Denford Triac VMC w/ ATC conversion

Hello,
This is my first post. However, I am not new here. I have been on since 10/2012 reading about all great stuff that you guys accomplish with your machines.

After reading Chaz's Triac VMC conversion post, I decided that I am going to attempt to do mine. Thanks Chaz!

http://www.mycncuk.com/threads/8800-...near#post70146

I have a stepper one (not a servo), but the conversion should be close to Chaz's. The first thing that I did was weld up a stainless steel table on adjustable casters. The funny thing here, with all my great wisdom, is that I first made the table tall enough to fit the control box under it! Unfortunately, my thinking in the beginning was totally off. I didn't realize how bad it would be to work on a step stool to get to the machine. So I bit the bullet and removed the machine off the stand (what a pain) and cut the table legs down and re-welded it. It looked much better the first time. :disturbed:

A pic of the second version of the table...

Attachment 16683

The control box...

Attachment 16685

...so with the machine in place, I started my research on the following:
1) New steppers for X,Y,Z
2) Stepper drives
3) Controller
4) Mach 3 (reading that Mach 4 isn't not quite up to par)

What I have so far are the new steppers. I found ones with 9.5mm shafts that would fit right in. Unfortunately, I didn't check the shaft on the Z axis, which is 12.7mm and keyed. So, I ordered a new pulley to fit the new Z stepper. Given that the shafts are small and are not keyed, I decided it would be better to use a coil pin on the pulleys to hold them in place. I felt this was especially needed because of the weight and torque on the Z axis.

The McLennan RS steppers:
X = 34HSX-208
Y = 34HSX-208
Z = 34HSX-312

Below are the specs...

Attachment 16684

Now I need to find drives that can do microstepping and a good ethernet controller (with sufficient speed and good resolution). I have read many forums posts regarding Mach3/4 with controllers such as CSMIO/IP-S, ESS (smooth stepper), PoKeys57CNC and many more! I have downloaded all the manuals and read them thoroughly. Some are very limited. I think the CSMIO/IP-S would be a good fit. I hope to hear Chaz's feedback on it and others' too.

Any thoughts and opinions would be greatly appreciated.
Best,
--mike

Hi Mike,

Got your PM and I'll gladly advise.

The Cslabs controllers are great piece of Kit and thou expensive worth it IME. Chaz's needs where slightly different to yours how ever because he had DC servos with Drives that would only take Analog signals. So he bought the IP-A. Has it happened he ended up replacing motors/drive with AC servos that would take Step & Dir signals so could have used the IP-S. That said Modern AC servo's also take Analog signals so he could use the IP-A controller anyway. Infact the IP-A controller is the better option because of the abilty to close the Loop all the way back to the Controller rather than just between drive/motor.
This means the Control software knows where the machine is at all times. So if for instance you push the Table while control software isn't in charge it still knows the exact location of the tool in relation to the table at all times because of the Encoder feedback to the Controller.
So In affect you only need to HOME at the beginning of each Session and untill power is removed the Control knows where it is at all times effectively meaning you don't need to HOME after every E-stop or control Reset.

Now with normal Steppers you haven't got this option because you don't use Encoders. Also most Stepper drives only take Step & Dir signals and the IP-S Doesn't take Encoder feed back only the IP-A does and this can't produce Step & Dir signals.
So you have either the IP-S or the IP-M to choose from. Now with Steppers the IP-S is really way Overkill because you don't need all the Mhz pulses it can provide. In performance terms it would be like running NASCAR in First gear all the time.
However it does have some features the Lower Spec IP-M doesn't that would really be useful to Mill owner, like Backlash Comp and more I/O capabiltys.
That said backlash Comp isn't very good on any machine and your better spending money to remove the mechanical backlash than relying on software to remove it. In which case I wouldn't buy the IP-S just for this reason.

The IP-M is more than capable of running steppers to there Max with overhead to spare so in performance terms easy does handles the job.
In I/O terms it's got less to start with and can't be increased using the I/O module like the IP-S/A can. (Or it couldn't but that may have changed with software updates.?)
But it still comes with resonable amount anyway so provided it meets all your I/O needs then for a 3 or 4 axis Mill using steppers the IP-M would be my suggestion.

