IH Build

[SMW Precision]

[QUOTE=IndHobby;220790]Personally I like optical isolation, so lets look at the basic issues.

Geckos and Limits.

Geckos are optically isolated onto themselves.
Optcal Limits are optical isolaters, take one apart and look.

If you follow the basic schematic and steal the control power from the PC you will be isolated.

Insofar as the motor and coolant. The SS relay I use is optically isolated as well and can be driven right from the PC. Solves that problem as well.

We do use Bob's VFD controller board, it can be used by itself, and to my knowledge can be isolated as well just by using a seperate PS on the VFD side.

I have helped troubleshoot more than one breakout board problem over the phone and it starts getting real old real fast, so have a hard time reccomending them to a customer.

The other problem I have with a breakout board is the real estate it takes, they are huge. If we are concerned with chips or coolant getting in the wrong spot lets not see how big of a target we can provide.

My first machine is all wire to terminal connections, just hard wired, nothing fancy. And it looks so bad I would never post a picture of it, it was just a concept and I always figured I would go back and rewire it later. Later never came. But she just keeps chuging along, about once every 6 months I blow out all the stray metal chips and leave it alone.

The point I'm making is that simplicity is a quality onto itself, simplicity usually brings robustness another quality I like.

Aaron,

Thanks for the comments they help! The other Key question that has come up is wattage capabilities of PC for supplying Gecko drivers boards. Are they a mismatch?

From; Richard
The documentation that I have states very clearly that the Geckos need 16ma of current to work properly. In the TTL world, 16ma is a lot more current than most TTL circuits need. In fact 4ma to 10ma are much more common than 16ma. Parallel ports, in my opinion and with my experience, can't be counted on to supply 16ma. That means that the step/direction signals might not pulse when they're supposed to. That also means flakey behavior and possibly missed steps.

Having recently been through computer power supply issues this is of concern to me. There are huge variances between computer power supplies in quality and wattage capabilities. From the bottom end $15 PS on up!

What are your thoughts on PS issue?

[IndHobby]

Thanks for the comments they help! The other Key question that has come up is wattage capabilities of PC for supplying Gecko drivers boards. Are they a mismatch?

***Not an issue, a PC has far more juice than you will need, all the little parts were talking about here draw less than a second hard drive.

From; Richard
The documentation that I have states very clearly that the Geckos need 16ma of current to work properly. In the TTL world, 16ma is a lot more current than most TTL circuits need. In fact 4ma to 10ma are much more common than 16ma. Parallel ports, in my opinion and with my experience, can't be counted on to supply 16ma. That means that the step/direction signals might not pulse when they're supposed to. That also means flakey behavior and possibly missed steps.

*** USE a SIIG CyberParallel card, off the top of my head they will source 20ma and sink 50ma per line. It's what all of us guys use.

Having recently been through computer power supply issues this is of concern to me. There are huge variances between computer power supplies in quality and wattage capabilities. From the bottom end $15 PS on up!

*** Please do not cheap out on the powersupply, spend the $50 for a good one and it will last you forever.

[Richards]

Aaron,

I guess that we'll have to agree to disagree on breakout boards. After reading your posts, I downloaded documentation for Bob Campbell's board and then reread the documenation to the PMDX board that I use. Both sets of documentation seemed to be standard stuff that would work if followed. Granted, because all of our backgrounds are different, what seems standard to me might be something else to others - and vice-versa.

Some of the boards are large, others are fairly small. The PMDX that I use measures 3 x 4-1/2 inches, which, I believe is small enough to fit in most boxes. Speaking of boxes, I've always assumed that anyone adding CNC controls to a mill would use an enclosure that would protect the electronics in a liquid environment. Maybe I'm expecting too much, but it just seems logical that, if there is water present, the enclosure has to keep the water and chips out. Enclosures come in all sizes. Getting one big enough to include a breakout board would also guarantee that the case would be big enough to dissipate the heat generated by the Geckos and the power supply.

The SIIG board that you referenced would probably solve the low-current problem with parallel ports built into motherboards. I visited the SIIG site. Their documentation was good, but it didn't list source and sink current capabilities of the card, other than to state that the card adhered to standards. I think that we agree that a built-in parallel port might not work properly. But, wiring things to a DB-25 connected is not fun. My old eyes have a hard time seeing solder bridges between adjacent connectors. Buying a proper crimp tool for $125 so that crimped pins can be used instead is about $40 more than a breakout-board. I've built litterally hundreds of cables over the years using DB-25 connectors and making wiring errors was a common problem. One of the more tiresome tasks was to beep out new cables using a meter that beeped when shorts were present. (There was a day when terminals, such as the Wyse-50 and Televideo 950 were connected to computers via an RS-232 cable. Since there was no wiring standard, it was easier to make a custom cable than to try to buy cables ready-made.)

