Messages

11

General Mach Discussion / Re: Mach3 4.0 Rev Update

« on: December 17, 2011, 07:35:06 AM »

Awesome! Surely it will take some time for the new programmer to get up to speed with Mach3, but this is very good news.

I think Artsoft should consider 2-tier pricing for home and business licensing. As a commercial user I would gladly shell out a few more bucks for this software. Think about it:

CAD package - $1000+/license
CAM - $1000+/license
machine controller - $160/machine with free updates?

And all I need for hardware are standard stepper drivers and a BOB? Go ahead and raise the price for us, but let the hobbyist enjoy for the nominal fee. The functionality and customization that Mach3 allows deserves a higher payout from those of us that are saving $30,000+/yr/machine on machining costs when compared to outsourcing the parts. And yes, my shop does save close to or in excess of $100,000/yr with 3 CNC's - no exaggeration. This extra money could then be used to improve and stabilize Mach3, and would be more than worth it to all of us commercial users - assuming that the product does improve as a result.

So, by all means, instead of releasing Mach3 rev4, release it as Mach4 with new 2-tier pricing. I would strongly recommend to the owner of my company to upgrade all of my machines to Mach4, paying the new licensing fee, without even thinking about it.

My 2 cents.

Other commercial users, please share what you think about this. I'm sure I'm not alone on this train of thought. I want Mach to be as good as possible and am willing to pay for it. Do you agree?

12

Newfangled Solutions Mach3 Wizards / Re: Machine jumps when in program and looking at last wizard

« on: December 17, 2011, 07:13:10 AM »

Mach3 takes liberties with the Windows OS. Microsoft designed Windows to be an absolute control freak and it will not release that control unless you massage it in just the right way. As mentioned before, much more expensive controllers use independent hardware to achieve smooth motion and a feature rich environment in Windows. Mach3 is all software, baby! No expensive hardware at all. To be honest, as far as I know Mach3 is the ONLY exclusively software based motion controller operating on the Windows platform. Making it work at all was in itself a great feat.

If you told him a Ford Focus was just as good as a Ferarri, would he look at you funny then. 'Cause it's pretty much the same thing.

Actually, I see it more like a Ford Mustang vs an F1, but with the Mustang, at a fraction of the price, they also throw in a complete workshop to tune it up as you see fit. However, you still need to know or learn how to use the tools to tweak and tune the performance and make new parts to produce a car that is truly drool worthy. In Mach3, the same is true - you get a decent base model and if you're willing to learn (or pay someone) you can have one slick machine. That being said, don't expect it to outperform the F1 on the track, but you can still get way more functionality and enjoyment out of the Mustang.

Feedhold function in Mach3 stops at the next convenient and safe spot to guarantee that you can resume without problems/lost steps/sync issues. Perhaps it could be better with the developers spending more time on it, but I guess it is what it is. Likely this was the easiest way to make it work within Windows.

I don't think people were trying to beat on you, but instead trying to impress on you that Mach3 was designed for the hobbyist - it just so happens that it's good enough to be used in production roles like in my shop - and I wouldn't change to a commercial controller at this point. With Mach3, where there's a will there's a way. With commercial controllers you are locked into the functionality that the developers decided you need. To each his own.

Please don't hesitate to ask more questions about the limitations and any quirks you find - many people have had quirks that are not normal and can be fixed, but if the experts here are telling you it's normal, please resist the urge to compare Mach3 to a commercial controller, as Mach was not designed to be a commercial controller replacement.

13

General Mach Discussion / Re: Tangential arcs

« on: December 17, 2011, 06:40:04 AM »

Thanks for posting this here. I will try it tomorrow!

14

General Mach Discussion / Re: External E stop

« on: December 12, 2011, 05:06:48 AM »

Perhaps a limit switch is active. That would result in only being able to move in one direction.

15

General Mach Discussion / Re: tool compensation

« on: December 10, 2011, 05:35:05 AM »

There is only one way to do this using tool compensation. I have no idea if cabinet vision is capable of it nor how to turn it on, but the CAM program must be able to output the GCode with tool compensation for it to work.

Once you have enabled tool compensation in your CAM program, you can use Mach3's Tool Table to input the diameter of the tool. Note that Diameter Wear and Height Wear are not implemented yet, AFAIK it won't be until revision 4 is released.

The reason the CAM package must be compatible is because Mach3 needs to know what direction to apply the offset in. It accomplishes this by using left and right compensation indicators in the GCode (G41 and G42), always with respect to the direction of the cut (ie with respect to the tool). The alternative is adding the compensation by hand - but this isn't recommended as you need a lead-in move for Mach3 to properly determine the offset direction, and it can be a pretty painful process for anything but a very simple part.

Hope this helps a bit.

16

General Mach Discussion / Re: Tangential arcs

« on: December 09, 2011, 06:22:00 AM »

I also have and do experience this problem, although it hasn't presented a major problem so I haven't pursued it.

Can anyone who knows the inner workings of Mach3's trajectory planner weigh in on this?

Can others confirm whether they also have seen this issue?

In fact, I see hesitations before/after all of the canned cycles. For drilling, it is of no concern, but it does affect cut quality with arcs.

I usually let our production workers run the Mill and Lathe, but I do see it on the router when I happen to run it. I will check the mill for the problem and let you know. The lathe it can be hard to get time to play with - it's too busy, but I'll see if I can check it too.

Anyone with some insight? It would be very much appreciated!

