Electronic Lead Screw Project

[Brian OldfordParticipant @brianoldford70365]

I've followed this thread from its inception. I plan to drive my Super 7's lead-screw via 90 degree skew gears thus containing the stepper motor completely within the change-gear compartment. To design and make the mounting requires the removal of the banjo but I became side-tracked with other more pressing tasks.

Now I'm back on track I'll try and put up some photos of my progress,

Edited By Brian Oldford on 14/04/2020 18:48:28

[Roy VaughnParticipant @royvaughn26060]

The odd missed step isn't crucial except when screwcutting, but it shouldn't happen, so the root cause (overspeed or overload) needs to be fixed by the operator. A lot of missed steps leading to a crash is another matter. If it's due to operator error that's exactly equivalent to the geared lead screw situation with the same consequences and solution. If it's due to electronics failure I imagine that the banjo will be back on the lathe in short order. The lead screw will track the spindle even after the emergency stop button has been hit, but what if there's a power cut? Put the ELS on a UPS? It all gets a bit complicated. So I agree the software best solution is to signal the alarm on the display and leave it at that.

Roy

Since it should never happen the only corrective action is to fix the source of the problem, e.g. overspeed or overload.

I envisage alarms arising from the odd missed step due to over-speed or overload, hard to notice otherwise and only important when screwcutting . If something major happens, e.g., the lead screw stops while the spindle carries on, leading to a crash, I imaging the whole lot being decommissioned and sold for parts on Ebay!

[JimmehParticipant @jimmeh]

I think its just a placeholder for code to be added at a later date if so desired. When testing my system (CL57T driver + stepper with encoder) I physically forced the servo out of position and it just cuts the power to the servo. Cycling the power restores everything back to normal. Likewise if you spin the rotary encoder really fast one way and then the other the system errors and power is cut. The consequences are obviously bad, but I think I'd rather the stepper powers down if there's a problem.

On a separate note I would like to try and see if I can get the whole system to power on/off with the lathe itself. I am conscious that if there is an electronic fart the carriage could move unexpectedly and I would like the E-stop to help me!

I'm almost done with my installation. I just have to adjust the lathe side cover so it fits back on and mount the control panel somewhere accessible. Its really quiet without the change wheels rattling away, and I actually use it for fine feed now. Before I never bothered as it was too much of a pain in the ass swapping everything around.

[Roy VaughnParticipant @royvaughn26060]

Jimmeh

I have the same setup. I haven't tried those tests. Maybe the behaviour of the stepper controller could be changed so let the software take control by doing something over the data port. Probably in the all too difficult category.

Apologies for the junk in the end of my last email, I can't get rid of it now.

[Johnboy25Participant @johnboy25]

Hi everyone. I’ve been following this build on James Clough’s YouTube channel (Clough42) this is an on going project and there have been updates which I have included & modified in software. There are ‘third parties’ actively adding options they consider to be useful. I have seem mods to include more safety features. This is not a ‘finished’ project and there are safe guard in place. Further info can be seen on github forum. Addition features are planned but James is currently contributing to the Corvid19 response producing PPE for hospital staff which he’s making on his 3D printers.

I have been building this system along with a friend who has a Southbend 4.1/2”. He wants to screw cut metric threads. He’s managed to get his hardware up and running and it has transformed the versatility of the lathe. I’ve been repairing/renovating the gearbox prior to fitting on my Boxford 4.1/2” metric lathe between making parts for the Loco project. I’m using this lockdown period to maximise workshop activities and trying to keep me out of harms way.

Stay well & keep away from this virulent viruse

[Johnboy25Participant @johnboy25]

Jim

This is a mod that has been produced to use the on/off button To disable the  servo Drive.

I also intend to incorporate an E-stop activated by the lathes control system.

Edited By Johnboy25 on 18/04/2020 09:28:44

[Brian OldfordParticipant @brianoldford70365]

After a minor hiccough with the first attempt I have managed to squeeze the stepper motor into the change-wheel housing of my S7. The first attempt left me unable to close the door without modifying it.
Mk2 means the door can be left unmolested so the lathe can be returned to stock if my efforts prove to be a monumental disaster. The next job is to mount and make the drive for the rotary encoder.

 

Edited By Brian Oldford on 19/04/2020 20:01:40

[Phil GrantParticipant @philgrant54580]

Nice job, I need to look at mounting mine at a right angle to the leadscrew.

[ZanParticipant @zan]

Cracking. That’s a great idea. How will you shield it from all the oil dripping off the tumbler? The bottom inside of mine always has a thick layer of oil on it not excessive oiling, just doing the standard when I remember…..

[Brian OldfordParticipant @brianoldford70365]

Posted by Zan on 19/04/2020 21:15:11:

Cracking. That’s a great idea. How will you shield it from all the oil dripping off the tumbler? The bottom inside of mine always has a thick layer of oil on it not excessive oiling, just doing the standard when I remember…..

