Alpha losing position

[gabepari]

So, this is the third time this has happened in the last 6 months. Cutting along, small part moves and gets jammed between the waste and the bit, no big deal, just one part messed up. Then after a few sheets, the operator notices that we're not cutting square any more. I Check some things, the x car is racked by more than 1/4". I don't think it jumped a tooth, we should have heard it. And if it's jumping teeth under torque, that's unacceptable.

I know we've discussed this before, but I'm not so sure that the Alpha is immune to losing position.

I'm going to do a little more figuring when I have time, but I know the pinions are not loose, and everything else is tight, oiled and running smooth.

Oh well, only ruined $1000 worth of wood this time. {insert little man hitting himself on the head with a big mallet}

Gabe

[knight_toolworks]

when it happens why not square it up to make sure? I know when quit a run I always zero the machine and the z to just make sure.

[Brady Watson]

Gabe,
You cannot aggressively cut with an Alpha 1:1 and expect it not to lose position, or deviate from the programmed path. This is a sticky & complex subject. Many people expect a 1:1 Alpha to run @ 10 IPS while cutting 3/4" material in one pass, with their ramping set at the default settings and they complain about the motor going into Alpha-mode (causing a divot where the tool makes a 90° turn). It seems to me that one of the X motors took the brunt of the impact, while the other kept moving. You have about 3 seconds before an Alpha will fail the driver and stop the machine. It will be using all of it's torque to attempt to get back into position, but if the required amount of torque exceeds what the motor is capable of delivering, you are going to have a loss in position on one of the motors. The operator should learn from this experience and stop the machine, pull it against the stops to ensure that it is square, and continue production. It is a good safety feature that the motor stalls.

Expecting an Alpha to be absolutely invincible & impervious to losses in position is like asking your 4X4 with ABS to stop on a dime on a frozen lake. There are limits to everything. It would be a worthwhile investment to upgrade to the PRS motors that put out 2X the cutting force of the ones you have. I have run mine @ 12 IPS and have never experienced a deviation, like the 1:1 versions.

-B

[conceptmachine]

Getting out of square is the biggest problem with a bolt together machine!It wouldn't cost that much more to at least have a rigid Y car.
-shawn

[Brady Watson]

Shawn,
It's only a problem if you crash the tool. My Alpha still cuts dead square on after 15 months of cutting. If your tool constantly goes out of square, something isn't right.

-B

[harryball]

You're killing me Brady... I really want to do the upgrade to the new motors and you just keep making it sound better and better! I neglected to ask Chris, when you upgraded did you have to trade in your old motors and controllers or did you keep them?

Robert

[gpari]

Brady,

this is this first time it has happened since I have installed the 3:1 boxes (home-brew). First thing I checked was all the grub screws and such to make sure nothing slipped. they were all good. Next thing I'm doing is making index marks on the pinion gears and belts so I can be sure that nothing skipped.

It was moving 5ips at about .625" deep in 3/4" ply with HPL. It wasn't a violent crash, enough that my operator didn't even think to recheck square. BUT, trust me, any little hiccup and he's checking it from now on.

My machine cuts dead square as well, and always has. Just need another sticky note on the operators desk.

I understand the reason that one motor can be unsynced from the other and keep on going like nothing happened, I just don't like it, and wasn't expecting it to be the case when purchasing the machine. The work around is not a big deal, and I'm not ranting, just blowing off some steam after ruining a bunch of panels.

We single pass 3/4" MDF at 10ips with the 3:1 boxes with no problems, but for some reason the ply we were cutting was really tough. Even the table saw bogged down on the stuff.

Thanks,

Gabe

[Brady Watson]

Robert,
You return your old motors and drivers in exchange for new 7.2:1 motors on the X & Y axes. You keep your original 1:1 Z motor. I'm telling you...it's an absolute steal at the price that ShopBot offers the upgrade. You'd pay around $1800 for ONE 7.2:1 motor and driver if you bought it from Oriental. ShopBot is extremely generous with this upgrade since you get 3 sets of motors/drivers, 3 pinions ($50 ea), 3 allen keys to really crank down those pinions properly, plus whatever else you need to do the job properly.

The benefits that you will receive far outweigh the cost...You will pickup 2X the resolution of the 1:1 setup, plus nearly 3X the cutting force. I weigh about 210#s and I REALLY have to try hard with all my weight against ANY axis to get it to move and go into Alpha mode. Since the pinions are 30T instead of the 1:1's 20T, you get higher life on both your rack AND pinions. Since there are more teeth engaged in the rack, there is less force per tooth against it, and things wear much better than the 1:1 setup.

