The Tuff Writer

In-House or Off-The-Shelf CNC Control?

(image from CNC CookBook by Bob Wharfield)

There seem to be two basic schools of thought to machine and control integration:

1) Build the iron and choose a control (Doosan, Samsung, Ganesh, DMG Mori).

Most manufacturers design and build the machine (often referred to as the "iron") and then partner with a control company such as Mitsubishi, FANUC or Siemens for the electrical components as well as control systems. These companies then work to customize the control to some degree to the specific machine or line of machines (i.e. specific G and M codes for the particular application). This allows them to pick the best control for the specific task and focus on building the best machine they can. Sometimes a machine may even be available with multiple versions of the control depending on where and how the machine will be used (some DMG Mori machines for instance are available with the customer's choice of Siemens, Heidenhain or FANUC).

There are some advantages here. With the machine maker focusing on the iron while the control company works on the interface, everyone is doing the job they are best at. Companies like FANUC have a well earned reputation for consistency and reliability in their control interfaces. In addition, with each company focusing on their part of the puzzle, updates and improvements to both the mechanical and electrical side often occur every few years. Since electrical components are off-the-shelf and often used across multiple machines and CNC manufacturer brands they are also usually very well thought out, serviceable with good parts availability interchangeable. In addition, since the controls are, to some degree standardized, once you learn a control it becomes much, much easier to run any other machine with the same brand of control. This means that a shop with FANUC controls will be able to choose any CNC manufacturer that uses that control and be able to significantly flatten the learning curve for running the new machines, even if they are all of different brands (though the custom G and M code implementations will mean that there will still be differences). In addition if you like a control but want to change CNC makers it's very doable. Much easier to stick with a control you know and like and pick the right iron for the job.

There are some shortcomings to this approach as well. Chief among them is that by separating the control and iron manufacturer you also split responsibility. I've heard of stories of frustration from shop owners chasing down a machine issue only to be told by the CNC builder that it's a control issue while the control company points the finger at the custom implementations and pushes the blame back on to the builder (wrenches get thrown at some point and violence ensues). Another issue is that with custom implementations of a each control there will be small idiosyncrasies between not only machines but definitely between brands of machine with the same control. By way of example, my CAM post-processor works beautifully for normal operation but does not work at all for sub-spindle pick-off and synchronization as the Lynx G and M code implementations for the newest iteration of the control (specifically the spindle synchronization and chuck interlocks) seem to be unique to this machine. The other problem I've run across is that machine manuals are somewhat inconsistent as the CNC builder does not have the depth of knowledge of the control while the control company writes manuals for the control itself and not the specific implementations. Occasionally finding the right information can become a bit tricky (finding research data for my Master's thesis seemed easier by comparison than finding certain G-code examples for my particular machine).

 

2) Build the iron and use your own control (Haas, Mazak, Hurco).

 A few companies choose to not only design and build the machine but actually create their own electrical cabinet and control interface. This is a much more complicated undertaking as now both sides of the puzzle have to be developed, integrated and tested under one roof.  The two most notable players here are Haas and Mazak although their implementations are worlds apart. While most of the electrical bits are still off-the-shelf the critical components are now designed and supplied by third-party vendors specifically for the implementation. I've had the pleasure to speak with the designed of the Haas ColdFire system and was amazed and how much work and thought was required to develop that system and be able to implement it across their entire line of CNC machines with only "minimal" customization for each machine. A benefit here is that the buck stops with the company that made the machine. Control implementations are designed for one line of machines only and tend to be extremely consistent (I can personally say that Haas does an amazing job here). In addition, there's only one place to go for support and whether the issue is mechanical or related to the electronics or the control, you can be assured that these are the guys to call. For that reason, the CNC builder in this case often has a greater depth of knowledge of any specific machine implementation and how to solve even complex problems. The manuals also tend to be much better here (again, Haas has the best manuals I've seen in the CNC world) because both sides are made in-house. Mazak does an impressive job with their control as well although they have a tiered system with the Smooth G, Smooth C and the full 3D control implementation based on the complexity and / or price of the machine. Even so, the basic philosophy and usability of the control is kept consistent. 

It's not all sunshine and daisies though on this side either. For starters, keeping everything under one roof means that the CNC builder now has to handle everything. Development time occasionally takes a bit longer (the Haas NextGen control came out on mills a few years ago and is only now becoming available on lathes). Since only one brand of machine uses the control, these systems are not always as robust (smaller sample size for testing and implementations) and in my experience seem to be a bit glitchier. In addition, all parts and components are now sourced from a single place this way. On the one hand, this can make things easier as the CNC builder is now the place to guy for mechanical, electrical and software parts and support. On the other hand, if they don't have the part available or decided to discontinue a component, you're out of luck and better hope that you can find something in the after-market or on eBay (I've heard of this being an issue with older Mazak machines). Finally, the implementations themselves may be somewhat idiosyncratic. Don't like how something is implemented in the new Haas control? Too bad because that's how it works on every Haas machine from now on. Love the Mazak iron but can't stand their control or keyboard layout? Either look elsewhere or learn to love it because that's what you get buddy. My friend Brian has a Hurco lathe and while the machine itself is decent and makes good parts, certain control features which are standard on almost every machine control on the planet (such as the ability to restart from the middle of a program) are simply not available. In addition, he's had to re-install the control software multiple times already.

Sometimes shops will stick with a control their shop is familiar with and compromise on the iron. I've visited places which are exclusively Haas or Mazak shops because they don't want to have to retrain staff on a new control and worry about changing their existing programs or having to dial in new CAM post-processors. Something definitely worth considering as so much time is spent working directly on the control.

 

Different control design philosophies make a difference too.

The control on the Mazak machines as well as their decision to implement a proprietary control language known as Mazatrol (although the machines will also process standard G-code) is an excellent example. Mazak machines have an excellent reputation for quality but the Mazak controls are also known to be... well, different. Many Mazak shops swear by them and run all Mazak machines once they become proficient with Mazatrol. On the other hand, other machinists swear at them and find the interface work-flow to be.. well, foreign (considering the control is designed and made in Japan, that makes sense). 