The tumbler will be removed as stepper reversal can done by electrical switching. If not by the SW, by an additional switch to alter the winding energisation.

[Brian OldfordParticipant @brianoldford70365]

Posted by Phil Grant on 19/04/2020 20:51:12:

Nice job, I need to look at mounting mine at a right angle to the leadscrew.

What machine have you got Phil?

[ZanParticipant @zan]

Lol been in the workshop too long! Ignore my idiot posts! Who needs tumblers with a step or drive?

[Phil GrantParticipant @philgrant54580]

Posted by Brian Oldford on 19/04/2020 21:58:03:

Posted by Phil Grant on 19/04/2020 20:51:12:

Nice job, I need to look at mounting mine at a right angle to the leadscrew.

What machine have you got Phil?

Small Chinese job from Warco Super mini lathe

[Andy CarruthersParticipant @andycarruthers33275]

Hi Phil, I have a WM180 and WM240, sadly not available to work on right now – I am very interested to see how you implement the ELS Mod as I expect I can blag your work, please!

[Phil GrantParticipant @philgrant54580]

Andy,

I think the layout on the W180 and 240 are quite different than my Mini lathe but you're welcome to use what I have, pictures are in my albums, I'm now looking at converting to a right angle drive arrangement to reduce the width of it all.

[Roy VaughnParticipant @royvaughn26060]

I have just got my implementation up and running on the lathe, all is operating as it should. During the build I found that screened cable was needed to stabilise the display even without obvious sources of interference nearby. As the control panel is not yet permanently mounted I can move it around to test its residual susceptibility to interference. I've found that when the cable is passed close to either the lathe's motor or the stepper, there is still occasional interference. It gave me a fright the first time it happened with the lathe turning and the feed engaged, when the system switched itself unbidden from a slow metric feed to coarse tpi. Obviously I will arrange the final installation to minimise the chance of interference but the fact that it could still occur with potentially unfortunate results convinced me that something needed to be done. I've therefore made a small modification to the code to disable keypad input when the spindle is turning. Has anyone else made any safety mods to the software?

[Brian OldfordParticipant @brianoldford70365]

Posted by Roy Vaughn on 02/05/2020 11:33:23:

I have just got my implementation up and running on the lathe, all is operating as it should. During the build I found that screened cable was needed to stabilise the display even without obvious sources of interference nearby. As the control panel is not yet permanently mounted I can move it around to test its residual susceptibility to interference. I've found that when the cable is passed close to either the lathe's motor or the stepper, there is still occasional interference. It gave me a fright the first time it happened with the lathe turning and the feed engaged, when the system switched itself unbidden from a slow metric feed to coarse tpi. Obviously I will arrange the final installation to minimise the chance of interference but the fact that it could still occur with potentially unfortunate results convinced me that something needed to be done. I've therefore made a small modification to the code to disable keypad input when the spindle is turning. Has anyone else made any safety mods to the software?

Perhaps you would be kind enough to publish those code mods at the Github repository.

[Martin ConnellyParticipant @martinconnelly55370]

Mach3 has a key debounce setting to help with this type of issue. You can set a minimum time that a key must be pressed before it is acted on.

Martin C

[Tony Pratt 1Participant @tonypratt1]

Posted by Roy Vaughn on 02/05/2020 11:33:23:

I have just got my implementation up and running on the lathe, all is operating as it should. During the build I found that screened cable was needed to stabilise the display even without obvious sources of interference nearby. As the control panel is not yet permanently mounted I can move it around to test its residual susceptibility to interference. I've found that when the cable is passed close to either the lathe's motor or the stepper, there is still occasional interference. It gave me a fright the first time it happened with the lathe turning and the feed engaged, when the system switched itself unbidden from a slow metric feed to coarse tpi. Obviously I will arrange the final installation to minimise the chance of interference but the fact that it could still occur with potentially unfortunate results convinced me that something needed to be done. I've therefore made a small modification to the code to disable keypad input when the spindle is turning. Has anyone else made any safety mods to the software?

Just my opinion but I like the flexibility of changing thing when the spindle is turning.

Tony

[SillyOldDufferModerator @sillyoldduffer]

Posted by Tony Pratt 1 on 02/05/2020 16:00:01:

Posted by Roy Vaughn on 02/05/2020 11:33:23:

… During the build I found that screened cable was needed to stabilise the display even without obvious sources of interference nearby. …I've found that when the cable is passed close to either the lathe's motor or the stepper, there is still occasional interference. … Obviously I will arrange the final installation to minimise the chance of interference but the fact that it could still occur with potentially unfortunate results convinced me that something needed to be done. I've therefore made a small modification to the code to disable keypad input when the spindle is turning….

Just my opinion but I like the flexibility of changing thing when the spindle is turning.

Tony

+1 to Tony. But also, it goes against my grain to cure a hardware fault like interference with a software fix. It's not mad, but I'd prefer to fix the interference first. There's a hint of a related, but different, cable problem in 'I found that screened cable was needed to stabilise the display'.  Likely the signal is being distorted as well.