PRT Alpha owners can pick up a little more smoothness and quiet down their tool by adjusting the V.Fil rotary switch on ALL Alpha drivers in the control box (4 of them people) to a value of '8'. TURN OFF THE CONTROL BOX 1ST! Then, rotate the arrow to the 8 setting on all drivers & then close the box and turn it on. DO NOT mess with the current adjustment. Just the V.Fil one, as shown in red below. Merry Christmas...


3704


Gabe, you got one in on me while I was typing. Sorry to imply that you were running the tool to hard...and glad to hear that you are grooming your operator to do what you would do if running the machine!

-B

[conceptmachine]

Brady,
I personaly don't have any issues with getting out of square because i don't push the bot beyond it's capabilities but i do ocasionaly check to see if it's still square and all the bolts are tight.I don't have my y car welded as you do either so i don't have that added security.There are alot of people that buy a bot that have no experience with cnc and don't know how to properly set one up and operate it so if you at least had a rigid y car this would eliminate alot of frustration for some people.
-shawn

[Brady Watson]

Shawn,
My 6yr old PRT has a welded gantry. My Alpha is 100% as delivered from ShopBot - the same bolt together deal as everyone else with a PRT Alpha.

-B

[wcsg]

I just noticed this thread about the V.FIL. I popped my greasy head into my control box and noticed mine are set on A, is A higher or lower than 8?

[Brady Watson]

Higher...The dials are base16, meaning A-F represent in this case #10-16. 0 is lowest & F is highest. A is too high and can cause problems. Anything in the 6 to 8 range is a good setting.

The V.Fil dial is there to adjust mid-band resonance inherent in all stepper motors.

-B

[henrik_o]

What about the new PRS systems? Is it advisable to change the V.FIL setting on these as well?

Edit: all five drivers are currently set to 6.

What exactly does this accomplish, technically speaking? Is there anything that could go wrong?

[weslambe]

Still lurking.

I have had both bolt together and welded frame machines. The machine that I just built has some of both but the gantry isn't welded and neither are the main rails. The reason fof this configuration is that I CAN make changes as needed if needed.

A solid welded machine is a solid welded problem if they aren't perfectly aligned to begin with. Quite frankly, you could have hinges on both ends of the gantry instead of bolts or welds if the motors are running correctly and have the torque needed to hold position as they run.

A couple of physical stops to get the machine square are essential.

Mike Richards,
Yes, I have issues running mach 3 at 100 kh. Very choppy. I can run my new machine at 65kh and get about 10-12 ips. I did get a flaky gekko 202 that craps out pretty frequently but I am getting a replacement from them. So far so good. It's dang close to being a perfect machine. I still have some tweeking to do but expect great things. I may post some new pictures on my website soon. www.doorbot.net (http://www.doorbot.net)

[richards]

Wes,
Do you have the power supply D.C ground connected to chassis ground (along with A.C. ground)? Dirk Dunham and Pat Fulghum discovered that they had to tie the D.C. ground to case ground to get their indexer to work properly. At first I didn't have that problem with my indexer, but I started getting the same kinds of problems that Dirk and Pat had told me about in their emails. As soon as I tied D.C. ground to chassis ground everything started working perfectly for me.

It seems that the Gecko stepper drivers use hard anodizing on the case as an insulator. The metal screws that we use to bolt the stepper driver to a heat sink can cut through that anodizing and cause problems. Marris has always recommended connecting the D.C. ground to chassis ground to eliminate some of the 'flakey' problems that are hard to trace. (There's been an intense discussion on the Gecko forum on Yahoo groups concerning how whether to tie D.C. ground to A.C. ground to chassis ground. Those who've followed Marris's instructions and tied the grounds together aren't getting the 'hard-to-diagnose' problems that others sometimes get.)

[Brady Watson]

Henrik,
Leave them @ 6. They smooth out mid band resonance. Yes...too high and it can cause problems.

-B

[weslambe]

Mike,

I don't really know what I've got going on in the controller box. Stan Holt (Holt Control Systems) built my controller box. I supplied the parts, he put it together. Great guy, great company.

About my Gekkos. Only one of the x axis drives craps out on me. The machine is stout enough to drag the other side along, but it does cause the piece under it to get destroyed.

I posted several new pictures on my website a few minutes ago.

[richards]

Wes,
As you know, there are a lot of 'gotchas' that can creep into a CNC router to cause anguish. One of the most interesting things that I've ever done was to connect an oscilloscope to my Alpha and just watch various signals. All I can say it 'Wow'! There's enough noise on those signals to make me wonder how it even is able to run. It's a credit to Ted and the rest of Shopbot's crew that they knew how to make it all work together.

