PDA

View Full Version : Inductive Limit Switch problem



Kitwn
29-03-2020, 10:36 AM
I've just upgraded from using microswitches as limit/homing switches to LJ12A3-4-Z/BX type inductive proximity sensors. I'm getting some spurious trips from the Z-axis sensor due to interference from the spindle, a Huan Yang 2.2kW water cooled unit fed from a VFD. The spindle and sensor wiring go through the same drag chain for part of the run and I'd like to avoid having to separate them if possible.

One of the symptoms is that the red LEDs on the relevant sensors glow visibly, but very dimly, when the spindle is running. A ferrite ring on the sensor cable makes no difference, neither does a 1uF capacitor between ground and either the +V supply or sensor output.

Has anyone else seen this faint glow before and, if so, how did you fix it?

All help gratefully received.

Kit

Muzzer
29-03-2020, 11:16 AM
Are the wires from the VFD output to the motor shielded in any way?

JAZZCNC
29-03-2020, 11:28 AM
Has anyone else seen this faint glow before and, if so, how did you fix it?

All help gratefully received.

Kit

I've seen this before when the voltage is close to the minimum 6v. I've also seen it when wired wrong but can't imagine you've done that. Another thing that often happens is the switch can be marked up wrong ie: NPN when should be PNP or NC when NO.

What voltage and breakout board are you using.?

Voicecoil
29-03-2020, 01:02 PM
Are you using the cable attached to the sensors? When I was wiring up my machine I noticed that the cable wasn't errrm exactly up to the standards Auntie Beeb would have liked, so extended with screened twisted pair: not had any problems at all, even running the spindle up to 900Hz (12.5KHz carrier)

Kitwn
29-03-2020, 01:12 PM
I've seen this before when the voltage is close to the minimum 6v. I've also seen it when wired wrong but can't imagine you've done that. Another thing that often happens is the switch can be marked up wrong ie: NPN when should be PNP or NC when NO.

What voltage and breakout board are you using.?

The VFD cable is screened, earthed at the VFD and connected to the spindle body. Just writing that makes me wonder if there's an earth-loop because of that. Must try removing the spindle from it's bracket.

It all worked fine before I started the spindle! I'm using a basic cheapo Chinese parallel BO with diodes to isolate the 5V on it's inputs from the higher sensor voltages. I originally installed the sensors with 12V and have tried 24V but still get the same slight glow on the LEDs.

I have known issues with the BO inputs not being pulled down far enough by the sensors. I can fix that, it's the slight glow and my inability to fix it with filtering that worries me most at the moment. I was wondering if it was a known issue. If somebody is getting the same symptoms but still has the sensors working reliably I'll be happy.

JAZZCNC
29-03-2020, 04:02 PM
The VFD cable is screened, earthed at the VFD and connected to the spindle body. Just writing that makes me wonder if there's an earth-loop because of that. Must try removing the spindle from it's bracket.

That would be my first move. It's a much debated thing but I never ground both ends and I never have issues with these cheap sensors. Thou like voice coil mentions I cut the shitty cable off close to the sensor and use screened back to the board and ground it.

Cube3
29-03-2020, 09:49 PM
I had very similar problems until I did this.

https://www.youtube.com/watch?v=DJfiOqaeFDg

Simon

Voicecoil
30-03-2020, 11:45 AM
It all worked fine before I started the spindle! I'm using a basic cheapo Chinese parallel BO with diodes to isolate the 5V on it's inputs from the higher sensor voltages. I originally installed the sensors with 12V and have tried 24V but still get the same slight glow on the LEDs.

.

Mmmm, just had a thought..... any chance you could post a little pic of the exact arrangement please? The way I'm reading the above, you're using only the bottom 5V of the voltage swing, and hence will only get the noise immunity of a 5V system - having a series resistor or resistive divider would be better. However I may be reading it wrongly!

Kitwn
30-03-2020, 01:07 PM
That would be my first move. It's a much debated thing but I never ground both ends and I never have issues with these cheap sensors. Thou like voice coil mentions I cut the shitty cable off close to the sensor and use screened back to the board and ground it.

That will be test one. How many layers of cling-wrap do you think I'd need round the spindle to insulate it from it's bracket if that works?

Interestingly I get the same amount of glow from two sensors. The Y one which goes to a junction box and onto screened cable before joining the VFD cable in the drag chain, and the Z which has the original cable running in a common length of drag chain before being connected into the same (8way) screened cable as the Y.

Voicecoil,
I do take the point about cutting off the original cable short and going screened, I shall probably do that anyway. Possibly with some balanced audio cable if I have enough, it's thinner than the other stuff I have in stock and Auntie would surely approve.

Simon,
I'll have a look at the full video tomorrow, but I've tried winding a few tuns of the cable onto a ferrite ring at the sensor end and adding capacitors as described earlier at the joint into the multicore. Much to my disappointment it made no difference. Next will be to hook up my oscilloscope and look at the noise level with different solutions.

I'm between 12 hour shifts tonight (and very glad to be still working) so it will be Wednesday before much more gets done on this. Thanks for all the input everyone.

