[boydage]

Ive wired my one up to the letter as per manual. Went through all the jog startup stuff no problem.

Gear ratio is set. Currently have it SIG/PUL through a couple of 2222 transistors and they are checked out as working. Im just having trouble working out the DI stuff, especially the pulse command inhibit input function that I think is the issue possibly as the motor shaft is free to spin. A free spinning shaft means to me its not looking for a pulse at all almost like its in the wrong mode but I cant find any other settings.

So a quick question asking is someone familiar has time to give a real fast rundown on how the DIs are set, and if anyone knows about how this pulse inhibit command is sorted? I dont understand the logic at all nor the tables in chapter 7. Gee it would be nice if it was like a VFD with the setup, I cant even understand the literature and I am definitely not a novice engineer. Well. I am a novice at setting up a industrial servo drive anyway.

I have a programming cable on its way but it would be nice to at the very least get this motor to control through my system which is Mach 4 ESS.

Many thanks in advance. At least the motor ran nicely on jog-test-mode. Definitely a step up. Just it is not even looking at my inputs for now.

[joeavaerage]

Hi,
I use Delta B2 series servos. I have one Enable input, an Alarm Reset input, and an Alarm output....so two DI's and one DO.

The drives come from the factory with a number of default DI's including an Emergency Stop....and it's a real pain in the arse. If you do not have wired input to take the DI high, it Estops the servo
indefinitely. No amount of resetting the drive will help. So first of all get rid of it, that is to say program it 0x00.

Do you have a programming cable and the Delta setup and tuning software? If you don't then get it, you need it. If you don't have one, just wait until you do or borrow mine. Programming the drive by pushing buttons is a recipe
for disaster, don't bother.

If you want I will email one of my parameter files and that would allow you to clone a drive. That would at least set the Di's and the one DO.

Craig

[joeavaerage]

Hi,
the first pic is from the manual and shows the two DI's you need.

In Delta's terminology they are SON, a servo enable input, and ARST, an alarm reset.
Note that the setting value is 0x01 and 0x02 respectively.

I have programmed just one digital output ALRM, a servo alarm.

But when you go to enter the the parameters you find that the settings are 4 hex digits, eg: 0x0101

What gives? The extra two digits in front of the 'setting value' determine whether the drive interprets the pin as active high or active low per attached pic.

The last pic is the parameter settings, well a subset of, from the X axis drive of my machine. This is saved on my PC as an .PAR file, which Delta's software reads and populates the parameter tables, very
convenient.

Note that I have used the default pin for the DI's SON and ARST, DI1 and DI5 respectively, so parameters P2-10 and P2-14 respectively. Note that all the unused DI's DI2,DI3,DI4,DI6,DI7,DI8 and DI9
are all programmed 0x0000, ie disabled. Note also that I have programmed bot inputs to be active low as signified by the leading '1' in the parameter setting.
I have just the one DO active, pin location DO1, parameter P2-18. All the rest of the DO's, DO2,DO3,DO4,DO5 and DO6 are programmed inactive, ie 0x0000. DO1 is active low as signified by the leading
'1' in the parameter setting.

Craig

[joeavaerage]

Hi,
just as a matter of interest, the screen shot is just a fraction of one page worth of parameters, and ther are five page, P0, P1 thru P4.

Pages one and two (P0,P1) each have 77 parameters defined, page 3 (P2) has 72 parameters, page 4 (P3) has only 11 parameters and lastly page 5 (P4) has 24 parameters,
or a total of 261 parameters. Programming all of those by pushing buttons on the front panel is not a good idea.

Download the Delta software. If you cant find it let me know and I'll post you the OEM software CD that came with my programming cable. I seem to recall I had to try many of the slightly different
versions before I found one that matched my PC.

Sometimes it can be difficult for the software to establish a communication channel to the drive. Once the settings are basically correct thereafter AutoDetect usually finds the correct port and channel promptly.
Thereafter reading from the drive and/or writing firmware to it is no trouble.