FANUC controlled machines are notorious for being difficult to learn with operations being fairly button press intensive. On the other, hand they're also known to be exceedingly reliable and allow for a very refined level of control (much of which is accessibly through un-lockable parameters and options as opposed through a user friendly interface).

The Haas control, especially the ColdFire and NextGen controls are legendary for their ease of use ability of operators to learn them quickly. A friend in the industry explained that as a domestic manufacturer, Haas designed the control in a very "American" fashion. It might sound weird but if you've traveled abroad you will see that cultural norms and differences are often represented in user-interfaces all over the world (inch vs metric, right-hand-drive vs left-hand-drive, etc) so this actually makes a lot of sense.

 

So which is better?

I'm not sure there's a simple answer to that. I personally am a huge fan of the Haas control (although there are a few things which I find mildly annoying about the new NextGen version) and find that their integration with the machine is nearly seamless. That said the robust nature and reliability of the FANUC control on the new Doosan is impressive. While the interface itself may not be my favorite it's obvious that the control is extremely powerful with incredibly granular (though often obfuscated and hard to find) control capability. 

Ultimately, it boils down to the maker and the implementation. More than one person has told me that Samsung is making some decent lathes but they still don't have their control implementation dialed in even though their machines share the same FANUC 0i-T control as my Doosan where that control is absolutely is rock solid. As always, support and follow through are key when making a purchase that can cost as much as a house.

Pick your poison.

 

 

 

 

 

The Doosan Decision...

After selling our beloved Haas lathe late last year, we were quickly on the hunt for a replacement. Since many of our bits and pieces (i.e. pens, swivels, etc.) are turned parts a lathe is essential for the shop. As of the end of the year we had narrowed our search down to three finalists: the Haas ST-15SY, the Doosan 220 LSYC and the Mazak QTU-200MSY.

With Haas threatening but never making good on their promise to release the new dual-spindle ST-15SY we weren't able to go with our preferred CNC manufacturer. This left us with a close decision between the Doosan and the Mazak. The Mazak has slightly better specs but also cost more to purchase as well as to tool up, though the biggest strike against it was the control. The Mazatrol Smooth G interface is a 1/3rd keyboard proprietary interface that runs Mazatrol (although it can also run G-Code) and is used only on the lower level Mazak machines. Higher level machines run the Mazak Smooth G full keyboard control which still has some quirks to it and is designed around the proprietary Mazatrol language. The Doosan uses the much more common FANUC control and this newer version has the full keyboard layout. Making the decision a little easier was not just the less expensive machine price of the Doosan but also the lower cost of tooling as well as availability of off-the-shelf post-processors. It was a tight race but we decided to go with the Doosan from Adams Machinery in Tempe Arizona.

This is happy Jack taking delivery of his shiny new Doosan lathe:

Although we purchased the machine in December, it took a good 7 weeks for it to actually show up at the shop. The machine itself was out of Doosan in California with the main holdup being the chip conveyor out of Illinois. The ATS high-pressure coolant was actually delivered earlier and sat in the garage waiting for the machine. As always, the amazing people at C&M Rigging did the delivery and rolled it into the shop for us. They are truly wonderful and I cannot recommend them enough. 

Setup and install:

Brian from Adams did the install and he's a fantastic service tech. In fact, his walk-through of the FANUC control was the primary deciding factor in our choosing the Doosan over the Mazak for purchase. Paul, the Adams sales rep., did us a solid and insisted that Steve Allnut, the Applications Engineer from Doosan California, fly out for a week of training. This was a brilliant idea because the Doosan manuals are complete shit. They are some of the worst technical manuals I have ever read in my life. Written in some Google translated nightmare of Krenglish (Korean-English) they are nearly unintelligible and downright infuriating to read. 

The Necronomicon - Klaatu barada nikto:

When Steve showed up he let me rant about the manuals for a few minutes before gently taking them from me, nodding his head and handing me a Doosan America made "Cliff's Notes" version written in proper English. He them told me not to worry, that this manual would contain all of the information I was looking for and to his credit.. he's mostly correct. Steve Allnut is not just an amazing Applications Engineer but also quite possibly the finest CNC / Machining instructor I have ever run across. Had I known that this level training would be provided with the purchase of the machine my decision making process would probably have been abbreviated by a few weeks. He adds that much value to the entire process. With 3 days of training as well as Steve's help with creating a few templates and printing out his Doosan and FANUC PDF manuals, we were getting ready to make some parts. 

The man, the myth, they legend.. Steve Allnut of Doosan:

Unfortunately, we had our first hiccup at this time.  On the first day of training the lathe decided to mark its territory by peeing hydraulic fluid all over the shop floor. This is a new machine and all of the other machines are Haas so it's expected that it might be nervous. I didn't think about it much. Steve also found that one of the tool holders was out of spec (the tool would NOT fit in at all) and needed to be warrantied out. The tool holder was replaced in a few days by Adams and Doosan. Unfortunately, the last day Steve was here the machine leaked fluid once more. Hm...

Not to be dissuaded I started setting up the first job, D22 Large Tritium Carabiner swivels. This would be a part which would take advantage of both the main and sub-spindle allowing us to drop a completed part right off the machine (I get tingly just thinking about it). 

Spindles - two is one and one is none:

After installing the Royal QC collet chucks (get some, trust me) we were ready to go, almost. Upon start-up and slow-down the sub-spindle would shriek like a scalded cat. Service came out again and it turned out that the sub-spindle drive belts were not properly tensioned at the factory. This was starting to get a little.. uh, interesting. You see, once they were tightened with the machine up and running it quickly became obvious that the sub-spindle wasn't balanced. In fact, the shimmy was shaking the entire machine so hard you could feel it in the floor across the shop. There was almost no measurable runout (.0002") at the collet face or with a gauge pin and even after removing the chuck the shimmy persisted so the problem had to reside within the spindle. 

This time even Brian couldn't get everything running properly so Doosan called out the "big guns". They sent in the service tech the other service techs call, "The Iceman". Ok, maybe not quite but they did fly out a specialist from the east coast to work on the machine. It took a day but by the time Mohhammed and Brian left the machine was running beautifully (smooth, like butter). Doosan even sent the regional Sales Manager, Eugene, to verify that everything was taken care of properly. 

