Official Luthiers Forum!

I recently had the opportunity to go head to head so to speak with a PLEK machine.
Here is what I found http://howardguitars.blogspot.com/2017/01/man-vs-machine-or-is-plek-better-than.html

_________________
Brian

You never know what you are capable of until you actually try.

https://www.howardguitarsdelaware.com/

Well done, Brian - fight the good fight!

_________________
"Act your age, not your shoe size" - Prince

Great write up! The quality of the pictures and the detailed descriptions of what you do and why is really helpful, beyond the comparison to plek. I know plek is sold as the cats meow. I have zero experience with it. In general though the goal of automation is to remove or reduce the need of skilled workers. I am sure there are quality tradeoffs made in the machine design itself and then the setup and operation of the machine to reduce the need for skilled workers and to maximise output.

The danger is that while some men can beat the machine, the machine can beat most men and the machine's work becomes the acceptable standard.

_________________
http://www.Harvestmoonguitars.com

Nice article Brian. Plek being a relatively new technology perhaps it's almost a matter of time that they jsut get so good, and cheap, we cannot compete with them. Ultimately the machines will take over the world

I almost always use a beam too unless I simply cannot get the FB straight but that brings up a point that goes against that philosophy. Take setting up a nut for example. Of course there are many ways of doing it but most are done by either scribing a line across the face of the nut at fret height or measuring the fret height and slotting to that measurement and so on. This is all done with the fret directly in front of the nut or maybe the first couple. Everyone want the nut slots to be at the same height of the frets in front of it but no one measures the 14th fret for example and then transfers that to cut the nut slots.

IOW in order to not get buzz on any given fret the first few in front of it need to be the same height, just like the way the nut is set up. Of course this is all within reason, if a fret end has popped off that's a different story. So why do we need to be so accurate as to use full length beams and get the 1st fret within .0001in of the 14th? Especially when the 14th has even more angle up to the saddle.

Thanks, helpful write up.

Excellent write up Brian.

I'm helping a friend build a 4 string fretless electric bass. Neither of us have built or worked on one. Do you have a write up on setup by chance?
I'm wondering what clearances at the nut, 12th fret and 24th fret we should we shot in for.

Any help would be appreciated

_________________
Joe Beaver
Maker of Sawdust

It's all geometry. As the frets get higher they get closer, thereby counteracting the increased string angle. And in reality under playing conditions we want the neck to actually have a slight bit of relief to allow for the vibrational envelope of the string when played open, this also takes a little bite out of that increased string angle on all the frets past the center of the truss rod (not the center of the neck necessarily).

_________________
Brian

You never know what you are capable of until you actually try.

https://www.howardguitarsdelaware.com/

Good going Bri!

We were offered a PLEK free of charge and turned it down. Some may say WHAT you turned it down. Yep we have always believed that we can do superior work to PLEKs with our hand methods AND that this is a real value-add for the clients who's perfect action depends on superb fret work.

So why wouldn't we take the PLEK anyway if for no other reason than to hang out our laundry on it? Two primary reasons. First you never really own a PLEK it's a licensed thing with proprietary operating systems that require regular updates and support from PLEK. PLEK charges for this.

Second we believe the design of the PLEK to be fundamentally flawed..... For all of time it's been a cardinal rule of machinists and tool and die makers, CNC, etc. that your measurement tools be completely independent of the cutting or business end.

PLEK uses the very same rails that position the cutting tools to also do it's measurements. Not good and kind of not unlike breathing one's own air so to speak in so much as any errors in the rails will be duplicated, or not, in the resulting measurements.

A PLEK in our experience never was superior to a highly skilled Luthier with very well developed skills AND methods but it is faster, kind of cool if you like machines and I do, and an excellent way to bankrupt your business when you find out that you need to keep it humming for hours every single day to pay for it's self. That's before you make money too. I figured out that you really need to be doing multiple fret jobs nearly every day to pay for one of these things.... For f*ctories a PLEK is a different matter and it does not come back from lunch stoned or drunk either. dang thing....