The LED&KEY display is based on a TM1638 chip. It communicates with a computer with a fast synchronised serial data stream, which is encoded and decoded at each end. It's data rather than plain switching. In addition to power, there are CLOCK, STROBE and DATA signals. My LED module runs at about 300kHz, a radio frequency that doesn't carry well on ordinary cable, being prone to radiate and receive interference and distort. The application  really needs a data cable.

I suggest the problem is made for an Ethernet Patch Cable and RJ45 sockets.

Ethernet cable contains 4 twisted pairs and a shield.

So Power (+ and -ground) on first pair, STROBE & ground on second, DATA & ground on fourth, CLOCK and ground on the fourth. Then the POSITIVE, STROBE, DATA, and CLOCK signals are each independently balanced against ground to prevent interference and cross-talk. And all four pairs are further shielded behind an Aluminium foil wrap. Carrying high frequency signals on twisted pairs allows them to go further along a cable without losing shape and amplitude.

A standard ethernet patch cable can be bought: plug pin connections described here.

However, at the sockets, the pins are not Ethernet standard. Instead the ground side of each pair should be brought together and grounded at both sockets. Then the other 4 pins are run to POSITIVE, STROBE, DATA, and CLOCK; any pair can be used provided the computer and display ends match!

External RF interference to power and control circuits is almost always due bad practice inside the victim. It's the electronic equivalent of a leaky roof. Better to fix the roof than catch drips in buckets!

Dave

 

 

 

Edited By SillyOldDuffer on 02/05/2020 17:20:04

[Roy VaughnParticipant @royvaughn26060]

Thanks for the information SoD. As an ex-professional programmer (many years ago now) but latterly a user of microcontrollers for hobby purposes, I am familiar with Cat5/6 cabling but never in non-standard (non-balanced) usage you propose or using STP rather than UTP. The idea that solving a hardware problem in software is possible mad comes as a surprise, in the microcontroller world the software often (usually) has to work around and protect against "hardware" artifacts, that is, the real world. For example, contact debounce is simple in software and saves one or two components, which can be worthwhile if you are pushed for space or weight as my projects mostly are. I agree that curing interference at source is preferable and will definitely try your recommendation. However, seeing the lead screw take off in high gear is hard to forget and I will be keeping the software fix in place in case the hardware defence is ever breached.

The fix hardly deserves a mention on Github, it's simply a jump around the input reading statements in the Userinterface.ccp module if the spindle speed is greater than zero. Pm me if you want further info. As well as defending against interference, the fix also protects against finger trouble on those little keys. A refinement, if you really want to alter settings with the spindle in motion but still provide some protection against the worst consequences, would be to disallow change in more specific circumstances, e.g. to disallow switching between feed and thread, to inhibit the up/down buttons unless in feed mode, by making changes dependent on the rpm (I think there is some stuff on Github along these lines), etc.

Thanks all for your thoughts. Off to Ebay now to buy an Ethernet patch lead.

Roy

[Michael GilliganParticipant @michaelgilligan61133]

Posted by SillyOldDuffer on 02/05/2020 17:14:58:

[…]

External RF interference to power and control circuits is almost always due bad practice inside the victim. It's the electronic equivalent of a leaky roof. Better to fix the roof than catch drips in buckets!

Dave

.

Brilliant summary, Dave

MichaelG.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Roy Vaughn on 02/05/2020 19:16:46:

…

The idea that solving a hardware problem in software is possible mad comes as a surprise, in the microcontroller world the software often (usually) has to work around and protect against "hardware" artifacts, that is, the real world. For example, contact debounce is simple in software …

Roy

Hi Roy,

I apologise for 'mad', it was a very poorly chosen word! I was thinking of my dodgy efforts at overcoming hardware problems in software, especially interference, not your valid examples!

Been playing with a couple of different 433Mhz transmitter and receiver modules and Arduino this week and one of the transmitters (claimed 90m range) generates enough RF to upset the Arduino. Scope screenshot shows 3V of noise superimposed on an Arduino 5V digital output pin when the transmitter is sending!

And the el-cheapo 5V wall wart I first used to power the receiver is noisy enough to deafen it.

Someone who shall remain nameless spent a day attempting to fix the software rather than sort out shielding, short wires, and a better power supply.

Dave

[Roy VaughnParticipant @royvaughn26060]

No offence taken Dave. I've ordered the bits for an STP cable to replace my existing screened-but-non-twisted pair cable and we'll see how they compare. I can't provoke any interference at all today so I expect it will be inconclusive, but extra margin would be beneficial.

[Brian OldfordParticipant @brianoldford70365]

A couple of photos of where I've got to with the Rotary Encoder mount and drive arrangements.

 

And a photo of the quadrature outputs on my osilloscope.

Edited By Brian Oldford on 09/05/2020 12:25:28

[Author]

Posts