To their credit (and primarily due to the efforts of Paul from Adams Machinery) Doosan / Adams not only fixed the machine but also extended the warranty to make sure that any similar issues are covered. I can honestly say that both Paul and Brian are stand-up guys and would absolutely recommend buying a machine from these guys. They're a great team and not only stand behind their products but support their customers and actually do what they say they are going to do.

CNC machines are complex animals. They are intricately designed amalgamations of control systems (FANUC) and iron (Doosan) which can hold amazing tolerances and work tirelessly to make precise, repeatable parts. That said, their complex nature also means that a little setup as well as fine tuning may be necessary along the way. 

When buying an expensive machine we've learned that the equipment itself is only half the of the equation, the post-sales support is the other half. Partnering with a solid CNC manufacturer and even more importantly, a solid dealer like Adams Machinery can mean the difference between a smoothly functioning install and a service nightmare. Obviously, we're still just getting up and running on the new lathe and only time will tell how everything shakes out but for right now, we're enjoying that new-lathe-smell of our Doosan and excited about getting all tooled up and dialed in over the next few weeks. 

D22 carabiner and newly machined swivel fresh off the Doosan:

-jack

 

Shopping For A New Lathe...

With the sale of our beloved Haas ST-20SSY, it is now time to start shopping for a new lathe. Ideally, considering the fact that our pens are long, slim turned parts with deep holes, we should be looking for a Swiss machine, and we are but.. before we get there we still need a more traditional lathe capable of running larger parts. In the past year we've had to run a few parts from 1" bar stock or larger and we still do a lot of short run parts and prototyping. In terms of setup and small runs, a more traditional lathe is still something we can do without in the shop.

So what features are we looking for in our new lathe? To narrow things down a bit, let's look at price range. Total price for the machine, delivery and rigging has to come in under $150,000. That's still a lot of money so less is better. We need a dual spindle capable machine in order to be able to run parts complete (or as Jeff from Magnum likes to say, "done in one" operation). We also need a Y-axis and live-tooling. This means that we can machine off the C-axis center line with either radial or axial tool holders. In terms of tool holders, we're looking at a BMT (Base Mounted Tooling) turret. These have the advantage of greater stiffness / rigidity with the live-tooling drive and coolant runing through the turret giving the lathemore clearance. Size isn't a huge factor as our parts tend to be pretty small and even the smallest dual spindle machines will easily accommodate the largest parts we usually come up with. The control is also important although most common machines have fairly robust controls (FANUC, Mitsubishi, Siemens and Mazatrol Smooch C) and all of them will be a learning curve. Availability of a post-processor is a consideration but it's very likely that we'll have to work with a vendor (or work with our CAM provider) for a custom post. Finally, service and application support in Phoenix is a big deal. With out brand new Haas Mill we've had to rely on Ellison Machinery for warranty work a few times already and this lesson has taught me that I'd hate to have to rely on an out-of-town provider. 

The contenders:

1) Haas ST-15SY

This machine was demoed at IMTS last year. A fairly small dual spindle machine with BMT-45 turret, it uses my beloved Haas control and looks like a perfect fit. Unfortunately, while I keep pestering Haas (seriously, I've got rock star salesman Handsome Chris Hansen from Ellison Machinery on speed-dial) the machine is still not available. The original story was that it would be ready a few months after IMTS 2016 but Haas has decided that they would go with the NextGen control and do some field testing first. 

That said, we do know a few things about this machine. It will use linear guides like the ST-20SSY and not fully boxed ways (a more rigid but more expensive solution). Since rigidity, finish and tolerance were never an issue on the old Haas, this shouldn't be an issue. Linear guides are also faster and easier to service although as mentioned, they are less rigid and less tolerant of holding up to machine crashes (conversely they are much easier to replace in the field). For the relatively small parts and light cuts we do in the shop, this isn't a problem (we've been able to repeat to within .0005" on the old lathe which is well within our needs). Live tooling will most likely be handled by a 5 hp and 6,000 rpm setup like the current ST-15Y. The new BMT-45 turret will most likely be a 12 station setup and be able to support additional stations through the use of additional offsets and used of the Y-axis in multi-tool holders. The main spindle will be 20 hp which means the sub should be somewhere between 5 hp and 10 hp. Rumors also state that unlike the DS-30SSY, the new ST-15SY will be able to support both positioning and full contouring on the sub-spindle. The NextGen Haas control has been out for a few years on mills but is still waiting to be released for lathes. Most of the glitches for mills have been worked out now and it's a fairly solid control with a phenomenal user interface. Since Haas support is provided by Ellison Machinery we're fairly confident that we know the level of service we can come to expect from them. Standard features include everything from rigid tapping and thread repeat to 1 GB or program memory and the ability to run directly from the USB as well as remotely monitored via Wi-Fi. C-axis indexing on the current ST-15Y has an unimpressive resolution of .02 degrees and hopefully will be upgraded in the NextGen control. Typically every CAM company supports post-processors for Haas controls and while it may take a few months for solid posts to become available, they should be easy to come by within weeks of the machine hitting the market.

List price is estimated to be in the $125,000 range. Unfortunately, with no current release date scheduled we're stuck looking elsewhere for our lathe solution.