The last time that I priced them a civilized one, not the Mr. T starter set PLEK, will hurt you to the tune of around $115,000.00 to get it running and then there are substantial charges for trainings, support, etc.

Lastly what happens when a tool with a very limited distribution and proprietary software has it's mother ship company go teats up? You lose updates, support, bug fixes etc and eventually the use of the machine.....

That's why we turned down a free PLEK.

And.... you know something else too. As Brian well knows some of his customers are "his" customers because great Luthiers develop loyal followings. Just like the "romance" involved in having a big name Luthier build you a custom guitar some people prefer a well respected Luthier to be the only one with his or her hands on their ax and promising to not teach it any new tricks. These sorts of clients are not really keen to think about their baby strapped into a 50 shades of PLEK machine.... maybe....

Why are the words to "John Henry" running through my head?

Brian Howard was a fret sandin' man . . .

I think the message that the PLEK is a flawed machine is the one that resonates best. I don't think we should be as bold as to say that a machine will never be as good at this job as a person. We just don't know that. But we do know that this particular machine has flaws in its design that, in fact, on the ground, cause it to do inferior work, compared to the work of a highly skilled person. Players will see the truth in that when they experience it, as Brian's customer did.

I got a lot of crap on another forum when I wrote that PLEK's were very limited and could not compare to an experienced and skilled luthier. I do not think the machine benefits from experience, nor does it learn nuance. A real person can do that, and most of the craftsmen who've done levels for years can beat a PLEK - I'm sure of it.

_________________
"Act your age, not your shoe size" - Prince

Thanks for the write up....

As Hesh points out.... The place where a Plek makes you money is at the factory.... Not in the Luthier's shop.... Say you do 10,000 guitars a month.... Well - the $115k only costs you $11.50 each guitar the first year.. That's a lot less than 1 hour of paid employee time per unit..... Makes total sense - as poor fretwork is a huge factory warranty/customer complaint issue.

The factory also has a huge advantage with relatively few but consistent models. Every standard profile neck on factory new J45's are identical... Frets are the same, neck profile the same, fretboard the same, etc... That saves a LOT of setup time. Just clunk it in the holding fixture, indicate it in, and off you go.....

On the other hand... For the luthier...
1. People pay him for the time to know how to do it right.
2. Not nearly enough volume to support the unit's cost
3. It takes very specialized training and is very operator dependent. And let's face it - most repair Luthier's aren't trained CNC machine operators where they would already have a basis for understanding the operation.
4. And maybe the biggest trouble item.... Every guitar is different... First is a 1986 Korean ovation.. Next a 1975 D28, a $100 Rogue, a 2 year old Taylor, a Mexican Fender squirt, a 1950 Gibson LG2, a 1971 Harmony Stella, a 1908 Washburn, next a 5-string bass.... No two are the same... So the setup takes relatively a LOT longer.

One thing I saw with CNC at work was....
It's generally faster for a good tool and die man to make ONE piece using standard machining techniques... But it's generally faster for the computer to do quantity work...

Hold on just a minute everyone. I've spoken to Hesh and Dave Collins about the PLEK, and if I recall correctly, Dave suggested that the PLEK could do frets that were the equal of anything done by an extremely competent repair guy,...if it was run by an extremely competent operator.

That doesn't mean that it's appropriate for the average repair shop. Hesh and Dave are extremely competent repair guys, so a PLEK would just be redundant.

I guess my point is that it's just another tool. Appropriate in some circumstances, not so much in others. And as with any tool, its effectiveness depends on who's operating it.

Steve

Brian nice job.

I traded into a new "made in Nashville" Les Paul fresh off their PLEK. It was pretty rough, fret finish was poor. I have no idea what they consider a proper string path, the nut slots bound the strings and were on the high side. Took me about four hours to get it up to speed. Plays good, sounds good

Our experience echos that of others here - it is surprising how often we see instruments in the shop that have been PLEK'd by either the factory or by other shops and are still in need of significant work.

_________________
A constellation only takes shape when one maps the whole.
- Beth Brower

A couple of points to resist or make.