My thought is that wiring a parallel port directly to devices means that I have to have higher skills and knowledge than simply connecting wires to screw terminals. In other words, plugging one end of a parallel cable into my PC and the other end into my breakout board, and then connecting cables between my breakout board and the various devices that I'm controlling seems pretty basic to me with little chance for error.

I haven't ordered your CNC kit yet, so I don't have your wiring chart, but I'm sure that you've written your instructions clearly. Your web site is outstanding and the machines that you demonstated for me at IH show that you really know what you're doing. So, without looking at your documentation, I would assume that anyone following your instructions would have no problem connecting electronics in a basic system. The potential for problems would seem to be adding other input sensors or output devices that might not be opto-isolated to the system.

We agree that the Geckos are opto-isolated and that your limit switches are opto-isolated, so we don't have to worry about them, but, those other devices that people add to their mills that might/should cause concern. A breakout board solves that problem.

Adding opto-isolators to everything might be redundant, but it would guarantee that everything is truely isolated. The danger in that approach is adding extra isolated power supplies for each level of opto-isolation. A big problem, when using opto-couplers is that one side of the circuit has power and the other side is dead. Some people try to get around that probjem by using a common power supply to power everything. Bad idea. That defeats the isolation. Adding a couple of small 5V power supplies is easy. Jameco has inexpensive supplies that are the size of a credit card and less than 2-inches tall.

Building a good CNC system takes a lot expertise from a lot of disciplines. You have the knowledge and experience to do it all. I have some experience in electronics. Others may have more experience or less than I. Sharing experience and concerns might help the unwary get past the gotchas that seem to be lurking just beyond the line of sight, waiting to mess things up.

[SMW Precision]

Building a good CNC system takes a lot expertise from a lot of disciplines. You have the knowledge and experience to do it all. I have some experience in electronics. Others may have more experience or less than I. Sharing experience and concerns might help the unwary get past the gotchas that seem to be lurking just beyond the line of sight, waiting to mess things up.

Richard,

I could not agree more! I have been thinking about this discussion. Too a great degree I think everyone is right. One of the things Aaron did a couple of years ago was design an excellent heatsink system for those Gecko drivers.
I think this may be why operation problems are not occuring.

[wildcat]

When I was reading about isolation I also read that the Geckos and limit switches are isolated. However, the wire leaving the devices are not. For instance, the isolation in limit switches does nothing if a short occurs long the length of wire back to the enclosure. A number of bad things would have to happen for this to become catastrophic but the breakout board seemed to provide cheap insurance.

[IndHobby]

Metal cases for all the “stuff” is pretty expensive, and for the most part an old PC case works just fine. Unless something is reasonably priced folks will use what works, a NEMA enclosure runs about $250 (yea it’s big) a old PC case is about $20.

Insofar as skills to wire one or the other, who knows, I honestly think a direct wire is easier. I’ve done complex, and I’ve done simple. I prefer the simple.

For the most part I get only a few wiring questions so I guess the basic schematic is doing ok:

One funny story:

I had a customer who had an axis running away. We went over possibilities, he checked this and that and it still ran away. I sent a new Gecko and encoder (I think he swapped them both) it still ran away he checked every wire, it still ran away.

After about 7 hours on the phone over a week period he mentioned he was color blind. I asked if anyone else was home. He got his daughter out of bed and brought her down to the shop. After introducing myself I had her tell me the color of the wires in order. After she told me the color of the eighth wire I had it. He had swapped the green and brown wire (Index instead of Phase A). He was running in 30 minutes.

The moral of the story?
It is easy to fix simple, when you have a product from one vendor, connecting to a product from another it gets real messy real fast.

[IndHobby]

Yes Wildcat, if enough catastrophic events occur at the same time all hell can break loose, at which point if they are catastrophic enough they will happen on the isolated side of the circuit, or the other side. Whatever side you are trying to protect. There is some weird law out there which states that eventually the sum of your fears will come true, it is just a matter of time.