Craig

[boydage]

Hi Craig. Awesome. TBH it was you who recommended this system to me and its fantastic. Many thanks for the response I managed to sort it all out. I was a little too rapid to run past the input wiring diagrams. Hey if you would share your perameter file I would be most appreciative? I totally get you will have spent some time on setting yours up to maximise the finctionality, if I can benefit from that I will be in your dept all the way from Nelson.

This is a servo for a 1/60 4th axis. Using a 1.5kw Delta motor. I would be real keen to get your input on what you, would setup for a 1/60 geared ratio, steps/rev with the 160000 encoder on the motor? Ive been playing around to see what the best relationship is between Mach 4 to the drive. Currently I seem to be getting a #9 alarm if I set the speed up too high. Looking at what is actually happening, it appears the motor is unable to keep up, for a period after removing my finger off the jog button the motor continues spinning whilst it catches up it you know what I mean.

So much to consider. Oh. Im not fussed this is a public forum my email is [email protected]

[joeavaerage]

Hi,
so I presume that the servo is rated to 3000rpm. The maximum rotational speed of the A axis will be 3000/60 =50rpm.

If you were planning to put a chuck on the fourth axis and run continuously rotating tool paths, they could take a while to complete. For instance, let's say you have a part which approximately rotationally symmetric
and you devise a tool path which requires the part rotate 500 times. The tool path will take ten minutes to run no matter how much or little material you need to remove. This is the downside with high reduction rotary axes
when you want to run continuous rotation tool paths. High rotational speed of your rotary axis is very desirable for continuous rotating tool paths.

The contrary issue is resolution. My fourth axis is a trunnion table. When I'm doing a three axis tool path I require the trunnion table to be dead flat. In fact over only as little as 100mm and you might require that the
tilt of the table be 0.01mm or less over that 100mm, then the table must be set to within 20 seconds of arc. If you wish the table to be flat over say 300mm then for the same 0.01mm then the table needs be position within 6 seconds of arc.

This is the two contradictory requirements. To achieve highspeed you must sacrifice resolution, but in those instances when you want high rotational accuracy then you need high resolution which means lower speed.

Lets start with the servo tuning numbers and Mach4 settings you might make if you were wanting best possible speed, ie you want the servo to go at its max speed of 3000rpm, or 50 revs/second.
You have adopted single ended signaling which means you can signal the drive at 200kHz max, according to Delta specs. Note you really should be using differential signaling because then you could use the full
500kHz that is available. You'll get tired of single ended signaling very quickly. The number of step per revolution can be 200,000/50=4000 Steps/rev or one Step is 5.4 arc minutes. The resolution of the fourth axis
is 55.4 arc min /60 =5.4 arc seconds which is not too shabby.

So the Mach4 setting need to be something like:

Sep/Unit =666.666666 (from the above we have the resolution of the fourth axis is 5.4 arc second per step or 3600/5.4=666.66667 steps per degree)
Max velocity= (50 x 360) = 18000 units/minute
and try the acceleration at say 1800 unit/sec²

Now we need the numerator N and denominator D to program the drive to give us 4000 Steps per with a native encoder resolution of 160,000


4000 x N/D =160,000
so N/D = 160000/4000
=40

So the simplest is let N=40 and D=1, but equally N=80 and D=2 would work as would N=1560 and D=39

Ok we have found some Mach4 settings and a suitable numerator and denominator for the drive. This will allow max speed of the servo and therefore the axis. But what happens if we don't really want or need
high axis speed, say for a trunnion? Then we want good resolution and will take whatever speed we can get. Lets say we want to be able to set the axis to one arc second. Given the reduction that would mean that we
require the servo resolution to be 1 arc sec x 60=1 arc min. There are 60 x 360 = 21600 arc min per rev. So the servo resolution has to be 21600 steps per rev.

Our max signaling speed remains 200kHz so the max number of steps/second = 200,000/21600=9.26 revs/sec or 555.555 rpm Note how we want high resolution but that has meant that we cannot run the servo at
maximum speed WITHOUT exceeding our signaling limit.

Mach4 setting:
Steps/unit =3600 (from our target of 1 arc second of the axis per step or 3600 / 1= 3600 steps per degree)
Max Velocity=555.5555/60 x 360=3333.33333 units per minute
and try acceleration at 333 degrees/sec²

Now we need to find the numerator and denominator for the drive.