This entire thread kind of points out that the unexpected can happen, and usually does, even after a machine has been running reliably for a long period of time.

The Gecko G20x stepper drivers have 'pots' to smooth out the stepper's motion. The Alpha stepper controller has a 'pot' to do the same thing. Proper grounding of a machine can eliminate some of the problems. Proper shielding of the signal conductors from the power/spindle conductors can sometimes help.

In your case, your X-axis driver might just be faulty, or it might be picking up enough RF that it can't tell the difference between a valid signal and signals imposed by other electrical devices (especially the spindle - or perhaps a neighbor down the street that's running some electrical device that outputs a lot of noise). It's hard to tell without swapping parts around a little and turning things off and then on again to see what happens.

Your machine has some very nice features. I hope it all comes together quickly for you.

[pete]

I could be wrong here, but isn't it a bit of a cheek to use a CNC machine manufacturer's forum to promote another brand of machine?

[wcsg]

naw, this forum is not like that here.

[richards]

Peter,
If you were Ted, and were handed tens of thousands of dollars of free research and development time and materials for the cost of a few posts on your company's forum, what would you do? After looking at the latest PRS machines, I can't help but think that some of the banter that was passed around on this forum helped solidify the needs and wants of the Shopbot community. Some of the things that were 'hot' topics on the forum found their way into the PRS machines. That doesn't mean that the forum 'developed' those ideas and designs, but it showed that there was a lot of interest in those ideas and designs.

Wes has just about completed a prototype of a machine that incorporates the best ideas that he has about what he thinks a CNC router should be. It may be a very good design. It may be a faulty design. Only time and testing will tell. For those of us who have built a lot of prototypes, his machine would be the 'alpha' model - not to be confused with the Shopbot Alpha product. An 'alpha' model is the first working model that incorporates most of the things that were desired during the design phase. Sometimes an 'alpha' model actually works. Most of the time, however, the 'alpha' model primarily shows where the flaws are in the design. In any case, when any flaws in the 'alpha' model are identified and corrected, the 'alpha' model becomes the 'beta' model. The 'beta' model is seen and evaluated by trusted customers and testers to compare to machines and models that work. Then more changes are made, designs are refined and, if the 'wow' factor is great enough, the design enters limited production. All of that costs a great deal in time, labor and materials. Having someone offer the fruits of his own labor as an example of what he feels is needed, shortcuts that process - and we all benefit. I think that we all should be grateful to Wes and others like him who use their own time and money to try new things. If they're right, in time, we'll all benefit. If they're wrong, they will have saved us from trying the same things.

If there are two things that defines most of us who own a Shopbot, it is the fact that we are curious and that we are willing to experiment. Let's let Wes experiment. He's willing to share his ideas and his design. If the thousands of Shopbot owners each found a way to improve the machine in some small way and then offered that improvement to the community, think how much better off we all would be.

[weslambe]

Peter,

I don't think it's that cheeky. I'm not selling a machine yet and may never do so. Also, I was a shopbotter for several years and probably would still be one if it weren't for my faulty Dell computer. I think the new PRS is an awsome machine and wouldn't have bought one if I didn't think so. (that damn dell pc) oh well, can't turn back the clock.

Mike,
The red light on my faulty gekko actually goes off and won't come on again until you do a power cycle on the controller. I don't think that fits the static/grounding issue but thanks for your guidance. BTW, this is my 5th machine. I have incorporated the best from the previous 4. It makes very nice cabinet doors so far. BTW, the machine is finished and running. The only hold up is the gekko. What I have been able to do so far has been remarkable.

[richards]

Wes,
I must be blind because I can't find the reference in the Gecko G202 manual that talks about the L.E.D. I know that I've read about it somewhere, but I can't remember where. However, as you know, the L.E.D. is on whenever the stepper driver is operating correctly. Other than driver malfunction, there are three conditions that would cause the drive to shut down:

1. No power on pins 1 and 2 There is a known problem with the BLUE header connectors that Gecko furnished with some drives. They found that some of those connectors were not making good contact - and giving intermittant errors. (My four G202 drives have the BLUE connectors and they seem to work fine, but I only use them on my test bench.) The other eleven Geckos that I have use the BLACK header/connector. A broken or cracked wire could also cause this kind of problem.

2. Short circuit "The G202 is protected against short-circuits across the motor windings, from one motor winding to another and from winding to ground." Check the header/connector with a magnifying glass to see if one strand of one conductor worked its way into another connector. This has happened to me.

3. Overheat The drive will shut down automatically if it gets too hot. Drives pulling 3A or more must use a heat sink with heat sink compound.