PS I have this cunning design for a low cost DIY optical fibre interface that would solve all these problems but I really would like to get the machine upgrades finished right now and actually make something with it!!!

Kitwn
31-03-2020, 01:19 AM
Funny how the brain works. There I was, lying in bed reading a book when I suddenly realised what an idiot I was for suggesting screened audio cable for this application. Not enough wires!

JAZZCNC
31-03-2020, 09:42 AM
That will be test one. How many layers of cling-wrap do you think I'd need round the spindle to insulate it from it's bracket if that works

2 wraps should do it, don't want to waste it you'll need it for when the Loo roll runs out.!

Voicecoil
31-03-2020, 10:52 AM
Funny how the brain works. There I was, lying in bed reading a book when I suddenly realised what an idiot I was for suggesting screened audio cable for this application. Not enough wires!
It depends on the polarity of your switches and how you're wiring them. NPN NO in parallel (possibly not the optimum choice) could work fine with twin screened, likewise PNP NC in series could work, as both have a commoned 0V - assuming your 24V ground is connected to chassis somewhere of course. Otherwise there's plenty of smallish 3 & 4 core screened cables available, I've a few bits of Mogami 2790 which have come in useful.

Kitwn
31-03-2020, 01:36 PM
2 wraps should do it, don't want to waste it you'll need it for when the Loo roll runs out.!

I'm really not letting my imagination go there! A DIY bidet is more likely.

Kitwn
31-03-2020, 01:54 PM
It depends on the polarity of your switches and how you're wiring them. NPN NO in parallel (possibly not the optimum choice) could work fine with twin screened, likewise PNP NC in series could work, as both have a commoned 0V - assuming your 24V ground is connected to chassis somewhere of course. Otherwise there's plenty of smallish 3 & 4 core screened cables available, I've a few bits of Mogami 2790 which have come in useful.

I'm really not happy with the idea of powering up electronic devices on the 'just in time' principle that series connection of these sensors involves. In practice the four I have are all independent anyway, X1, X2 , Y, Z.
The ones I have are NPN NO which seem to be the most common on eBay. I'm not sure if I have much 4-core screened left but have plenty of 8-core. At present I'm running the supplied 3-core cables to a junction box on the back of the Z axis where the Y & Z cables connect into an 8-core screened cable back the the controller.

dazp1976
31-03-2020, 02:42 PM
When I come to do mine in the future I'll wire them in series for each axis and use solid core CAT7 cable.
Where I have to split it down from 8 wires into pairs as I get closer to the machine I'll heatshrink them.
CAT7 has each pair shielded already. It's just a case of carefully getting the outer sheath off without damaging the foil on each pair.

Atm I'm just using the soft limits.

Neale
31-03-2020, 07:18 PM
I'm really not happy with the idea of powering up electronic devices on the 'just in time' principle that series connection of these sensors involves. In practice the four I have are all independent anyway, X1, X2 , Y, Z.
The ones I have are NPN NO which seem to be the most common on eBay. I'm not sure if I have much 4-core screened left but have plenty of 8-core. At present I'm running the supplied 3-core cables to a junction box on the back of the Z axis where the Y & Z cables connect into an 8-core screened cable back the the controller.

As per the previous post, combining these switches in series isn't a probem as you combine the switches at each end of the same axis - don't mix axes. As long as you have each axis (including A and X if it's a dual-axis machine) going to a separate input, that also lets you home more than one axis at a time. You need to do that anyway if you have a dual-axis machine and want proper gantry squaring on homing.

Personally I run spindle, motor, and switch cables all in the same cable chain; they are all CY and I use 24V signalling which helps with noise rejection. Don't have noise problems.I do have a solid earth wire running through each cable chain connecting the two end parts to avoid any need for continuity via the bearings. Haven't noticed the LEDs glowing. I did wonder if you had bad switches; I bought a cheap box of 10 from eBay when I built my machine and one was the wrong type (only had two wires coming out which was a bit of a giveaway), one was bad when I fitted it, and one failed shorty afterwards. Been OK since, though.

cropwell
01-04-2020, 12:49 AM
If you are looking for 3 core shielded you might find an old USB cable which is shielded (some are, some are not).

Kitwn
01-04-2020, 01:19 AM
When I come to do mine in the future I'll wire them in series for each axis and use solid core CAT7 cable.
Where I have to split it down from 8 wires into pairs as I get closer to the machine I'll heatshrink them.
CAT7 has each pair shielded already. It's just a case of carefully getting the outer sheath off without damaging the foil on each pair.

Atm I'm just using the soft limits.

Nice idea for fixed sensors but I don't think solid core cable will last long in a moving drag chain.

Kitwn
01-04-2020, 01:30 AM
Neale,
X1 and X2 are for squaring the gantry. The only axis with two switches is X2. I've left the old mechanical microswitch at the other end of that one just in case the soft limits fail to stop a crash for some reason. Shouldn't ever be needed in practice.

