil, though, and the asked prices for what does exist are completely unreasonable compared to non-cnc equipment. I'm not sure I'd dare buy a used cnc machine, either.

The way I'm looking at it, the 'bot at its present price could be a great deal, but I need to be sure it can do what I need it to do and that there are no hidden catches to operating it. I'm not asking for an automagical problem solver that can run my whole shop and make me my coffee as it polishes my shoes, but it must fit in with what we do and do it well. If it reduces some manual work but adds some instead, then I need to know that so I can evaluate and plan for it.

Sorry to go off on a tangent here, I guess I'm still formulating most of this stuff in my head


In closing, thanks a million for taking the time to answering, this is such a great resource made possible by such great people. Without this forum, I'd have tallied off the 'bot as a clever but probably under-delivering hobbyist machine. Shopbot Inc. couldn't buy this kind of advertising with money. It really is that great. I just hope you regulars fully realize that, what this means to newbies like me

[henrik_o]

Also, I'd like to re-iterate my Q about "on-sheet 'guide cuts' (that correspond to a lengthwise tenon on the edgebanding material)", I probably did not phrase that very well.

Inspired by James' snazzy graphics, here's a pic of what I meant;


7593

I first encountered this technique when doing complex (curved) edges. Through experience, we've learnt that this really really helps when doing manual edgebanding. We can put four 2440mm rips into one glue assembly jig at a time, operating two such jigs simultaneously. I think this procedure gives an industrial grade edgebander quite a run for its money, at least in my operation.

If I could get good accuracy (+/- 0.25mm over 2440mm, approx.) doing the sheet cuts on the 'bot, that would be quite something. The depth of cut would need to be 2-5mm, depending on the application.

[billp]

Henrik,
There ARE a number of router bits used in place of a biscuit (Lamello) machine. They usually have a 3-4 winged cutter, and a changeable bearing to determine how deep your depth of cut will be, so you can adjust the machine to a 0,10, 20, etc. I believe you could easily adapt one to do either biscuit holes or full length splines/mortices depending on the thickness of your material ( I forget just how much of the bit is hanging below the actual cutter). Instead of relying on the bearing you would probably be programming the actual path of the tool so you don't create too much side force on your router's bearings...In this case instead of "zeroing" your bit with the zero plate/routine I'd guess you would have to eyeball the proper depth to position the cutter properly, and determine your bit positioning...

[richards]

Henrik,

Cycle time per sheet depends entirely on the complexity of the cuts. Most of my cabinet designs take 5-10 minutes per sheet for the perimeter cuts and 3-10 minutes per sheet for the holes. Perimeter cuts are usually 3.5-inches per second (limited by my Porter-Cable router) at -0.65 depth. Finish cuts (0.020 - 0.040) are usually 12-ips at -0.77 depth. Z-Axis drilling speeds for 5mm bits is 1-inch per second move speed and 6-inch per second jog speed. I jog down to 0.10, drill to -0.25, jog the bit up 0.10 to clear debris, jog the bit down to -0.20, drill to -0.45 and then jog up to 0.30. Jog speed between holes is 30-inches per second. Removing the cut sheet, vacuuming the table, and reloading a fresh sheet takes about 5 minutes. That time could be cut drastically with a helper. My rule of thumb is to plan on cutting 3-4 sheets per hour for ordinary cabinet work. Those rates could probably be doubled if I used a spindle, a 15+ hp vacuum, rollers to load a spoil board and material to be cut, rollers to unload the spoil board with material, etc. If maximum efficiency is needed, the handling of the material should be planned as carefully as the actual cutting of the material.

Using a slot cutter to groove the edges of the material would be a good alternative to doweling. You might have to raise the material on pods to give sufficient clearance for the nut and washers on the slot cutter. (I would think that you could cut a lot of biscuit slots in the three minutes it might take you to change bits on the Shopbot.)