One final quote from the G202 manual: "When a short-circuit condition is encountered, the G202 goes into protective shutdown by turning off the winding outputs. The G202 outputs stay in protective shutdown until the power supply is turned off and discharged, then turned on again. Should this happen, please check the motor wiring and the cable going to the motor for errors." So, the fact that the drive sometimes acts normal would make me think that there might be a connection or heat problem. And, as the quote points out, once the drive faults, power has to be turned off and drained from the capacitor(s) before the drive will reset.

I didn't mean to give offense to you about your machine. You're using some design features that I've wondered about for a long time, such as the linear rails and a solid sheet of vertical steel for the Y-axis. But, if you're like me, you'll run the machine for a while and then you might wonder enough about one aspect or another to try a slighly different design. For instance, here in Utah, the temperature yesterday was a perfect 80-degrees. Today we have snow on the ground. With that big of a temperature shift in my unheated shop, I would have mounted several dial indicators on each axis to see if all the metal was expanding and contracting at the same rate or if one piece was expanding and contracting at a different rate. You might not have those kinds of temperature swings in your area, so you may never need to check. I have other 'problems' caused by the fact that I'm running my PRT-Alpha in a residential area. During the hottest times of the hot summer days, the local power company can't keep the voltage at a constant 117VAC. There are times that it gets suspiciously close to 95VAC, which is a near-brown-out condition that would cause some power supplies to over heat. In short, if I had a new machine of my own design, I would find myself checking things for a while until I had enough hours behind me to know what to expect. So, it's not a question about whether your design is good or bad, it's more a question of how everything works together as a system.

A company like Shopbot gets a lot of feedback in a hurry when they ship a new design. That doesn't necessarily mean that they have a lot of problems, just that they have a lot of eyes looking at whatever problems might exist. A new design with a single owner/operator would naturally take longer to exercise a machine long and hard enough to discover every possible quirk. And, to me, until every possible quick has been identified and classified as serious or minor, the machine would be classified as an 'alpha' model - in production, but still being carefully watched.

I have a very good friend from Greece who owns a large photo lab. At any particular time he has well over one-million dollars worth of equipment in his lab. Whenever I've asked him how a new piece of equipment is working, he always says, "So far, so good, but it's too early to tell for sure.". He gives that same answer for every piece of new equipment for at least one year.

[cnc_works]

Wes, I had the exact same problem with one of my Gecko G202 X drivers intermittently failing. My PRT 96 did not have the oomph to continue consistently and ended up losing steps and making sounds my ears did not want to hear. LED went out then came back to work normally after controller cycle...until the next unpredictable event. Replaced the G202 and my problems went away.

Donn

[weslambe]

Mike,
No offense taken. Thanks for the advice and input. I'll take any given. This really is one hell of a machine. I am pleased and really looking forward to building another one for my brother. The next one will have some improvements in it since some parts will be CNC cut. (didn't have one, now I do) I will also change the way that some of the parts are made. The process will be improved so that perfection will be a little easier to approach.

Donn,
Thanks. I'll bet that's what I've got going on but I'll check Mike's fix for sure. Now that I don't have a Shopbot I kind of feel weird posting here. I guess it's kind of like the feeling you get when you stop dating a girl but still remain friends.

I remain eternally grateful to Shopbot since without their prt I would have never gone this far in the woodworking industry.

[richards]

Wes,
I'm not against replacing the G202, but if the problem is somewhere in the cabling or the connections, it will still be lurking even if a new G202 is installed. One of the hardest parts of self-manufacturing a machine is the fact that you have to become the technical expert of every part of that machine. That can lead to a lot of head-scratching until the machine reveals its secrets.

When I opened the crate to take my first look at my PRT-Alpha, back in July of 2004, one of the things that really pleased me was the fact that Shopbot had pre-wired the Oriental Motor Alpha steppers. That made life so much easier for me then and it's one of the 'features' that gave me confidence in the whole machine. As little problems arose from time to time, I first looked at what I'd done before picking up the telephone to see what Shopbot had done. Right now, I don't remember a single problem that was caused by Shopbot, but I do remember quite a few things that I caused.

But, given a little time and experience, every machine starts to become an old friend. We learn its quirks and we learn how to work around them. (My customers have to deal with my quirks. When they do, we get along just fine. If they don't - well that's another story.)

One thing that I would recommend to everyone is to keep a journal. Write down every problem and the cure for that problem. That way, you'll only have to be a detective once. The next time something happens, you can just thumb through the book to see if you've faced that same problem before. If you have, you'll also have the solution to the problem.