I used NO switches to make interfacing 12V switches to a 5V BOB simple. Everything worked fine until I started the spindle. My circuit failed to work when I tried using 24V so a more complex interface will be required anyway. NC would have been better for the reason that broken wires show up immediately instead of causing a crash when you try to home the machine, but I'm not buying a new set now. I'm going to get the oscilloscope out and have a look at the noise, then some experiments will be needed to devise a reliable solution. I'll publish details once I have something useful to say. I might even make a video for everyone to watch while in lockdown!

Kitwn
01-04-2020, 03:29 PM
I didn't get time to look at the noise today but did manage to trace the input circuit for the cheap BOB. Picture below.
-
The inputs each feed one gate of a 74HC14D hex (that's six in one package) inverting Schmitt trigger. Exactly the technology I would have chosen for such an application myself!
There's a 10K pull-up resistor on the input and a 100pF capacitor to add a (very) little filtering.
Most interestingly, there's also an 1K resistor between the PCB header input and the screw terminals. This means that, when pulled low by an external circuit, using the screw terminals will leave the input to the trigger at about 0.4V higher than using the header connection. Since the low threshold for the trigger can be as low as 0.9V (typical value is around 1.3V with a 5V supply) this makes a big difference to how the circuit will behave, especially if you're not using a mechanical microswitch which provides a solid connection to ground.
-
Having just changed from microswitches to proximity switches with an isolating diode as well it's no wonder I'm having problems. The circuit below shows how I have the sensors wired into the BOB using a 1N4148 diode to isolate the 5V input from the 12V or 24V supply to the sensors.
The saturated output voltage of the sensor when triggered is about 0.3V. Add the usual 0.6V we see across a forward biased diode and that leaves 0.9V. BUT the current flow of about 0.4mA from the 5V supply through the 1K resistor lifts the input to the gate to 1.3V. This is right on the edge of the gate's lower threshold. All 4 sensors worked fine with a 12V supply before I fired up the spindle but very little noise is required to trigger a spurious trip. With a 24V supply the saturated output of the sensors rose less than 0.1V but this is enough not to trigger the inputs at all, even with no spindle noise.
-
Next move will be to connect the inputs to the PCB headers and see if the extra 0.4V of headroom clears the problem. I don't always get spurious trips so this might be enough.
-
I'm going to order a reel of 4-core screened cable from RS anyway, cut the tails off the sensors and wire them back to the connector panel without any other junctions.
-
What the diagram does not show is that I have 1uF capacitors across the screw terminal inputs for additional filtering. This was necessary to avoid spurious trips with the microswitches. These will also need to be moved to the PCB headers.
-
27695

27692

Doddy
01-04-2020, 10:00 PM
Next move will be to connect the inputs to the PCB headers and see if the extra 0.4V of headroom clears the problem. I don't always get spurious trips so this might be enough.


Interesting analysis. Re. the PCB headers, the only caution in a high noise environment is the protection of the '14 inputs to >5V spikes. The IC should have clamp diodes that should work to protect but worth bearing in mind.

I have to admit all my cheap boards have had opto isolated inputs which I think can only help for noise immunity for systems like this.

Voicecoil
01-04-2020, 10:20 PM
I didn't get time to look at the noise today but did manage to trace the input circuit for the cheap BOB. Picture below.
-
The inputs each feed one gate of a 74HC14D hex (that's six in one package) inverting Schmitt trigger. Exactly the technology I would have chosen for such an application myself!
There's a 10K pull-up resistor on the input and a 100pF capacitor to add a (very) little filtering.
Most interestingly, there's also an 1K resistor between the PCB header input and the screw terminals. This means that, when pulled low by an external circuit, using the screw terminals will leave the input to the trigger at about 0.4V higher than using the header connection. Since the low threshold for the trigger can be as low as 0.9V (typical value is around 1.3V with a 5V supply) this makes a big difference to how the circuit will behave, especially if you're not using a mechanical microswitch which provides a solid connection to ground.
-
Having just changed from microswitches to proximity switches with an isolating diode as well it's no wonder I'm having problems. The circuit below shows how I have the sensors wired into the BOB using a 1N4148 diode to isolate the 5V input from the 12V or 24V supply to the sensors.
The saturated output voltage of the sensor when triggered is about 0.3V. Add the usual 0.6V we see across a forward biased diode and that leaves 0.9V. BUT the current flow of about 0.4mA from the 5V supply through the 1K resistor lifts the input to the gate to 1.3V. This is right on the edge of the gate's lower threshold. All 4 sensors worked fine with a 12V supply before I fired up the spindle but very little noise is required to trigger a spurious trip. With a 24V supply the saturated output of the sensors rose less than 0.1V but this is enough not to trigger the inputs at all, even with no spindle noise.
-
Next move will be to connect the inputs to the PCB headers and see if the extra 0.4V of headroom clears the problem. I don't always get spurious trips so this might be enough.
-
I'm going to order a reel of 4-core screened cable from RS anyway, cut the tails off the sensors and wire them back to the connector panel without any other junctions.
-
What the diagram does not show is that I have 1uF capacitors across the screw terminal inputs for additional filtering. This was necessary to avoid spurious trips with the microswitches. These will also need to be moved to the PCB headers.
-
27695

27692

If it were me I'd be doing the following:

1) Add a decently low (<= 2K2) pullup on the output of the switch to 24V
2) remove the diode
3) Increase R2 to 10K
4) add a 2K7 resistor across the PCB header pins
5) up C1 to at least 10nF
6) remove R3

This should give a decent voltage swing on the switch output and a 0.1....+4.5v ish swing on the input to the 'HC14 with a bit of filtering and a 12dB attenuation of noise on the switch output to boot.

dazp1976
02-04-2020, 01:38 AM
Nice idea for fixed sensors but I don't think solid core cable will last long in a moving drag chain.