17

General Mach Discussion / Re: Steppers are too slow

« on: December 09, 2011, 06:10:17 AM »

Gents,

Hope I haven't started a forum war here, but I still have a bit of a problem understanding why a good quality Digital multimeter connected in line with the PS output to the drivers fails to read an accurate current flow compared to a hall effect clamp meter. When i look at the chopwave graph Ian posted at post 44, the current trace has fairly minimal peaks and troughs associated with the switch activity. I read this as the current flow through the coil and not necessarily a waveform of the driver input supply. I would have expected the circuitry within the driver would have provided some smoothing to the input current. Additionally with 4 drivers all connected together I would have expected some further smoothing of this wave form as one goes high and the other goes low.

Is my thinking correct here??

Stuart

Stuart,

Unfortunately, it isn't as simple as that. Measuring PWM power is actually a substantial problem. One major problem is the phase angle, which is not taken into account at all by a DMM, no matter how good it is. The second problem is the harmonics generated by chopping the DC signal. You can have measurable power up to the 100th harmonic in some cases, although at the power levels we're talking about here it won't be that extreme, and obviously the power going up the harmonics will drop exponentially. Nonetheless, unless these factors are taken into account you will not get an accurate reading. The best DMM on the market is therefore pretty much useless to obtain an actual measurement.

Power output from a PWM can be expressed as Power = V * I * cos(e). Without taking this phase angle into account, the only real use of a DMM would be to compare motors to motors, or drivers to drivers, but the actual value you get is quite erroneous.

In fact, an analog meter will give results closer to reality than either an averaging DMM or true-rms DMM, because it will respond to low frequency harmonics in much the same way as the motors. It will not, however, take the phase angle or higher harmonic power that is also being used/lost into account. Higher harmonic power, from what I understand, will not be converted to usable torque, but it is still being drawn nonetheless.

After a little more research into the topic, I'm starting to also question the effectiveness of a DC clamp meter for the same reasons as above. While I'm fairly certain it would be more accurate, I'm not sure it would be close enough to actually be considered accurate. A power meter, however, will take the phase angle into account and measures both voltage and current simultaneously. With this method you will be within 1% of real life power draw. This, however, doesn't really help with real time current calculations as voltage is not constant, but could still be useful with the right methods.

A digital storage oscilloscope will do a pretty good job, but you will have to know what you're doing to get useful measurements. Also USB oscilloscopes with decent software would also do a good job, and many cases better because of higher resolution sampling. Standard oscilloscopes are 8-bit, better ones are 12-bit. This is relevant to measuring the harmonics.

This is a complex topic because of, in short, the phase angle and harmonics, and is made worse by the very fast rise and fall times created by modern PWMs. In addition you may also have to contend with reflected power depending on cabling lengths and such, but not likely significant in this application. Unfortunately I'm not well versed enough to say for sure one way or the other, I just know the potential is there.

There is obviously much more to this topic, but hopefully this will be enough to provide some insight into why standard meters are so inaccurate at measuring PWM power.

18

General Mach Discussion / Re: Tangential arcs

« on: December 08, 2011, 05:18:30 AM »

Are you saying that the motors stall completely? Or is it a hesitation?

If you are completely stalling, the first thing to try is lowering acceleration and/or speed in the motor tuning window. If you've already slowed down the feedrate without any difference then I'd start with acceleration. You may be trying to drive the motors beyond the system limitations, which is a combination of your motors torque and inductance, as well as the power supply. Higher voltage power supplies are better able to deal with motor inductance and the back EMF that they generate.

19

General Mach Discussion / Re: Nesting troubles

« on: December 07, 2011, 07:52:05 AM »

Yes, one by one is very tedious. We cut up to 100+ parts at a time on our router. Takes a little while for Mach3 to read through the code for every copy when it's verifying the code, but it does work incredible well and saves us hours upon hours of time. I'll see if I can find an alternative method that bypasses this, but I won't spend much time on that - may not be possible anyway. I started with Olivier's code, but then rewrote it. I will work on removing the couple of bugs I've noticed and give it a once over before releasing it.

It may not be ready until the end of the year - work is pretty swamped right now with products that must ship by end of year, but I have 10 days off between Christmas and New Years, so worst case scenario I can do it then.

If there's a specific need before that let me know and I can email a copy as-is, but I'm not releasing it until I've looked at it again.

20

General Mach Discussion / Re: Steppers are too slow

« on: December 07, 2011, 07:36:18 AM »

Never seen a current meter work as you describe.  A shunt is nothing but a VERY small resistance, on the order of milli-ohms, typically provided by a simple strap of brass or copper, necked down at one point, and trimmed at the neck to provide the required accurate, small resistance.  ALL of the current flows through the shunt, and its resistance creates a small voltage, which is displayed on a moving-coil meter movement with, typically, 50mV full-scale sensitivity.  A DVM works in exactly the same way, except measures the voltage with an A/D converter, rather than an analog meter movement.

Sorry, I was actually thinking about an alternative measuring method when describing the shunt configuration and it wasn't complete by any stretch of the imagination - you are right about a typical ammeter having the shunt in series with ALL of the current. That being said the main point I was focusing on is that a DC ammeter samples DC voltage to infer the current flow, and cannot provide an accurate reading for this type of power signal.

If you want to find out how much current is being delivered to your motors via chopper circuitry (ie most stepper drivers), the DC Clamp Meter is the way to go - and the only reasonable option that I'm aware of - because it makes use of the Hall effect which responds to DC, AC and pulsed DC, as well as complex combinations. That's the point I was trying to make.