SVFeingold
Aluminum
- Joined
- Aug 7, 2015
- Location
- Santa Clara
Hey everyone,
I'm an engineer in Atlanta, one of the founders of a very small company making tennis gadgets (shot-stats.com if you're a tennis fan!). I've recently started considering the possibility of doing our own prototyping in-house, and even our own production runs for metal parts (7075) until the demand outstrips our desire/capability to make parts. Which it hopefully will someday. I'm hoping to get some advice from you much more experienced folks as a basic sanity check. This will be a long post as I type fast and tend to be long on the details, so grab a Snickers. I do apologize for that, I know many of you hate walls of text. But anyone who makes it through gets my deepest gratitude! And if your advice saves me from a costly blunder I'll be happy to send a tool your way.
The type of advice I'm looking for is stuff like:
- Is the machine a solid choice for our needs?
- Are the options I'm considering worthwhile?
- Are my fixturing ideas optimal?
- Is my cost estimate realistic?
- Is my "basic shop setup" realistic?
- Did I miss anything?
Bit of background on me, I'm a lifelong tinkerer/mechanical engineer by training/EE by studying. I've worked closely with machinists (or been the machinist) in just about every job I've ever had. I have a decent grasp on the basics although most of my machining experience has been manual (bridgeport or engine lathe) with the exception of a waterjet and laser engraver. I pride myself on being an engineer with well-developed practical skills...and one that machinists usually don't want to injure after they receive drawings.
To cut to the chase, I'm starting to finalize a quote from Yamazen for a Speedio S700X1 or S1000X1. We pay a lot for prototyping work at the moment (last single prototype cost just over $1k, more on that later) and more importantly we have to wait for it. If you want an idea of the parts made, here are some pictures of the current prototype. Imgur gallery w/ exploded view here.
All parts are 7075-T6. Magnesium was also considered but the processing/finishing is considerably more involved and good 7075 seems to have an even higher strength to weight ratio. I think even with such small parts a larger machine is better. My reasoning is that for prototyping, we can have multiple fixtures set up to make multiple types of parts without having to re-organize anything on the table. Big plus. For production we can fit many more parts into the machine so it can run longer without someone needing to feed it every ten minutes.
That set in the pictures is five machined parts. Protolabs managed to do four of them for just under $500 (cheapest and fastest I've found after getting quotes from 6 shops), but they have no undercutting tools (what?!) and one part had a dovetail. That part was done elsewhere, again for nearly $500. This isn't sustainable if we need multiple prototypes per week or 50 for a beta test run. Not to mention that those thin plates should have been 0.75 mm (or even 0.5mm) and protolabs won't do anything less than 1mm. There is a whole conversation to be had about China (I'd really like to do as much as we reasonably can in the US, hence this post) but suffice to say we spent 4 months in China at a program specifically aimed at hardware startups and navigating the manufacturing landscape. The last quote we got in China was for 4 units (much simpler parts too, total of 3 parts per unit) was just over $1500 for 7075 (supposedly...and no temper). These parts will get a lot more complicated before they get simpler.
As far as a machine, I've essentially settled on a Brother Speedio, though I hardly call that settling. By all accounts from this and other forums it is an excellent machine with a great control, and dead nuts reliable to boot. I considered the Haas DT-1 but with options, the price difference is just not compelling enough. The upsides to the Brother are very hard to ignore. The Brother is also more than fast enough to run reasonable production - I estimate we can do at least several thousand units per year without the machine really noticing, and without straining ourselves after sorting the fixtures/programs/tooling. If we were to luck out and find an experienced CNC operator I imagine that could easily go up into the tens of thousands or even higher. Of course there are still things like the electronics, packaging, assembly, etc that will be outsourced, but the machining cost is a big chunk of the BOM. I considered buying used but quickly abandoned that idea. For one, Yamazen is 20 minutes away from us in Atlanta. So far they have been extremely responsive, helpful, and patient. I've heard great things about them and I'd like to work with them.