21600 x N/D= 160000
so N/D =160000/21600
=7.407404407407

We will have to work a bit harder to find two suitable integers that will work.
I simplified the calculation to 7.407407407407=7 22/54
so:
N= (7 x54) +22=400
D=54

We had to work a bit harder but a few minutes with a calculator and the answer is plain.

So now we have two solutions that you can explore. On solution is if you want max speed but with a modest resolution of 5.4 arc sec.
The other solution is for better resolution of 1 arc second but limit the servo speed to 555.555rpm (alternately the max axis speed is 9.26 rpm)

You have to decide what sort of toolpaths you want to run and that may incline you to one solution rather than the other. Either way the balance of speed and resolution must be struck.
Note also that the max signaling speed is a significant property of the drive that sets the boundaries of what can be obtained. It makes sense therefore to use differential signaling
as it extends those boundaries.

Craig

[joeavaerage]

Hi,
I calculated the N an D at home last night on the calculator, and it is correct but this morning I did a simple excel spreadsheet I found a better solution.
The N and D I calculated at home was 400 and 54 respectively. But the Excel calculation shows that the Lowest Common Denominator is 200 and 27 respectively.

400/54 x 21600=160000
200/27 x 21600=160000

Craig

[joeavaerage]

Hi,
I have just detected and corrected an error in my assignments of the A and C axis drives.

As the drive comes from the factory the ALRM output is assigned to DO5 with a value of 0x0107. My X, Y an Z axes are correctly programmed such that should an alarm occur the output DO5 goes active.
There is a wire from that output back to my breakout board, and that in turn is detected by the ESS and Mach4 , Estops the machine and lights the on screen LED to indicate which servo has faulted.
This has been working now for 2-3 years.

Just recently I've added a fourth and fifth axis and I did not have the ALRM output correctly assigned to the DO5 output. As all my IO cables are identical this meant that my breakout board never actually saw the
alarm and therefore Mach didn't know about the alarm either.

This morning I deliberately caused my C axis to fault, ie prevented it from rotating such that the servo overloaded. The drive certainly faulted '006' but Mach4 never saw it. What gives??? This is what has caused me to investigate
and find my drive programming error. The drive's, both A and C now signal an alarm and Estop Mach4 should they occur. Small details can trip you up. I have updated all my parameter files should I ever need to clone a drive.

Craig

[boydage]

Hi Craig,

Last week I responded with a lengthy reply to see it disappear off my screen this morning. Must have missed the "reply" button. Well in a nutshell its all working and many thanks for the input. I am yet to use the axis for real, that will happen this weekend. I have a slight bit of thought about if I was to end up 6 feet under the ground. What my family would do with all my crap in my workshop. A CNC router that probably has cost over $30k to build - that would be sold for perhaps a grand. So I have started writing as built wiring diagrams. And as you will know they can become lengthy. Have you done the same?

Oh - I was going to ask if you would share what you are manufacturing that has such a tight tolerance of 0.01mm? Ive only seen tolerances like that in gas turbines so am curious. Anyways. Once again, many thanks for the words.

[joeavaerage]

Hi boydage,

So I have started writing as built wiring diagrams. And as you will know they can become lengthy. Have you done the same?

No, I draw the plans when I'm finished building it. What the point of having plans beforehand....you don't know how its going to turn out yet!

I was going to ask if you would share what you are manufacturing that has such a tight tolerance of 0.01mm?

None really. My PCB's have smallest features and separation between features of 0.2mm, so the repeatable accuracy needs to be about four to five times better or 0.05mm
in order to do a good job. The Z axis needs to be repeatable to 0.01mm in order for the copper to be cut and yet not so deep and to dig into the fibreglass underneath.
I vey seldom make parts that require 0.01mm tolerances, but that is what I aspire my machine to do.

Craig

[boydage]

Cool! I am guessing you are electronically trained? Absolutely no way I would remember what I have built. Relay on relay on connections.... I need to draw it all up prior because of my shocking memory.