I won't be going through the drag chain. The switches will be fixed to the machine casting so they won't be moving. The little angle plates they will read off will be axis attached. This way the cable won't need to be that flexible.

Kitwn
02-04-2020, 02:21 AM
If it were me I'd be doing the following:

1) Add a decently low (<= 2K2) pullup on the output of the switch to 24V
2) remove the diode
3) Increase R2 to 10K
4) add a 2K7 resistor across the PCB header pins
5) up C1 to at least 10nF
6) remove R3

This should give a decent voltage swing on the switch output and a 0.1....+4.5v ish swing on the input to the 'HC14 with a bit of filtering and a 12dB attenuation of noise on the switch output to boot.

After I put up the last post I thought for a while and realised that using a diode was the wrong way to do this and that a simple potential divider with the BOB input seeing the open collector of the sensor directly was a far better idea as you suggest. I have a spare sensor and a spare BOB (that's one of the advantages of using a cheapo board instead of a $900 one) so I'll try those component values and check the volts. Thanks for working out the values.

It's tempting to try replacing the surface mount components on the BOB but I think an external bit of Veroboard with all the new bits soldered on will be the best option.

Doddy,
Come to think of it, I've probably got a quad opto-coupler or two in stock though I'll need to print up a surface mount PCB to use one of those. I feel a rummage through my drawers coming on!

cropwell
02-04-2020, 11:54 AM
Does Mach3 disable limits when homing, or is this an option?

I am not able to look at my machine as it is currently on the floor in my office, which I am refurbing.
27706

Looking back through threads (to 2016), I sort-of answered this question, but left doubt. I use a NO inductive switch for homing, but a NC microswitch for limits and I just thought that a break in the homing circuit could cause a crash if the limit is deactivated.

Voicecoil
02-04-2020, 01:31 PM
Does Mach3 disable limits when homing, or is this an option?

I am not able to look at my machine as it is currently on the floor in my office, which I am refurbing.
27706


Wow, using CAD to plan your workshop refurb!

Kitwn
02-04-2020, 01:45 PM
OK, after serious consideration my priority has to be getting the #$%^*&^%$#@ing machine actually working so I can make some stuff with it. Playing about with circuit ideas is fine, but...

I'm not wasting a long weekend off work waiting for RS to send me some new cable and other components including 2mm pitch plugs for the PCB headers, so I've made a circuit board which connects into the existing screw terminals without the need to remove/replace surface mount components on the BOB itself. I expect it's going to work. If so, it'll do for now and I'll make a more reliable change to the wiring later. This machine is after all just a toy in my shed rather than the heart of a commercial operation.
Watch this space for details if it works!

cropwell
02-04-2020, 01:52 PM
Wow, using CAD to plan your workshop refurb!

Notice that Mach3 is already loaded :biggrin:
27707

Thankfully, I had all the cabinets and worktop (25mm moisture proof MDF, 800mm deep) delivered before lockdown. All the carcases are built and in position. Next job is the worktops, and they are heavy sheets.

Anyway, My question is about homing and limits and I cannot find any reference to the query in the Mach3 reference manuals.

Neale
03-04-2020, 07:34 AM
Anyway, My question is about homing and limits and I cannot find any reference to the query in the Mach3 reference manuals.

Mach3 certainly disables limits when homing if you use common home and limit switches. Seems likely that it does the same with separate switches - I'm sure I remember reading this in the manual somewhere but that was a long while ago.

JAZZCNC
03-04-2020, 10:41 AM
I have a spare sensor and a spare BOB (that's one of the advantages of using a cheapo board instead of a $900 one) so I'll try those component values and check the volts.

Nope, that's no advantage because if you had a decent board you wouldn't have this problem in first place and this thread wouldn't exist.!! . . . Throw the bag of shite in the Bin and buy a decent board. While you are at it spend a bit more buy one with Ethernet...:joker:

dazp1976
03-04-2020, 12:04 PM
Nope, that's no advantage because if you had a decent board you wouldn't have this problem in first place and this thread wouldn't exist.!! . . . Throw the bag of shite in the Bin and buy a decent board. While you are at it spend a bit more buy one with Ethernet...:joker:

I've got a crappy £5 board on my benchtop machine and it works well for me but I only have the estop connected no physical limit switches as yet.
Haven't figured out my spindle wiring either, still doing that manually :smile:
1 question from me is that,. If I have a NO estop I'll need NO limit switches yes? Or does it not matter if estop is NO and limits are NC?