For another, I'm not interested in a project. When I buy tools I always aim for quality and functionality first. I want a machine that works, period. I'm not just going to be running this machine, I'm also the guy designing the parts, the guy designing the electronics and getting the PCBs made, the guy going to China to find a packaging supplier, the guy interviewing new hires, the guy overseeing the software development, etc... I think most of you can sympathize with having too many responsibilities! It's enough constantly changing tasks that IMHO without the "good tool" strategy I would be spending half of each day fixing the cheap oscilloscope, then the broken CNC controller, then the cheap air compressor, then the worn out ballscrews, and so on. It's how my father was and is, and how I grew up, but nowadays I'd rather buy the more expensive air compressor instead of the cheap one I have to fix 10 times. I know I don't have to preach to you all about the importance of quality tools.
The cost of the machine, amortized over the next few years, is easily justified by looking at what we spend on prototyping now. Even if the monthly cost is the same (or more) we gain much more flexibility, control, and most importantly SPEED. After that, being able to use it for production is just a bonus. I'm good friends with a few guys in Oakland with their own startup (met them in China) that also bought a VMC (Haas) for prototyping. Despite the cost, frustration, and learning curve said they'd do it again in a heartbeat because there is just no other way they could have prototyped as much and as fast as they did. This is my main motivation.
As far as options I'm looking at:
- 16k spindle
- Coolant through spindle
- Increased capacity ATC (21 instead of 14)
- Probing system and toolsetter (Yamazen here works more with Blum but are very amenable and I'm interested in Blum vs. Renishaw thoughts)
- Manual pulse generator (no jog wheel on the control!)
- Top cover and possibly a mist collector
- Fog-buster system
A 4th or even 5th axis (3+2 type system) is something I'm seriously considering because the idea of cutting down on setups for prototyping really appeals to me. The fewer custom fixtures and setups I need, the faster we can iterate. My main concern here is working around the loss of usable Z-axis travel. My ideal machine setup (currently) is a 4th/5th on one end of the table, then a vacuum fixture for machining the very thin parts, then your standard Kurt-type vice (or two) for general use. Being able to reach all those things on a machine with a relatively small Z travel concerns me. So a column riser...eh...good idea? I'm concerned with the lowered rigidity of the riser as well, but truthfully we will not be stressing this machine. This is where I really need the advice of someone experienced with that type of thing to guide me.
As for the vacuum fixture, this is an area where I can't find a ton of information online for our specific needs. I have no prior experience with vacuum fixtures either. It seems to me that a custom vacuum fixture would be the best way to hold those thin covers. The machining operations are pretty minimal: contour, pop a few holes, countersink. From running some calculations, it seems machining forces ought to be low enough even with barely over 1 sq. in. of vacuum area. Ideally we turn off the vacuum and pick up the finished part without needing to sand/file, or do anything beyond sandblast it and put it in a box to send to the anodizer. Is this realistic? Is there a better way?
As far as other major equipment, the Brother will be going into our new workspace which is a hair over 1100 square feet (picture here). It is an office as well as my electronics workbench as well as our prototyping center.
In addition to the machine we need:
- An air compressor. Quiet operation is an absolute requirement, and sadly these compressors are very expensive. Unless the building allows us to stick one outside we have no choice. I'm very open to suggestions, so far it seems we will need a rotary vane/screw type compressor in order to run it inside and not want to poke our eardrums out.
- A horizontal bandsaw for managing stock and cutting workpieces. Open to suggestions here, I think a small, bog-standard Jet bandsaw would suit us just fine.
- A seat of HSMWorks. I use Solidworks for everything and I think HSMWorks will be more than adequate for anything we will reasonably do.
- General things like a workbench, tool changing station, inspection equipment, shop vac, etc...
Here's what the cost breakdown looks like:
- S1000X1: Just under $100k installed with listed options. Will be financed.
- Fixtures: Allocating up to $10k to start
- Tooling: Allocating up to $10k to start
- Air compressor: Expecting around $5k
- Horizontal bandsaw: $1500
- Workbenches/cabinets/shelving/toolholding: $3k
- HSMWorks: Expecting just under $10k
So that's it! If you made it this far I am impressed and grateful. Did I miss anything? Are my costs reasonable? Am I looking at something the wrong way? Are my assumptions flawed? Please let me know! I'm an information sponge and any thoughts I can get are extremely valuable. Thank you!