You still get all the great bennifit of Industrial quality controller with 24V general I/O. Analog I/O and Spindle Speed Control but at half the price. Software is excellent with frequent Firmware updates released. Support is good thou can be slightly slow to respond but when they do often resolve your issues. (which are often your issues not the controller or software.!!)
If your thinking to use Servos then Go with the IP-S or IP-A because you'll need the higher pulse rate they offer. Same if you need lots of I/O because you'll be able to use I/O modules for pretty much unlimited I/O.

Now the Other Controllers like ESS and Pokeys are very capable Controllers but they have what I consider BIG draw back in that they use Breakout board for I/O connection and use 5v logic signals which is prone to noise issues and can lead to all sorts of other head banging sessions.
The Only Breakout board I'd consider using with either of these controllers and Esp the ESS is PMDX 126. Now for us over here in UK or EU the price of importing the ESS and PMDX from States makes the combination more expensive than the IP-M.

For you then this combination is probably the cheaper option. It does give higher pulse rates than the IP-M and slightly more I/O but not 24V thou the PMDX allows for using 24V Prox switches.
In terms of quality and Software/support then nothing comes close to the Cslabs in this price range.

Pokeys is Ok but doesn't come close to IP-M or ESS/PMDX.

CncDrives UC300 is ok controller but USB connection lets it down and it still use 5v logic. Ethernet is much more robust.

Hope this helps on the controller decision.

Drives then I'd say go with Leadshine Digital drives EM806 or AM882. Even thou you live in Gecko territory the Leadshine is better drive IMO.

Jazzcnc,
Thanks for the reply.

It does validate my thinking on the Cslabs controllers. I may have to just go with the IP-S to play it safe in the future.

Also I did look at the PMDX products in my first round of research. They were working on the PMDX-426 that was to be an integrated board with a motion controller and an Ethernet interface. It was also to be compatible with both PNP and NPN 24 signals, but it hasn't come out, yet. Nevertheless, the integration does seem to be the best in Cslabs.

In terms of drives, I did talk to Gecko. They are coming out with a new G214 that has a maximum resolution of 256 microsteps, which is much lower than Leadshine. Also, it does not have differential inputs for the GND signal. So, I am not sure what direction to take for drives at this time. Once, I get all the mechanicals working, I can do more research on drives.

I will be working on the new steppers this weekend to get them installed. Is there a way I can test how bad the backlash is on the machine without having any drives? Can I make a makeshift pulse controller to test it directly to the stepper so I can measure the backlash?

I will post some pics later of the stepper modifications.

Best,
--mike

The Mesa Boards from America are good CNC-Boards as well. But it runs merely with linuxcnc, and the developer of the boards is often in the LC-forum.
And the possibilitys are nearly endless. If you want 100 or more IOs, no problem, they can make up to 9 controlled axes, a mix from analog and step/dir isn't a problem: But the correct installation of the SW isn't a easy task.
The test of backlash is rather easy, you need a dial indicator, adjust the needle to 0, and pull and press the guide of the machine. then you can see how big ist the backlash. I don't think it'll be big,
because afaik the triac have ballscrews.

Quote:

---

Originally Posted by stallone

In terms of drives, I did talk to Gecko. They are coming out with a new G214 that has a maximum resolution of 256 microsteps, which is much lower than Leadshine.

---

The G215 is out already which I think is the super seed of the G214. It's a Max 10 x Ms drive(2000 steps per rev) which also does Sub microstepping at lower speeds dividing each Micro step up further into 32 pieces. This gives equivalent to 320Ms drive (64,000 steps per rev) while at low speeds. When the speed starts to increase the drive kicks in another feature called Morphing. This gradualy switches the drive from Microstepping at low speeds where it helps with motor smoothness back to Full stepping at high speed to increase Torque where Ms isn't helpfull and Robs Torque.
So in affect the G215 range is between 320Ms and Full step over set Speed range.

Now this all sounds great but I can tell you because I had G203 drives fitted on my machine for 3yrs that the Morphing isn't really noticable and makes very little difference compared to good quality Analog drive. I changed to Leadshine Digital drives Both AM882 and EM806 and they Knocked the spots of the Gecko's in every department.
I saw a noticable difference in Smoothness at lower speeds and greatly increased performance at higher speeds. My Max Rapids went from 9mtr/min to 14mtr/min with no changes other than drives. Thou I didn't require those speeds and didn't use the extra speed this still translated into higher torque at 7.5Mtr/min I run at.
Motor heating was considerably less at lower speeds and the whole motor sound changed. This was instantly noticable the minute motors started to spin.