I plan on getting an ess in the future and use one board running axis/spindle and one board running estop/limits.

JAZZCNC
03-04-2020, 12:12 PM
I've got a crappy £5 board on my benchtop machine and it works well for me but I only have the estop connected no physical limit switches as yet.
Haven't figured out my spindle wiring either, still doing that manually :smile:
1 question from me is that,. If I have a NO estop I'll need NO limit switches yes? Or does it not matter if estop is NO and limits are NC?

I plan on getting an ess in the future and use one board running axis/spindle and one board running estop/limits.

You think it works well because you have nothing to compare it against.! When you get the new controller with higher pulse freq etc you'll see the difference.

Regards the N/O question then provided you use separate inputs then you can use different switch types. Obviously if wiring in series and sharing the same input they must all be the same type.

Also if your thinking to buy Warp9 ESS then I'd look at others like UC300 because they are better than ESS and you are not stuck with mach3/4.

cropwell
03-04-2020, 12:41 PM
Mach3 certainly disables limits when homing if you use common home and limit switches. Seems likely that it does the same with separate switches - I'm sure I remember reading this in the manual somewhere but that was a long while ago.

If that's the way it is, it ain't going to be changed now. If I had my machine available, it would be an easy test - set homing going and trigger a limit switch.

JAZZCNC
03-04-2020, 12:43 PM
Anyway, My question is about homing and limits and I cannot find any reference to the query in the Mach3 reference manuals.

So what's your question.?

Kitwn
03-04-2020, 02:57 PM
Well this has been an interesting day! I think the moral has to be: If you've got the test gear, use it!!

I did some quick mods in line with previous suggestions but to no avail. I was soooooo confident as well.

I finally got the oscilloscope out and looked at the noise on the limit switch inputs to the board. Clean as a whistle until I enabled the motors. Then there are short bursts of a surprisingly sinusoidal 6MHz waveform. Same on the 5V power rail. Same on the 5V ground rail!! I thought all the earthing was OK, Could be better, but OK. Clearly the wiring to the sensors needs to be completely replaced. At the moment it's modded from what was there for the microswitches. Obviously not good enough.

Problem is at that frequency a 20cm length of earth wire is too long. I measured 400mV (peak-peak) across the two ends of a wire les than that length from the chassis to the 5V ground connection on the board. The real killer comes from switching on the mains to the VFD. There's a transient on the Z sensor input that goes from 4.5V to below zero. It only lasts about a tenth of a microsecond but that's long enough. Obviously I'm not switching the VFD mains on and off while the machine is in use but it indicates there's a problem in need of fixing.

I did take a squillion screen grabs of the waveforms to show off with but the fact is that the controller needs to be rebuilt with more care taken over the earthing and the sensor cabling needs to be replaced with a separate, unbroken screened cable from each sensor (existing unscreened cables cut short as advised earlier). I suspect some of the wiring and existing filtering attempts are ringing and I doubt that even the most expensive new control board could be relied upon to work faultlessly in this electrical environment, so I'll stick to my well used $20 unit for now, as well as the old computer and parallel interface. Perhaps a complete new machine will materialise once I've retired. Whenever that turns out be now that the world has turned upside down.

I now have to wait for RS to deliver the goodies required for the improvements so will have to try being patient for a few days. Maybe I should try designing that low cost DIY optical fibre interface that would allow the motor drivers to sit next to the motors and no long wires through the machine except power rails. Well we must keep the old brain busy during these long spells stuck at home. Does anyone know if the professionals have moved to fibre connections for motor controllers and/or limit switches? It would solve a lot of problems.

Doddy
03-04-2020, 03:09 PM
Maybe I should try designing that low cost DIY optical fibre interface that would allow the motor drivers to sit next to the motors and no long wires through the machine except power rails. Well we must keep the old brain busy during these long spells stuck at home..

The first time I attached a scope to my spindle encoder on the lathe my jaw hit the floor with the noise from the 3-phase motor, so I appreciate where you're coming from. I'm surprised you're seeing that behaviour on short cables though comms theory was never my strong point - I'll believe your experience over my memory any day. I still think using cheap (or expensive!) BoBs with opto-isolated inputs offers a lot of advantages, if only for the pathetic frequency response that will filter any amount of HF noise as well as providing much better noise immunity level than HC-series logic.

In my case (the spindle encoder) - the design of the 7i76e Mesa card is intelligent and offers a differential input for shaft encoders with 130R input, which allows easy connectivity to a RS485 driver to allow transmission over a balanced twisted pair - helped massively. I appreciate this doesn't work in your case.

Muzzer
03-04-2020, 04:18 PM
Of course, most machines are massive lumps of cast iron and steel cabinets which form a good hard ground and provide a lot of shielding into the bargain. In contrast, if you are constructing a relatively spaced out router(?) using extruded sections with long lengths (high stray inductance), you lose many of those benefits.