How the PLEK is deployed is of course of value when evaluating it's work. It's been my impression that the lousy fret work that we see too from G*bson is likely because the instruments spend minimal time on the machine. I say this because I see the tool paths.... the lines milled in the frets. Similar to scratches they represent not enough passes to improve the overall finish of the frets. Do this with the business part of the procedure, milling the tops to precision levelness INCLUDING accounting for final fit and finish and it's likely that some manufacturers are simply speeding things through and that the machines are capable of much better.

By the way errors in fret work are easily detected and measured once we define errors over what distance or span. It is very possible to design a process to measure and compare the results of a skilled Luthier vs. a PLEK machine. This is what we have been wanting to do too for around 6 years now with no takers.... We did score one interested party who agrees with our assertion that we can do better than a PLEK (he can too) but he would not agree to let us publish our results.

Steve we do say that a PLEK is only as skilled as the operator but there are other limitations. The code is not open source and after talking with PLEK experts that have more than one of them they have told us that the code does not offer functionality for some of those "Nuance" things that Chris very rightly brought up. Remember Steve when we are applying finger top pressure, the touch if you will, to correct a specific anomaly of a specific instrument while either leveling the board or final leveling fret tops? That's nuance and that's often important with fretwork.

If you have an ax that has a neck that responds different under string tension than say at least for the fret work we human bags of mostly water... can observe this, correct for it with the "touch" and that this is not easily done with a PLEK.

Don great comments and it may be a reach to say that a machine will "never" be capable of exceeding hand work. But it's not a reach these days to say that. From polishing multi-million dollar lenses for Hubble to some of the smallest tolerance machining ever done the skilled craftsman is still relied on to do the details. I can have a robot operate on me but I would much rather sue and punch a human being when they botch the job.....

As someone who learned to program back in the day in Fortran 77, COBOL, Basic, Pascal, MAD (proprietary, stands for Michigan Algorithm Decoder) and several other computer languages on the processor side the only thing a PLEK has over we humans is it won't get tired doing multiple iterations of the same math and it's way faster. We humans are still capable of the same precision.

We humans can also do the entire job including discuss options with the client, address any issues that may result after the work is done, we are not dependent on anyone else to turn us on, strap the stinkin ax on our bench, or motivate us to do well, pay the electric bill, pay PLEK, get updates, change tools, maintain a suitable industrial environment, etc. Seems a PLEK has a long way to go before it can compete with just how exactly irritating a human being I can specifically be....

Don't get me wrong there are good places for PLEKs but I just don't think that it's in a dedicated Lutherie repair shop or small builder shop yet unless you really push product through with multiple employees and a huge amount of business AND people don't come to you because of your skill and "touch..."

It does seem odd that a PLEk doesn't address the problem of a separate motoring system. If luthiers find poor PLEK fret jobs,(I have), the machine must not know what it's is ultimately turning out. But there is so much development in 3D photography and digitizing things in free space, it must only be a matter of time before some optical/laser type system could be integrated into it.

I think the idea of a PLEK is so appealing because simple human observation is almost incapable of sensing how things really are. Especially at the level of accuracy required for good fret work.

After years of trying to make "perfect joints",good fret planes, level shiny finishes, etc.,and using dial gauges, straight edges, and microscopes, I see the world around me differently. Things I always assumed were "straight", "flush", "stiff", "hard",etc., just are not as much as I thought. If I were to explain how close in height frets need to be to work well, a typical customer would think I was unhinged, but if I slap a .002" piece of cellophane tape on a window and run their finger over it ,They would realize just how big a couple of thousandths can be.
Maybe the most powerful part of a PLEK is how it taps into peoples fantasies of machine perfection.

Good point, David. Lots of folks just don't understand that what makes the world work is very tiny tolerances. Some of the tool & die work I used to do is a case in point. I had to machine some stuff to half a thousandth tolerance or tighter.

For those who couldn't comprehend, I'd illustrate thusly: Pull a hair from your head (usually .004). Split it into 10 equal portions. Now, machine a part to 1 of those 10 portions (holding the .0004 dimension). Repeat many times until you have the correct number of parts. Make it .0001 too big, and the fit will be sloppy. Make it .0001 too small, and you won't get the parts together. That is pretty dang picky tolerances.