So other than the Sub micro stepping which I haven't experienced yet and if honest I'm sceptacle about how affective they'll be real world use due to most Low/Med quality Step motors inabilty to resolve to such fine Ms amounts. Then I'd still probably say go with the AM882 or EM806 given previous experience of the Geckos morphing being less than impressive. Thou I will say Gecko's service and after sale warranty is second to none, Thou I did kill 2 x the supposedly Un-killable G203Vs.!! They replaced both times and both were out of warranty period.

The Leadshine also have tuneable Resonance thru software which for some machines can be a life saver if affected by resonance at any part of the motors speed curve.

Thou If you want to buy some G215 and send to me I'll gladly check them out against AM or EM drives and compare.! . . . :hysterical:

If you know the leadshines, I have seen, they offer a microstepping of 1000, 2000 and so on.
Do you know, if this works good. My machine have at the moment with 400 Mikrosteps a resolution of
1/100mm or 0,4thou. I must determine, if its metric or inches, But because all screws are metric, I hope the spindles are as well. With 2000µstep, I'll get a resolution of 1/1000mm that should be much better, if I want to make a real circle, or a tight fit.

Quote:

---

Originally Posted by uli12us

If you know the leadshines, I have seen, they offer a microstepping of 1000, 2000 and so on.
Do you know, if this works good. My machine have at the moment with 400 Mikrosteps a resolution of
1/100mm or 0,4thou. I must determine, if its metric or inches, But because all screws are metric, I hope the spindles are as well. With 2000µstep, I'll get a resolution of 1/1000mm that should be much better, if I want to make a real circle, or a tight fit.

---

Depends on the figure your quoting is that 1000 PPR or 1000 x .? . . ie 200 x 1000.
The AM & EM have Max Stepping of 51200 PPR or 256x (200 x 256) They can be set in 2 ways Via Preset values With Dip switches on the drive Ie 400,800,1600 etc or using software in which case you can set to any Ms amount you like.
This parameter Configurabilty esp regards Resonance is what sets them apart from drives like Gecko's etc that leave you stuck with there preset values. In practise 99% of the time you don't need to change any of these parameters and software will Autotune it's self to the motor but in cases where you do they are life savers. Or if you want to tweak the machine to gain the best possible performance from the motors then you can if desired.

Also Don't think of Micro step as increasing Resolution that's not it's main function. It's more useful for motor Smoothness at lower speed. Having the Ms set too high just stresses the Pulse generator for very little to No gain in performance/resolution.
Having Ms set above 16x or 3200ppr is pointless with normal steppers has they won't resolve any hgher than this. It also has a Negative affect on Torque at high speed.
Ms of 8x (1600ppr) is more than good enough for any stepper system using Hybrid steppers and will give reliable stable system. If you want more resolution the either apply a ratio or fit screws with smaller pitch. Trying to Use High Ms will only make your machine unstable and lower peformance.

I know, that µstepping don't enhance the real resolution. But in the case of an exact circle, it'll enhance the circular form. And maybe I can reach a little bit better fit. with my little Emco, I have only
72 ppr, so i have a minimal distance from 0,0277mm in diametre. Thats a bit big if I want a tight fit for ball bearings or similiar parts.
Thats only, because the normal *4 or *8 µStep brings a bit strange results, that can't display on a
standard XYZ-display of a machine.

Quote:

---

Originally Posted by uli12us

I have only 72 ppr, so i have a minimal distance from 0,0277mm in diametre. Thats a bit big if I want a tight fit for ball bearings or similiar parts.

---

That's a strange Number for Stepper motor.? That's 5 deg per step.! Are you sure.?

Yes, thats a 30Year old machine. they used bergerlahr 5fase steppers, with a strange driver as well.
But as long as the machine run without problems, i don't change the control and drive parts of it.
The precision is rather limited, but I don't need often H7 or smaller. In the mill its another thing,
if I want a bore for a bearing, it must be within 1 or 2/100mm. I can't easy grind it to the next tolerance end.

Quote:

---

Originally Posted by uli12us

Yes, thats a 30Year old machine. they used bergerlahr 5fase steppers, with a strange driver as well.