For my machines, I've used Lapp Olflex shielded drag cables, which have shielded bundles within a shielded outer braid. I noticed that the original wiring in my Shizuoka (Matchmaker CNC system, using Parker hannefin servos etc) used simple screened multicore cables for stuff like the encoders but simple unscreened wires for most of the rest (limit switches, solenoids etc). The spindle motor and brushed servos were wired through flexible, grounded steel conduits, which would have contained a lot of the HF noise. I reused that system when I swapped out the electronics and it's been very well behaved.

I have the Yaskawa VFD in the same cabinet as the rest of the electronics but I bought the pukka Yaskawa-Schaffner EMC filter to go with it. I've not had any issues with that either.

cropwell
03-04-2020, 05:13 PM
So what's your question.?

Does Mach3 disable limits when homing, or is this an option? (post #25)

dazp1976
03-04-2020, 05:20 PM
You think it works well because you have nothing to compare it against.! When you get the new controller with higher pulse freq etc you'll see the difference.

Regards the N/O question then provided you use separate inputs then you can use different switch types. Obviously if wiring in series and sharing the same input they must all be the same type.

Also if your thinking to buy Warp9 ESS then I'd look at others like UC300 because they are better than ESS and you are not stuck with mach3/4.

Thanks. There's eperate connections for estop and each axis limits. 2 switches will be wired in series per axis.

In regard to the controller in the u.k. I can easily get an ESS or a UC400eth but the 400 has limited ports.
I've found a UC300eth-5lpt in Ireland and that's about it.

I'll still end up using the cheap boards on them though but that shouldn't matter too much, should it.

Voicecoil
03-04-2020, 08:20 PM
Thanks. There's eperate connections for estop and each axis limits. 2 switches will be wired in series per axis.

I'll still end up using the cheap boards on them though but that shouldn't matter too much, should it.

As mentioned before make sure the "cheap" board has Inputs that are opto-isolated.

JAZZCNC
03-04-2020, 08:55 PM
Does Mach3 disable limits when homing, or is this an option? (post #25)

It's an option. If you go into General config there's a setting called Home SW safety. Tick this and it will enable limits while homing. Obviously if sharing the same switch as limits this needs to be off.

JAZZCNC
03-04-2020, 09:18 PM
Thanks. There's eperate connections for estop and each axis limits. 2 switches will be wired in series per axis.

In regard to the controller in the u.k. I can easily get an ESS or a UC400eth but the 400 has limited ports.
I've found a UC300eth-5lpt in Ireland and that's about it.

I'll still end up using the cheap boards on them though but that shouldn't matter too much, should it.

And do you think you'll need more than 10 inputs and 24 outputs.? I'm also not sure if some of those Outputs on the second port can be configured as inputs in software.? Only if you plan on having an ATC will you probably need more than 10 inputs.

Regards the Cheap BOB then yes, of course, it will make a difference. It's like having 500HP engine and fitting bald tyres.? Pointless.!

Any machines only as good as the weakest link so fitting a good card with a cheap BOB that probably can't keep up doesn't make sense to me. This whole thread proves this.
I fit good electronics, which yes cost more, but I never have any of this shit to deal with and the extra cost is made back by the fact I'm not running around fixing machines or put another way the owner doesn't have any downtime which costs money.!! . . . . You get what you pay for.!

If your going to use a cheap BOB then save your money don't buy ESS or ETH400 and use Linux CNC. You have half a chance of good machine and won't cost anything other than it's a steep learning curve.

Kitwn
04-04-2020, 07:43 AM
The first time I attached a scope to my spindle encoder on the lathe my jaw hit the floor with the noise from the 3-phase motor, so I appreciate where you're coming from. I'm surprised you're seeing that behaviour on short cables though comms theory was never my strong point - I'll believe your experience over my memory any day. I still think using cheap (or expensive!) BoBs with opto-isolated inputs offers a lot of advantages, if only for the pathetic frequency response that will filter any amount of HF noise as well as providing much better noise immunity level than HC-series logic.


What gobsmacked me the most was the amount of noise before I ever put mains on to the VFD. It's enabling the motors that introduces all the trouble. And the fact that the same noise is on the ground of the BOB. I thought the screening and earthing was quite reasonable but noooooo!

One thing that all this does point out is how noise-resistant it is to use NC microswitches on the limits. Under normal operation all the limit inputs have a hard-wired, all metal connection short circuit on them. This is why I've never had any of this trouble before. Ironically I did briefly consider using miniature relays to act as the 12 or 24V to 5V interface but considered it unprofessional and was in fear of being laughed at by my forum fellows. Actually it would work a treat! I doubt if the variation in triggering delay would noticeably affect the accuracy of the homing and the low-pass filtering effect would be even better than your suggested opto-couplers!

Re the 'experience', which isn't really as great as you so kindly implied: When I was a trainee at the BBC, back when Boris Johnson had a poster of Prime Minister Thatcher on his school dormitory wall, one of the trainers gave a demonstration of how there's no such thing as an earth connection at radio frequencies. This was at the Daventry short wave station which is now long gone. We went out into the antenna field and he took a coin out of his pocket and dragged it at head height down the side of a steel support post which carried one of the feeders to an antenna. This post was about 20cm or so in diameter and solidly concreted into the ground. There were distinct, bright sparks between the coin and the post!