2) Doosan 220 LSYC

I keep joking that this is the Hyundai or Kia of machine tools but then, if you've driven a Kia recently, you may know that's not a bad thing. Doosan has been around for years and has a solid reputation for making some decent iron. Every forum and blog post I've run across mentioning Doosan lathes has been very complimentary. Everyone who has had a Haas and Doosan lathe ends up commenting positively on the Doosan (as a Haas fanboy, I'm not entirely sure how I feel about this). Adams Machinery in Tempe is the local rep and while they are clearly no Ellison Machinery, they talk a good game and have been in the business for a while and are also reps for Citizen Swiss machines (THE major player in the Swiss CNC game) which means they're no fly-by-night either.  This is a relatively small machine, still using the less rigid linear guides as opposed to fully boxes ways (essentially, a rail system for the turret) but that seems normal for machines in this size and price range (the base machine alone lists for $125,000). Since we never had any rigidity or tolerance issues on our old Haas lathe(and we run fairly small parts) I don't really consider this much of an issue. The turret is a BMT-45 like the upcoming new Haas which means that tooling is readily available and to help things along, the tool mounting uses precision ground surfaces allowing the self-aligning tool holders to be mounted without the need for additional dialing it (that's the sales pitch anyway). Another nice touch is that while the turret is only 12 stations, it has 24 indexing points (half-step indexing) meaning that when properly tooled up it is for all intents and purposes a 24 station turret (similar to the BOT / VDI combination turret on our beloved ST-20SSY). The system runs off of a FANUC 0i-T series control which is known to be an industry standard. The interface actually looks like an old Soviet era ICBM launch panel ("Mr. Bond, if you don't tell us the launch codes, we'll cut you in half with a laser.") but is known to be legendary for its reliability and granularity of control. In addition, the control already has every single option enabled (i.e. rigid taping, polygon cutting, thread repeat, conversational programming, etc.) but it does have certain goofy FANUC specific quirks such as being unable to run directly off the USB and being fairly limited in memory (although there is a cute little PCMCIA work around for this). On the plus side, C-axis indexing resolution for the machine is an impressive .001 degrees making machining of precise, complex live-tooled features possible. The FANUC control means that post processors aren't a problem and while there will be a learning curve, it's a fairly common system. The live tooling motor is only 5 hp but it does spin at up to 6,000 rpm (as opposed to the measly 3,000 rpm on our old ST-20SSY). Since we run a lot of small live tooling, this is a big improvement. The sub-spindle is 7.5 hp while the main spindle is a genuine 20 hp  and 125 ft/lb little beast. Another nice feature is the ability to upgrade the factory coolant pump to a 145 psi model straight from the factory for $1,460.

 

3) Mazak QTU200MSY

Mazak is a huge player in the CNC world and they have a LOT of different models. The QTU line is their economy series and the machines are actually made in Kentucky, not Japan (but then again, so is the Toyota Camry). Mazak is rep'ed by Magnum Precision in Phoenix and these guys are a serious player in the local machining racket. They have their own service and support staff as well as their own applications engineers. While Mazak doesn't have a great rep with small shops for support, Magnum seems to have their shit together and I have no doubt that will be able to offer world class support. The machine is still a linear guide setup with a Mazak take on the BMT turret desing. This version does require the tool holders to be dialed in during install and the tool holders are somewhat more expensive than the Doosan (although they do seem very robust). In fact, the entire machine is a bit pricier as the base model comes in at $149,000. The turret is only a 12 station setup but there is a very wide range of static and live tool holders available for the Mazak and with a little fiddling of the tool offsets, it's still possible to run dual stations at each turret index location. The machine is almost 2,000 lbs heavier than the Doosan which bodes well for overall rigidity and accuracy. It uses the proprietary Mazatrol Smooth C control (which run the same basic back-end as the Smooth G but looks like it's a much de-featured version). Still, C-axis indexing on the current QTU is listed as an incredible resolution of .0001 degrees. Mazatrol is Mazak's proprietary conversational programming language but the control supports standard G-code as well. Sadly, control options appear to be a la cart. Rigid tapping is not standard but a paid upgrade, so is polygon milling, thread repeat, etc. etc. Things can get pricey quickly although much of this can be handled in CAM anyway. Another wrinkle is that the CAM software we currently use does not have a post for the Mazak but a custom post should only cost $1,500 - $2,500. The panel itself looks very nice and modern but is not any more intuitive than the FANUC control and may in fact be a little more quirky. The abbreviated keyboard is also not a big selling point for me. The live tooling motor is 7.3 bhp and while it only spins to 4,500 rpm it does have more torque than the Doosan. There is a $2,600 option which looks to be a re-gearing that allows it to spin up to 6,000 rpm but then drops the torque to match the Doosan. Similarly, the main and sub-spindle are 20 hp and 7.5 hp respectively and Mazak has a reputation for delivering every since little horsie they promise. Mazak also has a factory upgrade available in case you want a 230 psi factory coolant pump although at $3,500 it's not cheap.

 

 

 

 

Selling Our First CNC Machine...

As a young and growing business, we've heavily investing in infrastructure here at Tuff Writer HQ. This means building out the shop, purchasing machines, support equipment and the such. In fact, we're up to three (all Haas) CNC machines right now. We're got two mills and one lathe (which is goofy since we're a pen company but we make it work).

I started will the mill because the Gateway Community College Industrial Design and  Manufacturing Technology program is primarily a mill oriented program. That said, when it was time to buy a lathe we decided to stick with Haas. Knowing that we would want to make more complicated parts we opted for the ST-20SSY model with Y-axis and live tooling. A hell of a machine, to be sure.

The learning curve on this was fairly steep for me, especially considering that I really had only run a lathe basic 2-axis in class for a few hours. Being familiar with the Haas control was a plus but there is nowhere near the volume of information online regarding turning as there on milling. In addition, the CAM software I was using (HSMWorks) happens to be very intuitive for milling but pretty much crap for turning. In addition, virtually all of of the documentation for the CAM package was, you guessed it, focused on the milling side. 

It took about a year for me to really start to feel comfortable with the lathe. Even buying tools was something that I knew nothing about. I purchased a basic tooling package and got to work. I read endlessly from catalogs and tooling web sites learning about insert types, preferred insert geometries and material specific inserts. Quite a bit of trial and error ensued. I cut a lot of ugly parts and made a lot of long, stringy chips. Finally though, I started making some decent looking parts. 

Next came threading. This was a major pain. Getting my threads to work properly was miserable and scrapped more than one insert and threading tool (here's a hint, you're going to need more than one pass to single point cut a thread). Finally even those started to come together. I got the threads to work but getting them to look clean was another matter. Cutting threads in steel turned out well but aluminum was another story. The threads would work well but they looked like shit (tiny burs front and rear). 