[beacon14]

Henrik

PartWizard will not allow you to save multiple toolpaths in the same file unless they use the same bit. However, there's nothing stopping you from either cutting and pasting in edit mode to creat one long file, or (more common, I think) to call another file from within a file.

As for not wanting to pause a program for a tool change for safety reasons, what about a short piece of code that would require you to enter, say, 1 then 2 then 3 to continue. I think that would bring the chances of an accidental "bump" causing a premature start to something on the order of being struck by lightning. But once the bit is changed, it would be quick and easy to continue the file.

Finally, I don't have much faith in using biscuits as the primary joinery method for building cabinets. First, they are time consuming to glue and assemble, and require clamping. Most importantly, I don't believe they are as strong as the manufacturers would have you believe for butt-joining sheet material. I would strongly recommend considering other options before settling on that one.

[elcruisr]

Henrik wrote

"Do you have your own product line, or do you mainly produce parts? What is your idea of profitability regarding different production scenarios? I realize that's a terribly vague question to ask, but if you could give me some hint it would be much appreciated."

We started out with our own product line of niche market commercial furniture. But after 20+ years in the wood business I've become a "mercenary" woodworker. If it's profitable I'm interested. I'd say the Shopbot shaved 25% of the labor time out of that line. In the last year we've shaved another 25% of the time with various procceses to improve work flow and assembly time. We also have expanded our cutting services for outside shops to a whopping 60% of our annual business.

Make sure you look very hard at the software end of this business. I could do some of what I do with Part Wizard but it would take me longer to program it. I also feel the need to be able to create, size, and position tabbing precisely for tightly nested parts. Ramped entries and lead in moves are also vital to many of our proccesses. For us this meant an upgrade and we chose ArtCAM Insignia. It works great for us but is far from the only choice. It would be good to get a handle on things with Part Wizard but be aware you may or may not need to upgrade depending on your application. I just have to generate complex toolpaths FAST as we are often handed emergency parts runs from other shops!

As for profitability of different scenarios that can vary widely with the shop and the product. It has increased our profitability quite a bit.


Gerald_D wrote

"Eric, mentioning those volumes, have you worn out your rails, racks or pinions yet? We can see noticeable wear/deformation of the rails, but not so much that we are losing sleep over it. (the hard v-rollers are "smearing" the softer angle-iron rails - the alpha's bent plate rails would be even softer)"

We saw some of the same "smearing" early on but it seems to have leveled off as the shape of the top of the rail "broke it" to the V-roller. No problems with it at all. Racks are holding up fine. Y axis pinions seem to be good for about 4 months at this level and the x axis for six months but we just replaced them together and consider it a maintenance item. We have never replaced a z axis pinion. You also need to keep things greased, we use high pressure gear grease every 30 days.

Eric

[henrik_o]

Bill P,

quote:There ARE a number of router bits used in place of a biscuit (Lamello) machine. They usually have a 3-4 winged cutter, and a changeable bearing to determine how deep your depth of cut will be, so you can adjust the machine to a 0,10, 20, etc.
Is this along the lines of what you're thinking about?:


7594

Onsrud has a series, the 30-00 (page 21 in catalog), using the 30-01 arbor, which is unfortunately not pictured.

And hey, looking around a bit, I realize the kind of edge assembly I meant is called a "t-mold" in English, right?. Woodworkers vocabulary (English)++


quote:In this case instead of "zeroing" your bit with the zero plate/routine I'd guess you would have to eyeball the proper depth to position the cutter properly, and determine your bit positioning...
You don't think there would be a way to do that programmatically (still using the zero plate)? If not in the sb-supplied software, then as a self-programmed 'plugin' (or sw post-processor or whatever the preferred term is).

Hm, I'm not sure I understand the whole zeroing routine anyway. Need to check up on that.

(By the way, as a total aside, does the sb have an open plugin architecture or something similar to that?)