On the other hand, the more complex a circuit the higher the possibility of it being its own problem. The goal is simple, simple, simple.

I was electronic warfare in the Army for 6 years. Worked on everything electronic you can imagine. Simple ALWAYS lasts longer than complex.

This is one of those electron flow/hole flow arguments (if you understand that you are a geek) there is no single right answer. Just shades of wrong.

The long and short: On a Saturday afternoon with a Marlboro in one hand and a beer in the other, I can walk someone through a wiring problem and get it solved over the phone. Throw a few breakout boards and it becomes a whole new problem.

And it’s not that I have anything against these breakout boards, they are just not my thing.

[Richards]

Sometimes it's a hard decision knowing what to do and whose opinion should be listened to. I'd rather go with an expensive case ($250) to protect a $10,000 investment than trust to luck with a $20 case. That's just me. I've lived long enough that scrimping on hardware no longer makes sense. There is probably no 'one way' to make things perfect, but I'll almost always go with either 'standards' or 'industrial accepted pactices' as my model. One thing that I've learned over the years is that you can always tell a pioneer by the arrows in his back. In other words, I'd rather be safe (and spend the dollars) than sorry. It all comes down to 'insurance'. I have a rather large life insurance policy to protect my wife and family. I'm betting that I will die before my time and the life insurance company is betting that I will beat the odds and live at least as long as they predict. It's the same way with health insurance. I pay $10,000 a year betting that either I or someone in my family is going to need major medical insurance. So far, I've paid out more than $70,000 in premiums to save $11,000 in actual costs. So, I'm wrong. Perhaps it's my nature, but I tend to play it safe and follow all accepted rules - regardless of the actual possibility for failure. In actual costs, spending an extra $300 to $500 on a CNC upgrade to protect a $10,000 investment seems reasonable to me. Others may think otherwise. Please note that I have the highest regard for Aaron and the machinery that he sells. My opinion on how to hook up that machinery are just that - my opinion. He has proven himself over the course of years and I'm a totally newbie - although I'm a cautious newbie.

-Mike Richards

[Ron111]

Ken,

It's been a good six months since you finished the mill, I also has looked at the enco since it is on sale so how is the mill working out. I would love to see some pic/s and hear about it's performance and any associated issues.

Thanks,

Ron

[SMW Precision]

Originally Posted by Ron111 Ken, It's been a good six months since you finished the mill, I also has looked at the enco since it is on sale so how is the mill working out. I would love to see some pic/s and hear about it's performance and any associated issues. Thanks, Ron

Ron,

Mechanicals are excellent in the IH kit. Where I had problems and got fixes were the electronics. All the bugs are out, Aaron as usual was right don't use the breakout boards. As well as a few other tricks part his and a few of mine.
Listed performance spec's are extremely conservative while the IH system will run literally forever at 105 imp it is capable of over twice that IPM movement rate. Tooling, setup and gcode all play a role in performance.
I personally cannot suggest doing the conversion on a RF45, mine was purchased before the first IH mill shipment and there have been a great number of improvements from that point on the Mill.
Current IH mill is several times more rigid than the RF45 with twice the work envelope. I also ran into problems on Z-axis slide connector to ball nut when cut for ball nut this under sized casting becomes way to thin. Servo motor shattered it like it was not even there. I made a mistake on limit switch setting and replaced it with one about 10 times heavier. Thank you Gene & Tommy for helping to fabricate a replacement. That thin casting also affected Z-axis accuracy. I believe further kits sent out for a RF45 will contain a replacement for that part.

[Ron111]

Ken,
Thanks for the insight, your right it would pay to get the more ridgid current IH mill. Are you where your are happy with your mill and how much use does it get now. But your signature it appears that you may be using the mill for so type of manufacture of a product. I guess that's what a lot of us who read the zone are interested in in whether some of these benchtop mills are being used is some daily use whether it is in producing small proction runs of this or that and who they are holding up on a daily basis. Maybe I should start a new thread and poll the cummunity.

Thanks for your input,

Ron

[SMW Precision]

Ron,

My usage is fairly minor; basic parts runs. Though I know of one shop that just ordered their 3rd machine. They run them daily for many hours doing general CNC machining.
Aarons original prototype mill is still cutting straight with 10 to 15,000 hours of operation. It cut all the IH parts for about 5 years and still is used to cut several speciality parts. There has been a lot of progress from that mill with a stepper Z-axis built from an Enco RF45.

Ken