I'm an engineer in Atlanta, one of the founders of a very small company making tennis gadgets (shot-stats.com if you're a tennis fan!). I've recently started considering the possibility of doing our own prototyping in-house, and even our own production runs for metal parts (7075) until the demand outstrips our desire/capability to make parts. Which it hopefully will someday. I'm hoping to get some advice from you much more experienced folks as a basic sanity check. This will be a long post as I type fast and tend to be long on the details, so grab a Snickers. I do apologize for that, I know many of you hate walls of text. But anyone who makes it through gets my deepest gratitude! And if your advice saves me from a costly blunder I'll be happy to send a tool your way.
The type of advice I'm looking for is stuff like:
- Is the machine a solid choice for our needs?
- Are the options I'm considering worthwhile?
- Are my fixturing ideas optimal?
- Is my cost estimate realistic?
- Is my "basic shop setup" realistic?
- Did I miss anything?
Bit of background on me, I'm a lifelong tinkerer/mechanical engineer by training/EE by studying. I've worked closely with machinists (or been the machinist) in just about every job I've ever had. I have a decent grasp on the basics although most of my machining experience has been manual (bridgeport or engine lathe) with the exception of a waterjet and laser engraver. I pride myself on being an engineer with well-developed practical skills...and one that machinists usually don't want to injure after they receive drawings.
To cut to the chase, I'm starting to finalize a quote from Yamazen for a Speedio S700X1 or S1000X1. We pay a lot for prototyping work at the moment (last single prototype cost just over $1k, more on that later) and more importantly we have to wait for it. If you want an idea of the parts made, here are some pictures of the current prototype. Imgur gallery w/ exploded view here.
All parts are 7075-T6. Magnesium was also considered but the processing/finishing is considerably more involved and good 7075 seems to have an even higher strength to weight ratio. I think even with such small parts a larger machine is better. My reasoning is that for prototyping, we can have multiple fixtures set up to make multiple types of parts without having to re-organize anything on the table. Big plus. For production we can fit many more parts into the machine so it can run longer without someone needing to feed it every ten minutes.
That set in the pictures is five machined parts. Protolabs managed to do four of them for just under $500 (cheapest and fastest I've found after getting quotes from 6 shops), but they have no undercutting tools (what?!) and one part had a dovetail. That part was done elsewhere, again for nearly $500. This isn't sustainable if we need multiple prototypes per week or 50 for a beta test run. Not to mention that those thin plates should have been 0.75 mm (or even 0.5mm) and protolabs won't do anything less than 1mm. There is a whole conversation to be had about China (I'd really like to do as much as we reasonably can in the US, hence this post) but suffice to say we spent 4 months in China at a program specifically aimed at hardware startups and navigating the manufacturing landscape. The last quote we got in China was for 4 units (much simpler parts too, total of 3 parts per unit) was just over $1500 for 7075 (supposedly...and no temper). These parts will get a lot more complicated before they get simpler.
As far as a machine, I've essentially settled on a Brother Speedio, though I hardly call that settling. By all accounts from this and other forums it is an excellent machine with a great control, and dead nuts reliable to boot. I considered the Haas DT-1 but with options, the price difference is just not compelling enough. The upsides to the Brother are very hard to ignore. The Brother is also more than fast enough to run reasonable production - I estimate we can do at least several thousand units per year without the machine really noticing, and without straining ourselves after sorting the fixtures/programs/tooling. If we were to luck out and find an experienced CNC operator I imagine that could easily go up into the tens of thousands or even higher. Of course there are still things like the electronics, packaging, assembly, etc that will be outsourced, but the machining cost is a big chunk of the BOM. I considered buying used but quickly abandoned that idea. For one, Yamazen is 20 minutes away from us in Atlanta. So far they have been extremely responsive, helpful, and patient. I've heard great things about them and I'd like to work with them.