Kitwn
04-04-2020, 08:07 AM
Dean,
For a commercial machine that has to earn it's keep I'm in complete agreement with you. I'd have gone bankrupt a long time ago if I were mad enough to try and use the machine I've built as part of a business. But for me it's just a toy I have in my shed and I can't justify spending large amounts of money on it for new parts I can probably do without if I apply some time and ingenuity to the problems that arise. With the $AUD currently worth no more than about 60 US cents, prices of most other board options are way too high.

In practice this $20 BOB has done everything I need from it for several years now and the current noise problem is nothing to do with the board itself. I'm actually quite impressed that it worked so reliably before, with NC microswitches to hide the problem, and has proved quite robust as well, not complaining at all when I accidentally put 12V on one of the 5V limit inputs.

Building and modifying the machine and learning from the problems that occur is part of the fun of making it for me, it's hobby in itself, though there is an ever-increasing list of other projects on hold that are waiting for it to be working properly again.

Voicecoil
04-04-2020, 01:16 PM
What gobsmacked me the most was the amount of noise before I ever put mains on to the VFD. It's enabling the motors that introduces all the trouble. And the fact that the same noise is on the ground of the BOB. I thought the screening and earthing was quite reasonable but noooooo!


That doesn't really suprise me - I remember when I powered up my first PWM audio amplifier - which was switching roughly the same voltage and frequency as a decent microstepping drive - and the radio in the workshop promptly went off. Holding a 'scope probe 6" above the speaker leads gave a few volts of noise in thin air, I ended up using screened speaker cable :listening_headphone:. Are the cables to your motors screened I wonder? As with the sensors when I extended my cables I cut them very short then extended with screened (2 pair) which might explain why I've not had any problems in a setup which seems somewhat similar to yours - I'm using basically the same type of sensor albeit in the M8 size. The motor cable screens I returned to the HT- pins on the drivers (always return your shit to where it's generated!) which then has a VERY SHORT ground lead down to the aluminium back plate of the enclosure which acts as a ground plane - being an RF man you'll appreciate the benefits of those.

Kitwn
05-04-2020, 03:47 AM
Voicecoil,
I thought I'd been quite careful with the wiring: the BOB is in it's own aluminium box, the outputs to the drivers are 4 core screened, the cables from the drivers to the motors are screened, the PSU for the BOB (it's the original ATX supply in the PC case that is my control box) is a separate unit from the motor PSUs, the VFD is separately mounted on the wall, the spindle power cable is proper screened cable designed for the job.

The weak point is the sensor wiring which was not all screened cable for the old microswitches and has not been totally replaced yet, but will be soon. All the grounding of every component in the control box will need to be looked at again as well. No going down to the pub this evening for me!

Kitwn
05-04-2020, 04:05 AM
I had very similar problems until I did this.

https://www.youtube.com/watch?v=DJfiOqaeFDg

Simon

Simon,
Apologies for not getting back to you earlier, I got somewhat sidetracked away from the original reason for this thread.
That's a very interesting video and I especially like the improvised current transformer for investigating noise, I hadn't seen that trick before but it will be useful in future. As I've discovered the VFD is not to blame (yet!), it's enabling the motors that generates the noise I'm investigating at the moment. Once that's been quietened down I can then see if the VFD adds any problems. It wasn't an issue when I had NC microswitches for limit sensors so my fingers are crossed on that one.

Voicecoil
05-04-2020, 09:24 AM
Voicecoil,
I thought I'd been quite careful with the wiring: the BOB is in it's own aluminium box, the outputs to the drivers are 4 core screened, the cables from the drivers to the motors are screened, the PSU for the BOB (it's the original ATX supply in the PC case that is my control box) is a separate unit from the motor PSUs, the VFD is separately mounted on the wall, the spindle power cable is proper screened cable designed for the job.

The weak point is the sensor wiring which was not all screened cable for the old microswitches and has not been totally replaced yet, but will be soon. All the grounding of every component in the control box will need to be looked at again as well. No going down to the pub this evening for me!

Well, you've done a thorough job then, strange that you're seeing so much induced noise - unless it's induced pickup of magnetic field radiation?. It will be interesting to see what you come up with as a fix.

Kitwn
05-04-2020, 11:58 AM
Well, you've done a thorough job then, strange that you're seeing so much induced noise - unless it's induced pickup of magnetic field radiation?. It will be interesting to see what you come up with as a fix.
Hadn't considered magnetic induction, probably not the cause at such a high frequency, but I'm willing to be proved wrong. It wouldn't be the first time!