After talking with my buddy Dave22 it finally occurred to me to check how straight the tool holders were with a dial indicator. Things really came to light here. You see, whoever owned the machine before me had no idea how to install the tool holders and they were nowhere near straight. I dialed them in to within .0002" over a 4" gauge pin. Then I went a step further and re-calibrated the automatic tool probe. Now things really started looking good. Part dimensions no longer had to be dialed in and surface finishes started looking great. This made prototyping new pens much easier and we started to run a few small production jobs for others as well. 

What we really needed though, was to be able to make pens (including titanium pens) in a more production oriented manner. First though, we needed to be able to reliably deep drill through titanium. If you've never tried, you're in for a real humbling experience. While I love milling and even turning titanium, drilling it is another story. Everything my instructors had taught me had so far proven to be legend and hearsay and not much of it worked reliably. I got a pretty good handle on drilling 5xD (5 times the diameter of the drill) but what I really needed was to drill closer to 20xD in a material that did NOT tolerate deep drilling. 

Fortunately, with a little help from Geoff at OnlineCarbide.com and a lot of patience, we finally saw some progress. In fact, we managed to punch a .250 drill through 4.250 of Grade 5 titanium. That said, it took forever. What we really needed was a 1,000 psi high-pressure coolant system and special solid carbide coolant-thru drills. Unfortunately, this setup would cost well over $10,000 to add to the lathe. 

After thinking (agonizing) about this for a bit it became obvious that what we needed wasn't just different tooling, but a different machine. I've spent the past few months doing serious research but the first step is was to sell the ST-20SSY. Last week, we finally did just that. A company out of Florida offered us a fair price and we watched as our beloved lathe got loaded up and shipped out.

Now comes the interesting part, buying a new lathe. 

 

 

 

One Year of Keeping It Real at Tuff Writer MFG...

Holy shit, has is been a year already? Seriously!?!

For those of you who don't know, after 16 years of working at the Maricopa County Superior Court as a Database and Applications Manager, last April I quit my full time gig to work full time at Tuff Writer. Now, the fact that we went through 5 CIOs in 5 years and the latest one was / is bat-shit crazy didn't hurt but Peachy and I had talked about my leaving .gov and had come to the conclusion that it was time to shit or get off the pot. So now, one year (and some months) later, here we are.

How's it been so far, you ask? It's been a wild ride. 

Couldn't have done it without the wife though:

In the past year we've hired a few people. A few have left (and come back). We've brought our first full time employee (other than Peachy and myself) on board and are already looking at adding a second one. We've purchased our first brand new CNC machine (making this 3 total) and launched yet another Kickstarter project (as well as collaborated on 2 more).

Brand spankin' new Haas VF-2SS:

I've gone from being worried whether I could make this work full time to regretting not having jumped ship sooner. Don't get me wrong, there's still a long (L-O-N-G) way to go, but the progress we've made in the past year would have been impossible if I was still punching the clock with the county.

Peachy and I went to a cafe this morning and spent over an hour going over upcoming projects and we've already booked projects out well through October. While being busy is not the same thing as being profitable, we're fortunate to be doing alright there as well. While we've continued to reinvest proceeds back into the business (guess that private jet will have to wait) we're doing alright and the black, though by no means as well as when I was drawing that sweet, sweet government dollar (no jet for instance but I did buy myself a DJI Phantom a few months ago). 

That doesn't mean that things have been simple or easy, but they have been exciting and sure as hell educational. One of the biggest questions I need to answer is where I run a manufacturing company that makes pens or a pen company that manufactures other stuff.. or possibly we're really running two businesses at once. A little help and mentor-ship would go a long way here but the local small business accelerator program in Phoenix wasn't a great fit for us (unless you're a hair salon, real estate agent or fitness gym owner, you're going to be a little disappointed) but finding people to talk to about some of the issues we run into would be nice. Just last week at BLADE Show I ran across a few guys in the industry and we chatted for a few hours. It was quite honestly one of the best round table discussions I've ever had since being in business. I probably got more valuable feedback in those two hours than I have in weeks (in my defense, I don't get out much). I wish I knew how to find something like this regularly as I imagine all of us enjoyed the discussion and got something out of it.

Your's truly (the handsome one), Will from Tactile Turn, Brad from Southard Knives and John from Grimsmo Knives:

Where does that leave us going forward? Well.. forward I suppose. We're definitely going to keep the pens coming. We've recently added a used Epilog CO2 laser and LIT Fiber laser for part marking and engraving (as well as other things we haven't even tried yet). We've got numerous collaborations in the pipe as well as products we're machining for friends in the industry. There's at least one more Kickstarter project on the horizon for this year and a video project we're working on. 

I suppose the biggest challenge will be focusing on the right things. Now, if I only knew what those are.

It's ridiculously hard work (what is this sleep thing you speak of?) and occasionally terrifying (wadya mean the new mill needs a new spindle?!?!) while always being challenging but who knew that running a business could be this much fun?

Alright.. time to get back to work!

-jack

 

 

Keeping It Wheel at Log Corral....

Since I've gone full-time at Tuff Writer things have not slowed down. The entire "I'll have more time" illusion has been thoroughly shattered by now. The hours are long, the pay is slim (although admittedly that's because I decided to buy a s#!t load more shop equipment including an additional CNC mill, let's call that a story for a different blog post) but the work sure is interesting and there's no shortage of it.

Fortunately, my buddy Dave22 is here to keep me from melting into the shop and becoming some sort of manufacturing monk (hey, I already have the hair for it). As a hospital neurologist he works 7 days on / 7 days off. Now, I don't mean that he has 7 works days, I mean the man actually works for 7 days straight. Thus, when he has the week off he finds that he often needs to unwind from the stress of his job ("Yeah... so that headache you've been having? Right, it's a tumor. You're not going to make it."). Needless to say, his solution is to go wheeling. It usually starts with a text asking me to find a cool trail and suggesting that we meet at the trail head at 7AM on Monday. Now, we go wheeling pretty much every other week and sometimes even again on the weekends. Hell, works for me.  