Anyway, thanks for taking the time to fill me up a little on this and point me in the right direction. I hope we can re-visit this issue in more depth as my edification develops.

[henrik_o]

Mike R,

quote:Cycle time per sheet depends entirely on the complexity of the cuts. Most of my cabinet designs take 5-10 minutes per sheet for the perimeter cuts and 3-10 minutes per sheet for the holes. Perimeter cuts are usually 3.5-inches per second (limited by my Porter-Cable router) at -0.65 depth. Finish cuts (0.020 - 0.040) are usually 12-ips at -0.77 depth. Z-Axis drilling speeds for 5mm bits is 1-inch per second move speed and 6-inch per second jog speed. I jog down to 0.10, drill to -0.25, jog the bit up 0.10 to clear debris, jog the bit down to -0.20, drill to -0.45 and then jog up to 0.30. Jog speed between holes is 30-inches per second.
That's very good hands-on info. What material/sheet size/thickness do you cut for cabinets? Is your drilling routine what they call "peck drilling"?

Wow, that's pretty good speed for a machine in this price range.

quote:Removing the cut sheet, vacuuming the table, and reloading a fresh sheet takes about 5 minutes.
Again, good to know. Also, you're all gonna think I'm crazy now, but this is more speculative thinking than anything: couldn't the 'bot do the vacuuming? If I built a dust collection shoe with two 'settings', one for cut dust collection and one for sucking under the entire shoe, then after the operator collects the parts he could switch the setting, start a vaccuming program, and as he carts off the parts to the next station, the 'bot jogs over the spoil board cleaning it up.

I want the world's most cost efficient cnc and the world's most expensive hoover, damnit


quote:That time could be cut drastically with a helper. My rule of thumb is to plan on cutting 3-4 sheets per hour for ordinary cabinet work. Those rates could probably be doubled if I used a spindle, a 15+ hp vacuum, rollers to load a spoil board and material to be cut, rollers to unload the spoil board with material, etc.
All good info, you're spoiling me here


For our inhouse products, getting that kind of efficiency isn't crucial. However, it would be very interesting to be able to bid for hivol/loval cnc jobs as well, and then that kind of efficiency might be a must. I will obviously have to think more about this.

quote:If maximum efficiency is needed, the handling of the material should be planned as carefully as the actual cutting of the material.
Absolutely, with the added sentiment that this holds true for all levels of efficiency, basically. By changing shop layout and handling procedures, one can cut production times by quite a chunk, even for highly manual labour intensive production. I've seen this myself, the important thing is to identify the bottlenecks.

I know to little about the 'bot to be here yet, but I hope to get to a level where I'm confident I am not one to strain out gnats and swallow camels.

For the last paragraph of your reply, I defer a little, wanting to whip up some graphics for y'all to have a good laugh about

[richards]

Henrik,

Sheet size is usually 49x97-inch particle board melamine or 49x97-inch MDF with veneer. All sheets are 3/4-inch thick. Drilling routine is basically a peck-type cycle. (You might search the forumn for excellent advice from Eric L. about drilling, ramping, speeds and feeds in various materials.)

If I were buying a Shopbot for hivol/loval jobs, I would probably start planning the purchase in a different manner. I would contact the cutter manufacturer (Onshrud) and get their advice on the optimum cutter(s) for the type of work being planned. After learning which cutter would be absolutely the best, I would contact Shopbot or Precision Drive (Colombo) to get their recommendations on the best spindle to handle that cutter at the specified RPM and feed speed. Next, I would again contact Shopbot to get their recommendation on the best vacuum hold-down system to hold the size parts you're going to cut. (Very possibily, you might need something larger than the largest vacuum offered by Shopbot.) Software would be next. I'm happy with PartWizard and AutoCAD Lite, but many others recommend more robust software. In any case get the software that works efficiently for you (be prepared to spend the better part of $10,000, or even more for a complete package if PartWizard and a low cost drawing package won't work). Finally, design/buy a good material handling system with rollers/fork-lifts or whatever you need to get the sheets to/from the Shopbot. When you're done, you'll have a "system" that will cut the hivol/loval parts most efficiently.