That is why grumpy old guys like me are suited to be grumpy old luthiers. We set the bar pretty darn high, and don't tolerate those who say we don't measure up.

_________________
"Act your age, not your shoe size" - Prince

A couple of thoughts. I had a client with extremely unreasonable demands of string behavior on a 30" baritone. He was in Nashville, so I sent him to Joe Glaser who Plek'd the guitar. Sometime thereafter, I called Joe to ask how bad my fretwork was. He said I was over the 95th percentile of all the work he evaluates. A couple of minor flat spots, but that's it. There wasn't much he was able to do. I found that encouraging.

On the other side of things, I know that Jeff Traugott has all his guitars Plek'd. I would say Jeff has some of the most discerning and demanding expectations of any builder out there. That Jeff finds value in it says a lot. Is any device or person going to be be able to account for everything? No. But they are going to get there more accurately. I'd bet on average the Plek is far more accurate than the average fret job or dressing.

A plek is a machine that can be programmed to be run by an unskilled person. I agree with Brian a good tech will do as good or better job. I have redone a few plek jobs. The only advantage is that the machine is the skill not the operator. Brian myself and many others have learned how to do a fret job. That is the skill.

_________________
John Hall
blues creek guitars
Authorized CF Martin Repair
Co President of ASIA
You Don't know what you don't know until you know it

I think that guys like Joe Glaser and Phil Jacoby would disagree, and disagree quite vehemently with that statement.

To get the most out of the PLEK, the operator needs to be trained to a very high level of competence. It is not a "strap it in and push a button" operation, although that seems to be how Martin and Gibson treat it. There is a world of difference between a guitar that comes out of Joe Glaser's shop after being PLEKked , and the average new Martin guitar.

The very best PLEKked fret jobs, are the equal of the very best hand leveled and crowned fret jobs.

The PLEK doesn't do the fret ends, of course ... that is still where the human hand reigns supreme.

When I was at Collings there were two Fadals. $80,000+ each. Designed to cut metal all day. One day it was decided to switch an operation from one to the other. When the fixture from one was placed in the other, it was discovered that the locating pin holes didn't line up with the other machine. Heads were scratched, calls were made, and it turned out that even though the ways and lead screws got a continuous, automatic lubrication, simple wear on the screw had put a .009"(!) discrepancy in travel from one end to the other. That's what I was saying in my previous post. Things (machines) are rarely just how we imagine. Smart people had assumed the machine was so massive, high tech, and expensive, closer monitoring was not necessary. In fact it had been used as a reference tool. Measuring points on objects for CAD drawings. Everything drawn was theoretically a little out of wack. You have to periodically check the checker!

The operative phrase there is CAN BE, he didn't say WILL or MUST BE.

As an old machinist who learned G-code at the knee of Jesus, I can tell you all this new "conversational" programming used in newer machines enables the complete idiot to program. They don't even have to read a print properly. And that's what a lot of companies want for employees, because they can pay them MUCH less than people with real skills, knowledge, abilities, and experience.

It does help a lot, when the programmer can actually read and understand a print, think in 3 dimensions, and plan a complex job from start to finish - including designing the fixtures for the job. When you can run any machine in the shop competently, and program rings around most CATIA jockeys - you get stuff done under budget and please every inspector along the way. Which is why (without bragging too much), at my last job at Beechcraft - I zoomed to the top of the pay scale in a less than 4 years, at a time when many guys with 20 years employment were still in the middle range of the pay scale.

Plus, I was handsome and easy to get along with.

_________________
"Act your age, not your shoe size" - Prince

Geez, how often do you get that kind of comedic timing in a piece of writing?

Thanks Chris, my laugh for the day.

Steve

You are most welcome, Steve - glad to entertain and please you.

_________________
"Act your age, not your shoe size" - Prince

Don't forget that things might get a little funky with a PLEK'd guitar just as much as one from Hesh's shop if it takes a trip in the back of the wrong tour van... as long as folks play wooden guitars a sudden change in temperature and humidity can undermine the greatest of fretwork masterpieces.