For another, I'm not interested in a project. When I buy tools I always aim for quality and functionality first. I want a machine that works, period. I'm not just going to be running this machine, I'm also the guy designing the parts, the guy designing the electronics and getting the PCBs made, the guy going to China to find a packaging supplier, the guy interviewing new hires, the guy overseeing the software development, etc... I think most of you can sympathize with having too many responsibilities! It's enough constantly changing tasks that IMHO without the "good tool" strategy I would be spending half of each day fixing the cheap oscilloscope, then the broken CNC controller, then the cheap air compressor, then the worn out ballscrews, and so on. It's how my father was and is, and how I grew up, but nowadays I'd rather buy the more expensive air compressor instead of the cheap one I have to fix 10 times. I know I don't have to preach to you all about the importance of quality tools.
The cost of the machine, amortized over the next few years, is easily justified by looking at what we spend on prototyping now. Even if the monthly cost is the same (or more) we gain much more flexibility, control, and most importantly SPEED. After that, being able to use it for production is just a bonus. I'm good friends with a few guys in Oakland with their own startup (met them in China) that also bought a VMC (Haas) for prototyping. Despite the cost, frustration, and learning curve said they'd do it again in a heartbeat because there is just no other way they could have prototyped as much and as fast as they did. This is my main motivation.
As far as options I'm looking at:
- 16k spindle
- Coolant through spindle
- Increased capacity ATC (21 instead of 14)
- Probing system and toolsetter (Yamazen here works more with Blum but are very amenable and I'm interested in Blum vs. Renishaw thoughts)
- Manual pulse generator (no jog wheel on the control!)
- Top cover and possibly a mist collector
- Fog-buster system
A 4th or even 5th axis (3+2 type system) is something I'm seriously considering because the idea of cutting down on setups for prototyping really appeals to me. The fewer custom fixtures and setups I need, the faster we can iterate. My main concern here is working around the loss of usable Z-axis travel. My ideal machine setup (currently) is a 4th/5th on one end of the table, then a vacuum fixture for machining the very thin parts, then your standard Kurt-type vice (or two) for general use. Being able to reach all those things on a machine with a relatively small Z travel concerns me. So a column riser...eh...good idea? I'm concerned with the lowered rigidity of the riser as well, but truthfully we will not be stressing this machine. This is where I really need the advice of someone experienced with that type of thing to guide me.
As for the vacuum fixture, this is an area where I can't find a ton of information online for our specific needs. I have no prior experience with vacuum fixtures either. It seems to me that a custom vacuum fixture would be the best way to hold those thin covers. The machining operations are pretty minimal: contour, pop a few holes, countersink. From running some calculations, it seems machining forces ought to be low enough even with barely over 1 sq. in. of vacuum area. Ideally we turn off the vacuum and pick up the finished part without needing to sand/file, or do anything beyond sandblast it and put it in a box to send to the anodizer. Is this realistic? Is there a better way?
As far as other major equipment, the Brother will be going into our new workspace which is a hair over 1100 square feet (picture here). It is an office as well as my electronics workbench as well as our prototyping center.
In addition to the machine we need:
- An air compressor. Quiet operation is an absolute requirement, and sadly these compressors are very expensive. Unless the building allows us to stick one outside we have no choice. I'm very open to suggestions, so far it seems we will need a rotary vane/screw type compressor in order to run it inside and not want to poke our eardrums out.
- A horizontal bandsaw for managing stock and cutting workpieces. Open to suggestions here, I think a small, bog-standard Jet bandsaw would suit us just fine.
- A seat of HSMWorks. I use Solidworks for everything and I think HSMWorks will be more than adequate for anything we will reasonably do.
- General things like a workbench, tool changing station, inspection equipment, shop vac, etc...
Here's what the cost breakdown looks like:
- S1000X1: Just under $100k installed with listed options. Will be financed.
- Fixtures: Allocating up to $10k to start
- Tooling: Allocating up to $10k to start
- Air compressor: Expecting around $5k
- Horizontal bandsaw: $1500
- Workbenches/cabinets/shelving/toolholding: $3k
- HSMWorks: Expecting just under $10k
So that's it! If you made it this far I am impressed and grateful. Did I miss anything? Are my costs reasonable? Am I looking at something the wrong way? Are my assumptions flawed? Please let me know! I'm an information sponge and any thoughts I can get are extremely valuable. Thank you!