Last week we hit Log Corral up north of Scottsdale past Mile Marker 212. We've been talking about making this trail for a while and I've driven part of it with Jeyr. Unfortunately (for Jeyr, anyway), the trail has a decent sized obstacle about 1/3rd of the way in and Jeyr's Taco is not really built for rock-crawling. Dave's beast of a Jeep on the other hand sure is, and mine will do in a pinch. Our friend Decho came along since we're trying to convince him to throw off the shackles that oppress his daily life, grow a pair, and buy a Jeep. 

Dave's 4DR JKU the Polar Express along with my more plebeian 2DR JK:

The trail starts right off of the HWY and in true Dave fashion, he immediately got lost. I don't mean shortly after we started, I mean immediately. Within 30 seconds he made a wrong turn, lost sight of me and was lost. Thank goodness for CBs and we met back up at the air down spot within a few minutes. 

The trail starts to get a little more rocky after crossing the dry creek but it really gets exciting just a bit further up.

Starting to get fun here:

An optional rock shelf is the first indication that things are going to get a little challenging. About 50 feet past that is a rock outcropping which requires careful wheel placement and decent clearance (rock sliders don't hurt either). The proof of this was laid out before us in the form of a broken down rock buggy. Think of this like an old skeleton near a watering hole indicating that caution may be advised. 

Rocks and buggy:

For anyone wondering where the buggy most likely went wrong, I present to you Defense Exhibit A, marked for identification:

Things got narrower but easier once we got over and past this scene. The trees made for a nice and colorful route and we rode on like this for a few miles.

Picture time:

As we closed in on the actual Log Corral the route changed in character. Things opened up and red rolling dirt lined the road with deep ruts. This made for a fun little drive and we got to flex the Jeep suspensions out like giant spiders making their way through the rolling desert sands. 

Flexing it:

Tipping it:

A short while later we made it to the actual Log Corral the trail is named after:

View of the Bartlett Lake from the corral:

We followed the dirt trail down the mountain to the beach (is in, drove like mad men who missed their chance to be rally drivers). This was fun and not nearly as bumpy as it could have been, but quite dusty. 

At the bottom we searched for a way across but could not find one. While searching Dave managed to drive too close to the water in the pea gravel and almost made it into the drink! He turned into the slide hoping not to tip but that quickly took him into the water and he found that he didn't have enough traction to make it back out. He hit the brakes and stopped but the big 4DR Jeep slowly kept sliding deeper and deeper into the water. With a bit of panic in his voice Dave yelled over the CB, "Where the hell are you? Get over here and winch me out, I'm sliding into the water man!" Considering that Dave's Jeep was sporting a brand new 3.6L engine due to having blown his previous motor in an unfortunate water crossing, I could understand his apprehension. That said, I had to make it through the same pea gravel without sliding into the water otherwise we'd both be in some sad s#!t with nobody to bail us out. I quite literally had to drive at a 45 degree able as the gravel kept trying to suck the Jeep down the hill and into the water. Fortunately, after 100 of so feet I hit terra-thank-the-f#@k-firma and was able to get across the muck.

I dug in, put some rock in front of the tires and threw Dave and Decho a life-line:

Once we got clear we looked around for a place to cross but quickly gave up and decided to head back the way we came in. The trail was fun in both directions and we got some more good photos of the broken down rock buggy on the way out.

Let's Go... Rock Climbing:

We made it off the trail by 11:00AM and hit Senior Taco for some lunch. Hell, I was even in the shop and making chips by 1PM. If you ask me, there aren't too many better ways to start off the work week.

Next time... Hackberry.

 

Tuff Writer MFG Update...

Ok... so, things have been progressing. In case anyone is interested, here's the scoop: A few months ago I finally quit my gov't job of 16 years to work full time at Tuff Writer. While I've been making and selling pens for over 8 years now, it was until now a part-time gig. At first solo then with Peachy and eventually we added a couple of employees. While we were still growing it became apparent over time that Tuff Writer would never reach its potential (whatever that may be) with us only running it part-time. Someone had to make the leap and while a nice fat government paycheck and pension are appealing, sometimes you need to step into the unknown and see what you can make with your own two hands (and a few carbide endmills). Where does that leave us? I'm glad you asked.

The shop:

 

I've spent the past few years at Gateway college in the industrial design and manufacturing technology program. The goal was to be able to do all of our own design and engineering work but it quickly grew past that. While we've always had a huge part in making all of our products (designing, finishing, assembling, etc), over the past few years we've been bringing more and more hard manufacturing in-house to the point where we decided that we should just bring (almost) everything in-house. Joe, Peachy, Savannah and I have spent the past year building out the shop and over the past six months or so we have been purchasing equipment. Not just band-saws, compressors, blasting cabinets, grinders and buffers but also CNC machines. What kind of CNC machines, you ask?

At the moment we've got a HAAS ST-20SSY lathe (Y-axis and live tooling... smooth like butter):

 

We've also got a HAAS VF-2SS mill (a typical day at the office):

 

These were a big investment (huge, actually... I can kiss my mid-life crisis Porsche goodbye) but it's something we thought we had to make happen. A few months ago I read an article in The Economist that had a fantastic quote by Elon Musk where he stated, "Companies create value by doing things that are hard." This really struck me because he's so very right. Value creation lies in doing things which are inherently difficult, in creating things which are hard to make. So many companies have limited specialties and focus on cost cutting efficiency or economies of scale, "focused core competencies" and minimizing their footprints and capital outlays. You know, business stuff. Many of these companies have lost sight of of why manufacturing actually exists (our industry happens to be an exception). That said, good companies are those who focus on doing the things that are difficult (General-Electric) while great companies are those who do things that are thought by most to be impossible (Space X). In order for us to grow and become the company we know that we can be, we needed to take these steps and do the hard things. Take risks, learn new skills, get our hands dirty and build something special, then do it again and even better than before.

So, thank you for your patience with us while we've been balls deep in this painful growth and build-out phase this past year and stick with us to see where this all goes (cuz I sure as shit want to see it myself).

Who knows... with your help we might just make something special here.

-Jack

Titanium Is The Devil...

Wow... yeah, titanium.