The cutter you posted is exactly the type that I use on a router table for slotting. You won't need the bearing if you use the cutter with the Shopbot. The Z-zero program could easily be modified for use with the cutter (just replace the default plate thickness with a value that includes the shaft/nut/washers/etc of the cutter).

[andrewm]

Henrik,

There have been a few times I have just let the shopbot clean up after its self with the dust collection system, but in the end the quickest way to get the dust off the table is to manually do it. To actually get all the dust means you have to move the bot too slow.

I usually scoop the dust into piles with the scrap as I am taking the parts off the table. The outer edges that are square can just be scrapped across the table to pile the dust up. We cut 3/4 MDF and use down spirals to keep the dust in the grooves which helps keep the parts from moving around. When we clean the table of parts that means a lot of dust. I have found that if I scoop the dust into a pile as I am taking the parts off, I can then turn around with a dust pan and scoop up 90% of the dust with the dust pan and dump it right into a can. A quick pass with a wide floor attachement on the ShopVac then picks up the leftover. This method works a lot faster than using just a vacuum. And with MDF dust, you want to get as much of that nasty stuff as possible as quick as possible.

Andrew

[henrik_o]

David B,

Thanks for taking the time to address some of my questions.

quote:PartWizard will not allow you to save multiple toolpaths in the same file unless they use the same bit. However, there's nothing stopping you from either cutting and pasting in edit mode to creat one long file, or (more common, I think) to call another file from within a file.
Good to know. I assume that with higher-end design packages, more options become available?

quote:Finally, I don't have much faith in using biscuits as the primary joinery method for building cabinets.
Hey, what do your customers do with their cabinets? Play rugby in them?


Seriously though, while I've taken charge of this business of mine fairly recently, in over thirty years of operation, we've not recieved one complaint on a cabinet joint; biscuit, dowel or otherwise. I don't know enough about the finer points of joinery to argue, but I trust my guys to do good joints regardless of type. We use biscuits extensively, and it just seems to work. It's not for all joints, for sure, but for our cabinets it's a good fit, if you pardon the pun.

quote:First, they are time consuming to glue and assemble, and require clamping.
I think gluing is easier with biscuits, but we do use the lamello dosicol dispenser system. Assembly, well, yeah, you're right, it's a bit more messy than dowels since one direction is free, but I think that's a fair tradeoff.

As for clamping, I'm not sure how to respond to that. You don't clamp your (glued) cabinets? Or don't you glue in the first place? Please fill me in on your procedure, I'm always eager to get new ideas!

Edit:
dosicol;


7595

[henrik_o]

Eric L,

Sorry for the belated reply. All your points are well taken and helpful.

quote:We started out with our own product line of niche market commercial furniture. But after 20+ years in the wood business I've become a "mercenary" woodworker. If it's profitable I'm interested. I'd say the Shopbot shaved 25% of the labor time out of that line. In the last year we've shaved another 25% of the time with various procceses to improve work flow and assembly time. We also have expanded our cutting services for outside shops to a whopping 60% of our annual business.

Interesting. Was it a big change for you to move towards producing parts for external contractors?

Your 60-40 outsourced v. own production seems a good balance; I’d like to achieve something like that eventually.

Anyway, I’ve asked you and others so much about what you do, maybe it’s time I said something about what I do and what I’m looking forward to, especially concerning a possible ‘bot purchase, and maybe you and others can point out to me where I’ve got my ass on backwards (of which I’m, all things considered, quite certain).

We are currently very much a custom low volume/high value shop. Me and two employees plus an apprentice.