Hey this Delta 1.5kw motor had a lot of AL009 when I initially commissioned it. One thing I did in desperation was increase the deviation setting until it stopped alarming. And infact I ended up going full scale 160000000 to see if it worked. Last thing done at night but it managed to run without AL009 and I forgot I had left it so high. My mistake. Thats the background when I say my 4th Axis loses position intermittently. It did last night on a 90mm part that I am flipping 180deg. As you will already know a 45mm depth of cut even 1 deg will leave a shocking result let alone 9deg (approx) did last night. Ruined the part to be frank. But, intermittant is the issue I have also successfully machined parts with this axis nice and accurate. Acceptable infact.

I pulled the deviation setting back down and decreased the speed but unfortunately with these settings I either get it to run and its not accurate, or, it wont move and just AL009 alarms. This 4th Axis is tight, but a Nema 34 Stepper can move it no problem so I thought a 1.5Kw motor would have no trouble. So where too from here? I honestly have no idea what to do and am reaching out for some ideas. Apart from the possibility there might be a defect, Craig are you better schooled up on increasing torque or current to assist positioning?

For info I have it set for speed and when it runs it goes very well. The higher resolution was a little slow. I may look at that again tonight but ideally this motor would run to specification, is why I purchased it over the Nema 34 Stepper. What would you advise?

[joeavaerage]

Hi,
I call the setting 'following error widow' which I think rather more descriptive of the setting. Should the servo lag behind the commanded position by more than the 'window' it faults.
It is easy to imagine if the commanded position is advancing by 3100rpm but the servo is maxed at 3000rpm then the servo will lag behind and naturally fault out.
Reducing the demanded speed is a good way to give the servo time to stay within the error window. The other one is acceleration.

Lets say Mach calls for the servo to accelerate at 10000units/sec² when the inertia of the servo and fourth axis is such that it can only accelerate 8000units/sec².
Clearly you need to limit Mach's commanded acceleration to a value that the servo can deliver.

In a sense this is very much like tuning a stepper. Start with velocity and acceleration very low and keep upping the speed and/or acceleration until you find a combination that
stays within the error widow.

Craig

[boydage]

Hi Craig thanks for that. Your explanation was exactly what I thought it was or close anyways. Ok I will take another look tonight and have a bit of a play. It was fantastic when it was accurate. All that went away when a reasonably large part was destroyed past limits last night. The scrap alloy bin got another addition.

[joeavaerage]

Hi,
you have to go slow, not the servo, but Mach. You tune the servo for as much speed and torque as you can but in Mach you require a much lower speed (units/minute)
and even lower acceleration. That will allow the servo to keep pace and stay within the error window.

You may recall I advised you in the early days that a high reduction fourth axis can be a double-edged sword. Yes, you get very high torque authority with a high reduction but then get less speed and your
toolpath needs to slow also. Conversely with a lower reduction the fourth axis is faster and then coordinated movements between the fourth axis an the remaining linear axes are quicker. There is a balance to be
struck.

If you are doing 3+1, ie an indexing fourth axis then a slower rotary axis is less troublesome, but genuine simultaneous four axis is a pain with a slow rotary.

Craig

[boydage]

Gee I thought I was being cautious like I am not really a speed king.

My motor settings in Mach if we go left to right to save the typing are 666.666667 8000 140 and if I run it through a bunch of positioning moves it will lose about 0.5deg on the 4th axis. If I increase velocity to 12000 I lose a full 2deg on this code. Craig am I being too ambitious or is it possible I have a fault? As mentioned it is ok sometimes, but, I ruined a part last night because of this. I guess its not performing much better than a Nema 34 I had on this axis, infact the Nema didnt miss steps. Im a little stumped tbh

m11
g0b0
g0b90
g0b270
g0b180
g0b70
g0b0
g0b180
g0b90
g0b0
m10

[boydage]

Hi guys
Interesting discovery! It appears there might be a discrepancy between the manual calculation and the Excel spreadsheet results for the Lowest Common Denominator (N and D). Double-checking formulas and inputs could help identify the source of the variance.
[Spam link removed by Moderator]

That a fact? Would you care to share your findings? I have been playing around with a lot of setting options all the way through the resolution range and have had the same results. The 4th Axis will not return to zero after a series of movements. I have tried running it super slow with a super slow acceleration. Even at one stage it was set so slow I walked away whilst it was rotating through several positions. Only to return to find it was 1deg out.