---

Even stranger then because 5 phase steppers give much lower Step angles due to having 10 poles. This is why they were used because the gave greater resolution and lot smoother running because of smaller rotor steps. They also increase in torque when Ms rather than loose torque like 2 phase steppers do.
The down side being more PPR required compared to 2 phase motor.

The first Smaller leadshine Hybrid Servo motor used 5 phase steppers for this reason. 2Nm Hybrid servo motor gives a 3nm 2 phase motor a run for it's money and is lot smoother.

they were 3 phase actually, leadshine dont make 5 phase drivers.

Quote:

---

Originally Posted by Gary

they were 3 phase actually, leadshine dont make 5 phase drivers.

---

Ye had brain fart on that one Gary for some reason I had it in mind they where 5 phase motors but on reflection it was obvious given U V W wrote on top of drive. . :hysterical:

Don't misunderstand me, I want to change the control Unit of My Boxford machine. The Emco with the
strange Motors should be original. And only for the Boxford I need eventually the 2000 µstep from a
leadshine driver.

So I was able to get a little time with the machine this last weekend, but not much. I started to test the backlash on the Triac but I think I was doing it wrong. I am too inexperienced!! However, I did notice the end of the leadscrew (I think it is that) with the pulley, it moves back and forth approximate 2 degrees before engaging. I am not sure if this tells me anything.

Pic below of the end of x-axis pulley.

Attachment 16721

I did have time to measure the position of the stepper pulley on the shafts for the new stepper. The new steppers for x and y are much larger than the originals, but they fit!

Attachment 16722

Thanks for your help,
--mike

you can easily get rid of backlash on a triac by just tightening the tab nut on the end of the table. getting the right preload on the bearings can be pretty tricky.

ive gotten he backlash on my triac down to 0.02mm, which im pretty happy with.

Hi Andy,
Where would I find this tab nut?

Quote:

---

Originally Posted by andy_con

you can easily get rid of backlash on a triac by just tightening the tab nut on the end of the table.

---

...and what steps would I take to adjust it correctly. ...without screwing things up! :whistle:
Thanks,
--mike

you remove the pulley to reveal the tab nut and tab washer.

fold down the tab washer leg which stops the tab nut from rotating, then you can tighten tab washer.

only way ive found to do it, is to tighten put back together and check back lash, if still there remove everything and tighten even more. back lash should be checked using a test dial indicator

you remove the pulley by undoing the grub screw located in between the teeth

http://i187.photobucket.com/albums/x...ps9cep1rho.jpg

Quote:

---

Originally Posted by andy_con

you remove the pulley to reveal the tab nut and tab washer.

fold down the tab washer leg which stops the tab nut from rotating, then you can tighten tab washer.

only way ive found to do it, is to tighten put back together and check back lash, if still there remove everything and tighten even more. back lash should be checked using a test dial indicator

you remove the pulley by undoing the grub screw located in between the teeth

http://i187.photobucket.com/albums/x...ps9cep1rho.jpg

---

Ye, that's what my poor PM was trying to explain.

Anywhere where there are angular contract bearings requiredpreloaded and there are two in the end green part of the table

http://i187.photobucket.com/albums/x...psrkjko7jt.jpg

Thanks guys.

This is great. I understand it now. Basically, I am adjusting the ballscrew tighter by the tab nut and tab washer!

Attachment 16723

Not really, you're adjusting the preload on the pair of angular contact bearings in the green block that fixes the ball screw to the table.

Quote:

---

Originally Posted by i2i

Not really, you're adjusting the preload on the pair of angular contact bearings in the green block that fixes the ball screw to the table.

---

...the blue ones in the illustration below. I was trying to figure that out.

Attachment 16724

the drawing doesn't really represent the action, but that's where it is.

You cannot adjust the ball screw bearings, they are what they are.

When you have an angular contact bearing in your hand the middle wobbles back and forth.

So what your almost doing when you have a pair is pulling the middles together to get rid of that play.

Quote:

---

Originally Posted by stallone

...the blue ones in the illustration below. I was trying to figure that out.

Attachment 16724

---

This drawing is not correct as it is not showing the inner part of the AC bearing

The spacer should be on the inside race so it shows the lock nut squeezing the outer races together....if that's how its assembled? I thought they normally had the lock nut on the inner race with the ballscrew also on the inner race and the outer races locked in position?

The tab nut tightens the inner race of the bearings not the outter