Back when I used to design parts but farm out all of the machining I was often frustrated by the (what I perceived to be ridiculously high) quotes I would get back. Even accounting for material cost it was still just expensive to get parts made out of Ti. Now though, I'm beginning to understand why.

Aluminum is a wonderful material. Light-weight, strong and easy to machine. Corrosion resistant and even fairly scratch resistant when Type II or Type III hard anodized. It's also cheap, readily available and easy on the tooling. In fact, solid carbide tooling running only on aluminum seems to last forever as long as you don't cook it or break it. TiAlNi coated tools last even longer. Plus you can run aluminum as fast as your spindle speed allows (just make sure to clear those chips and keep it cooled). There are more parts machined out of aluminum than all other materials combined, that's how useful the stuff is.

Brass comes next. It's not quite aluminum but it's pretty close in terms of machinability. Heavier and much more expensive, it looks wonderful (especially when patina'ed). It scratches easily but it's actually pretty strong all things considered. The same tooling that works in aluminum also works on brass. Two or three fluted endmills cut nicely and take big, fat, juicy chips. A few people I know have run it dry with air blast but I tend to use flood coolant (with everything pretty much). Again, tool life is not an issue. That said, you do have to slow the small cutters down a bit or take smaller cuts than you would in aluminum.

Now we have mild steel. 1018 is literally the benchmark for steel machinability. This stuff is not corrosion resistant and I don't have a lot of use for it outside of cutting work-holding. Much stiffer although not necessarily stronger (depending on the alloy) than aluminum (for stepped jaws or soft jaws) it's a great metal for holding onto small parts and making for rigid work-holding fixtures. It still machines very, very well but it's no longer the same "hot knife through butter" ease of cutting that you get with aluminum.

Chromoly is a bit more complicated. All flavors of chormoly are structurally strong and stiff but machining it annealed runs about half as fast as mild-steel. Supposedly it machines better using solid carbide tooling and air-blast (no coolant) but I haven't really run enough to be sure. Running it post-hardened is another story and I haven't had the pleasure yet but I'm sure it's hell on cutters and slow as molasses. This is where we start to find that the sweet spot for proper feeds and speeds becomes more critical. Everything up to mild steel is very forgiving and even if you don't have your feed rate dialed in, can be cut slow, fast or everything in between. With chromoly that's no longer the case (especially post heat-treat) and running it too hot or taking too big of a cut will destroy cutters or inserts relatively quickly. Tool life really starts to suffer here if things aren't dialed in.

Following up is stainless steel. Obviously, there are many grades of stainless but 303 is the most common (and pretty easy to machine) though not as strong or corrosion resistant as say 420, which is often used for knife blades, or a host of others like 316 and 416 which are often found in watches. Some are heat treatable, some are not. Many work harden which makes machining a bit of a challenge as taking a too-small cut will actually be counter productive and quickly dull your cutter. The sweet spot for feeds and speeds narrows considerably here. Get it right and you'll get decent tool life, get it wrong and you'll find that your cutter or insert will show wear after even one part.

Now we come to Titanium. Lightweight, strong and corrosion resistant, titanium is a wonderful material for both structural as well as decorative parts. Like stainless it has some spring to it (even more than steel) and also work hardens. Unfortunately, it has an even worse characteristic which makes it difficult to machine, the tendency to retain heat. Unlike aluminum which transfers heat readily, or steel which sheds it in the chip, titanium tends to keep heat in the pocket, the immediate area being cut. This means that the tool heats up and beyond a certain point, the cemented carbide tends to soften and break down. Smaller radial cuts are the key here (axial depth doesn't have quite as drastic effect on cutter heating). A good, steady stream of coolant to clear the chips helps quite a bit as well as keep temperatures down but even so I've missed the mark and watched the pocket start glowing. While I've run HSS and cobalt tooling I find that coated carbide tooling works best here. Still, managing the heat means that unlike aluminum or mild steel, we have a very defined "speed limit" beyond which things get too hot and everything goes to shit in a hurry. The sweet spot here is very, very small. Run too slow and things take forever and you're just beating on the tools. Run too fast and you cook the tool and break it down. Miss your chip load and you either dull the tool or start chipping pieces off the cutting edge. Worst of all, even if you get everything right the tools still last only a fraction of the time they do in mild steel or chromoly. All this adds up to an expensive proposition with little margin for error or impatience. Still, the parts are light, strong and they polish out beautifully. Nothing good is cheap, right?

Is there anything worse than titanium? From what I hear some of the exotic nickle alloy steels such as Inconel are much, much worse. That said, they're strong and corrosion resistant but very heavy so I haven't had a use for them (yet).

So with all of the fuss over titanium is it really worth the trouble? I'll let you be the judge but considering that it's as strong or stronger than steel with only 60% of the weight and virtually corrosion proof... yeah, it's not bad:

 

Making Progress...

I haven't been posting to the blog as much as I'd like but it sure as hell isn't because things are slow. If anything, we're actually taking things up to a new level this year. The HAAS VF-2SS has proven to be a wonderful toy. I'm spending as much time as I can in front of it running parts but the reality is that there are not nearly enough hours in the day for this.

Burning the midnight oil:

That said, I'm making a lot of neat little toys and already working on some exciting collaboration projects.

Mini Kerambits:

To make matters worse (and yet so much better) we just took delivery of a HAAS ST-20SSY lathe.

My happy place:

While it's not hooked up yet (we're working on it) this is an amazing machine. Something that arguably will serve our needs much better than the mill. Why? Well, if you look at most of our bread and butter parts, they're round / turned parts. This lathe comes with live tooling and Y-axis which means we'll even be able to machine complex feature in one operation. Sounds great right? There might be a fairly steep learning curve here (plus tooling is going to cost a small fortune) and I think it's going to be a few months before we see sell-able product coming off of this thing, but final result should be pretty significant. It'll allow us to rapidly prototype new designs, make high-end low-production run products in house and produce designs which were otherwise too cost prohibitive to ever see the light of day. I can't say that I know exactly what that will look like but I'm pretty excited to find out.