Today I am going to look at everything. Mechanical is going to be a focus to see if that is my problem because I am out of ideas. I guess the issue for me is it has at one stage operated perfectly - or, perhaps that perfect result was a random success that I dont know now. I have a bit to go yet but am actually considering ripping the Delta off and putting a Nema 34 back on. Thats always in the back of my mind that what is the possibility the Delta is faulty? Surely not. Mechanical aside I just cant work it all out but now is the time to do some proper troubleshooting I guess. Just sucks because I thought it was setup fine and started using it for a job I have been putting off. (groan) (need a hug now lol)

[joeavaerage]

Hi boydage,
you are confusing the two calculations I did and mixing them together.

I did one calculation for maxium speed but lower resolution. The other was for higher resolution but required a lesser speed.

HighSpeed/LowResolution:
Mach4: Steps/unit =666.666666 steps per degree MaxVelocity =18000 degrees per min. Servo: N=40 D=1

HighResolution/LowSpeed:
Mach4: Steps/unit= 3600 steps per degree MaxVelocity=3333.33333 degrees per minute Servo: N=400 D=54

(Alternately but equivalently N=200 D=27)

Choose one solution or the other.
The high resolution solution is probably what you want. That is every one Step pulse from Mach4 rotates the fourth axis by one second of arc

It appears there might be a discrepancy between the manual calculation and the Excel spreadsheet results for the Lowest Common Denominator (N and D)

Then post it.

I'm also wondering if you are missing steps by virtue of not signaling the servo fast enough. You said earlier that you are using open collector signaling. I have personally tested Delta B2 drives
with open collector signaling and it worked OK to about 230kHz. I decided that was insufficient for me and use differential signaling, and that is perfect to at least 500kHz.

Just to be on the safe side back off the MaxVelocity in Mach4 by 50%. That should obviate any signaling error.

Craig

[joeavaerage]

Hi,
should it prove that by reducing the MaxVelocity in Mach4 solves or at least improves the accuracy then that means that your signaling to the servo is faulty.

What breakout board are you using? What is the bandwidth of the signal output?

To be honest I think with open collector signaling you will always struggle, 200kHz with an open collector is bloody fast and always will be prone to errors due to stray capacitance
of the cabling. These servos were designed and intended for differential signaling.

Craig

[joeavaerage]

Hi,
incidentally can you confirm that the reduction ratio is exactly 60:1?

Craig

[boydage]

Hi Craig. Sorry, its not me confusing the numbers I totally understood them, I was keen to hear Kokufms results though because all my calcs were the same as yours and one thing I know is to hear everyone out. Infact I found some better numbers after a bunch of work trying to find a match between the two in Excel but the deviation is the same always normally about 2deg which is huge really that's a 1/3 motor turn - massive.

My problem is no matter what I am using, whether its high resolution or low, whether the acceleration is high or extremely slow, I am still not getting the Axis to return to zero after a bunch of positioning moves. Its 0.5 to 2deg out every time even with the high res low speed options you gave me. So that's that part.

I am using a standard Chinese BOB that I have been using for years, well, two of them connected to an ESS. Craig I have never actually put thought to the BOB and its specs, I am not electronically trained, and they have just "worked" fine for many years. You have made me take a look and I see the specs of one that is the same "1x CNC USB MACH3 100Khz Breakout Board" 100Khz appears low? But then if that's the case, then slowing the speed and velocity WAY down like I have should have resolved my problem yes? It did not though so that was a bummer.

Ok. Some robust troubleshooting I would like to do today. Normally my day off work but I keep getting work calls for faults around the country. But I was going to double check the mechanical environment today. And yes I am wondering about the signaling through a couple of transistors although my colleague mentioned the ones I am using should have no trouble with signaling speed. I will check my earthing and shielding today too as it sounds like I need too. I am pretty anal with earthing though.

But still, slowing it all down should have sorted that.........

Can you tell me more about differential signaling and how I would go about it for my setup? I understand how it works but that's a very light understanding. Do I need to do some shopping Craig? AliExpress again ha