Our first leg, which accounts for about 20-25% of revenue is high-end custom furniture and cabinets. I don’t make any money on this, the labour costs are too high, but it’s terrific for ‘branding’ and it inspires confidence in our work; quite a lot of my sales in other legs have been closed right after I and the customer have had a tour of the showroom. This chair is a good example of this segment:


7596

Our second leg is in interiors, about 50%. Most of it for churches, where the new interior pieces has to conform to the existing milieux. This is often painstaking work, but it pays well.

Third leg is book storage/display solutions (we’re moving into (musical) notes storage/display as well). That’s 25-30% of revenue. Our niche is book cabinets with doors. We have an in-house ‘standard line’ and also do custom solutions. One example of the in-house line looks like this;


7597

Also, an example of custom work, a made-for-customer image of an existing room with our proposal inlaid;


7598

Now, enter the prospective shopbot.

For leg one, I don’t count on it to do anything. Of course, if it could, that’d be great, and I will experiment with extreme prejudice, but I am looking to scale this leg back to 5-10%, just enough to keep the workmanship current and add interesting items to our portfolio/showroom.

For leg two, I expect a modest productivity increase using the bot to cut basic structural parts and perhaps some décor items that we currently either don’t do or have to outsource. I guess it would very much come down to the design package/strategy I would use if deploying the ‘bot would be worthwhile or not. I am looking to scaling this leg down to 20-25%, enough to keep tabs on the market I’ve cornered in case I need to fall back on it, but being more selective about what jobs I do take on (these jobs generally pay well, but there is certainly a high return differential within the segment).

Leg three is where I’d expect the ‘bot to televise my revolution. That is, end my outsorcing (score one), increase my productivity (strike two), elevate my profit margin (strike three) and through lowering prices expand my market share (I guess I’m out, but I’ll bat that one anyway).

As I mentioned, it’s about book storage/display solutions, and we pretty much just do work that involves doors, since currently I can not compete with open shelf systems (remember, I live in the land that raised and foisted IKEA on the unsuspecting world…). With a proper production line, cnc included, I could go into the open shelf market as well –- setting my sights on semi-custom work (using standardized units but tailored to fit certain dimensions). So, what I want here is for starters a highly rational production of the basic cabinets + shelves involved. Our current products are needlessly expensive due to our non-rational production of these items.

Second, ideally, I’d want to be able to produce at least partially my doors on the ‘bot. I think I’ve got a reasonable tally on the structural parts and how to do that, but the doors are an open chapter to me. I have no idea if this could be done on the bot, I just hope so.

These three legs aside, and comments is as always very much invited and quite the point of my rantings, I obviously have a thousand and one ideas for new production and so on and so forth, but that’s stories for ‘nother night. (As I have alluded to it extensively in this thread, I’d very much look into doing cnc ‘mercenary’ work, yes, with vocabulary thanks to Eric
.)

Ah, I had meant to harp some more on this train of thought, but I hear anticipating little steps on the floor above, complete with expectant voices: I’d better go up and bring in the christmas tree.

Season’s greeting to everyone!

[beacon14]

Henrik,

Thanks for posting the photos, it's always a treat to see what our distant friends are doing. Your work shows a refinement of design and detail that is often missing, even in custom work.

Don't get me wrong, I use plenty of biscuits myself, but I have seen a few outright failures and a lot of telegraphing of the biscuit through to the surface. I use a narrow-crown staple gun to attach cabinet parts together, and follow up with 2" particleboard screws. I use two 4x8 sheets on sawhorses as my workbenches, on which I can usually assemble as many as 6 cabinets at a time. The last kitchen I built had 30 cabinets, including several angled cabinets, the upper cabinets were all 52" high; total assembly time including mouting hinge plates and drawer slides was around 13 hours. No glue or clamping. I think the weakest point in a biscuit-joined cabinet is the bottom of the upper cabinet. I haven't actually heard of anyone's stack of dishes coming crashing down onto the granite countertop, but I would have a hard time sleeping soundly knowing it could happen. Your mileage may vary. If it works for you, great.