Savannah and I have also taken up welding. Not as part of the business so much as fabricating our own tables and benches for the shop. I'm not sure if we're saving any money in the long run but we're having fun and learning a hell of a lot. It doesn't hurt that she's taken to it like a fish takes to water.

Savannah laying down some beads:

Peachy has gotten in on the fun as well with a Thermal Arc plasma cutter. She might look cute but the girl doesn't mind getting a little dirt under her fingernails either.

Peachy melting some metal:

 

Dave and I have also starter playing with TIG welding and I've get a tingle in my pants every time I think about buying a TIG welder for the shop.

Playing with a Thermal Arc 186 TIG welder in Dave's garage last weekend:

This doesn't even cover some of the ongoing projects. We just completed the Mini-Click Kickstarter Project (whoo hoo!) and already we're putting together Ti Bolt Flashlights for the Titanium Bolt Light Project collaboration. This means blasting, tumbling, assembling and testing. Fortunately the crew has a great attention to detail and we seem to be able to dial stuff like this in pretty well.

Savannah pressing Ti Bolt shuttles together:

First fully functional production Ti Bolt Flashlight:

I haven't even mentioned all of the assembly we now do on the in-house 416SS pen advancing mechanisms (have I mentioned how awesome they are?).

Joe putting together pen mechanisms like a boss:

This takes longer than you'd think:

Where does this leave us? Making progress. We've come such a long way in the past year it's amazing and we have such a long way still left to go. Hell, we're not even sure exactly where we're going to but it's one heck of an adventure, I can tell you that.

Stick with us and let's keep on trucking, enjoy the ride and see where the next year takes us. Just remember to keep your hands and feet inside the vehicle at all times.

-jack

 

In Pursuit of Quality...

My understanding of manufacturing has grown considerably during my time running Tuff Writer. Our pens have evolved from just a twinkle in my eye to an actual product and continued to evolve as our knowledge of design and manufacturing processes has matured. The funny thing is that the more we learn (it's definitely "we" now, Tuff Writer has grown beyond just me), the more processes we bring in-house, the more experience we gain... the more we realize that there's so much more left to learn.

In fact, we've added flashlights, carabiners and various other gadgets to our catalog. All the while, something we've never lost sight of is the pursuit of quality. We've worked hard to develop a keen eye for detail in the shop and an appetite for, if not the achievement of at least the pursuit of, perfection. That's not to say that we're perfect or that nothing ever gets past us (although we take it as a personal affront when it does and do everything in our power to make it right) but everyone in the shop takes great pride is making something special, something a cut above.

To that end, this seems to be a never ending education. I don't think we had any idea when we started purchasing our own CNC equipment what a wonderful and deep rabbit hole we were jumping into. That said, it's given me a fundamentally better understanding on what it means to make quality products.

You see... quality isn't just about hand finishing and hand assembly. The process starts all the way at the beginning during the initial design. So many components and products are beautifully designed and meticulously dimensioned for manufacturing processes that don't even exist yet. In other words, you can't expect a greater fidelity of the finished product than the resolution which the actual shaping process can deliver.

"Awesome, Jack... you wanna try talking English for a change maybe?" Hm... ok, let's try this.

Imagine that you're making a perfectly square internal hole in a 1" titanium plate which needs to slip fit a square block and hold .001" tolerances. In short, you're not going to machine this because a square internal corner simply isn't possible with an end-mill (remember, the mill is hold spinning cutters that are round and always leave an internal radius). You could waterjet the part but the kerf on the waterjet means that at 1" depth you're going to have tapered walls, no dice. A saw won't work because you can't get the saw blade into the square interior. 3D metal printing (i.e. DMLS - Direct Metal Laser Sintering) can make create geometry and thickness no problem but... with a tolerance of .003 on the high-end, there's just no way to hold the necessary tolerances. Grinding can easily hold the tolerances but it's not going to work for the internal square and the hole is just too small anyway. You're left with wire EDM (electrical Discharge Machining) which will use a very, very thin metal wire to cut through the metal.

And if you decided to design this out of Carbon Fiber... well, best of luck as EDM wouldn't work as it requires a conductive material (actually, I have no clue how the hell you'd make this out of CF but I'm sure someone smarter / more experienced could figure it out).

Now, if you can machine the 1" block to fit a square hole instead all of a sudden things become so much easier as you could loosen up the internal square hole tolerances and control the block instead. A 1" square block of titanium can easily be machined or honed to within .001" tolerances all day long. Heck, do you even need perfectly square internal corners or did the designer just think it looked pretty that way on the model?

All of this became especially apparent as we worked on the 416SS pen advancing mechanism. The dimensions were smaller and tolerances were tighter than anything I had worked on before. Fortunately, my buddy Chris was here (el hefe at Automatic Precision Inc. in Illinois) to help me understand what Swiss-type lathe turning was capable of. Even then, it took months of design work and modeling. Then we had to actually run the part and get the tolerancing right. In the end, we decided to have a custom ground shoulder end-mill made to be able to machine the race like we wanted it (bit of a risk, we didn't know if it would work like we hoped and still aren't sure how many parts we'll get out of the tool).

In other words, we've done our best to design quality into the mechanism and then continue to keep an eye on it through the entire manufacturing and assembly process. We've designed the new mechanism around the manufacturing processes available to us and sized it to fit within the performance envelope of the equipment we have access to (which, is admittedly pretty impressive).

Where does this all get us? For the moment we're golden (seriously though, have I mentioned just how awesome this thing is?) but then quality is a bit of a moving target. Just as cars have become much more reliable over the past few decades (remember the days when 100K miles meant the wheels were about to fall off and the engine needed a rebuild? No? I must be getting old) so will our expectations of quality continue to progress as time goes on. Without careful attention and constant vigilance, it's also possible to backslide (ever driven a Jeep Compass or Jeep Patriot?).

So what's next? We keep moving, keep learning.  The more we keep our hands dirty and the more tools we add to our design and manufacturing arsenal, the more cool $h!t we can design, manufacture and ultimately delivery.

Guess this means we keep adding toys to the shop and then figuring out what the hell we can do with them.

I'm in if you are.

-jack