I've still got some investigation I want to do. The controller is connected to the machine by screened multicore cables which have plugs on the end which fit into sockets on the machine itself. The two X axis sensors and the Z&Y sensors have one 8-way for each pair. I want to disconnect the cables and see if that makes a difference. This will help to decide if the noise is being picked up in the wiring within the machine itself where the Z and Y sensor cables run through the same drag chain as the motor wiring or whether the problem is within the controller itself. Both will be dealt with anyway but knowledge is never wasted. I'm also tempted to use the technique I learned from the video that Cube3 linked to to look at the noise on the motor cables with my laptop oscilloscope. But 6MHz???
While I wait for RS to deliver the bits required I'm doing some necessary tidying up. There isn't much room in my workshop with the machine itself pulled off the wall to get at all the wiring (note for new builders: If you have to put your machine up against a wall, do NOT put the drag chains and all the gantry wiring on the 'wall' side of the machine in order to keep it conveniently out of the way). The only place to put the trolley with the PC on is in the doorway to the rest of the shed (sorry, that should be 'studio' if my wife is listening) if I want to leave it all connected so the picture below had to be taken through the window!

27733

Voicecoil
05-04-2020, 12:30 PM
Hadn't considered magnetic induction, probably not the cause at such a high frequency, but I'm willing to be proved wrong. It wouldn't be the first time!
The weirdest things can happen with EMC issues, at 6Mz a couple of bits of wire running parallel for a few m probably makes a half decent transformer :teapot:


Voicecoil,
No going down to the pub this evening for me!

No chance of that for me either - 'coz of the virus they've been shut here for over 2 weeks :dejection:

dazp1976
05-04-2020, 03:00 PM
I get noise when my stepper motors enable. It affects my FM radio.
Now.. I have no switches and all circuits etc in a metal control box. The noise is from my estop (un-screened cable in use). :concern::confusion:

Kitwn
06-04-2020, 04:02 AM
The weirdest things can happen with EMC issues, at 6Mz a couple of bits of wire running parallel for a few m probably makes a half decent transformer
This picture shows the output coupling transformer of a Marconi BD272 AM short wave radio transmitter rated at 250KW (that's a whole Megawatt peak output). I cut my teeth on these things in the 80's. All that power is being coupled through this single turn, air spaced transformer. These are the coils for 15MHz I think, but the 6MHz ones are not much bigger. The peak voltage across the primary turn is about 40KV even though it looks remarkably like a short circuit to the uninitiated.

27734




No chance of that for me either - 'coz of the virus they've been shut here for over 2 weeks :dejection:

Same here. The local microbreweries are doing canned takeaways though, so I can still enjoy my favourite local brew at much less than the in-a-glass price they charge if you sit on an old wooden chair in their gravel yard getting swamped by midgies and other people's children. You have to look on the bright side!

Western Australia has now closed it's borders to all non-essential travellers and travelling between different regions within WA without a good reason can cost you $50,000.

Voicecoil
06-04-2020, 08:43 AM
This picture shows the output coupling transformer of a Marconi BD272 AM short wave radio transmitter rated at 250KW (that's a whole Megawatt peak output). I cut my teeth on these things in the 80's. All that power is being coupled through this single turn, air spaced transformer. These are the coils for 15MHz I think, but the 6MHz ones are not much bigger. The peak voltage across the primary turn is about 40KV even though it looks remarkably like a short circuit to the uninitiated.
27734
Thanks for posting that - awesome bit of kit even if it looks completely illogical to the uninitiated!

Kitwn
06-04-2020, 01:35 PM
Thanks for posting that - awesome bit of kit even if it looks completely illogical to the uninitiated!

This machine is 1960's technology. Some of the later, higher power designs, have water cooled the coils. A key difficulty with those was how to clamp the plastic water pipes onto the copper coils. Metal jubilee clips glowed white hot as soon as we switched the bloody thing on and plastic clips went brittle in about a week. We ended up using waxed cotton whipping cord, similar to to an old fashioned cricket bat handle. It's surprising the skills I had to learn as a BBC Transmitter Engineer back in the days when it was still fun.

cropwell
06-04-2020, 01:53 PM
This machine is 1960's technology. Some of the later, higher power designs, have water cooled the coils. A key difficulty with those was how to clamp the plastic water pipes onto the copper coils. Metal jubilee clips glowed white hot as soon as we switched the bloody thing on and plastic clips went brittle in about a week. We ended up using waxed cotton whipping cord, similar to to an old fashioned cricket bat handle. It's surprising the skills I had to learn as a BBC Transmitter Engineer back in the days when it was still fun.

Bet it made instant toasties. Seriously though, was all that RF energy hazardous to workers ?

Kitwn
07-04-2020, 02:16 AM
My standrad reply is "I've spent 40 years working in high RF fields and it nev, nev, nev, never did me any harm!" On one site I worked at you couldn't turn the flourescent lights off in the transmitter building. They wouldn't strike on their own but once lit they wouldn't go out until the maintenance break . They flickered in time to the program so you could set your watch by the lights: the time pips have a very obvious pattern.

High power wireless has been around for a while now and the only known risk from such non-ionising radiation is the heating effect. VERY rare that anybody gets exposed to that kind of power, only antenna workers in practice. Contact with the conductors can give you an electric shock as well, but that isn't unique to RF.

Some people claim to have been harmed by mobile phone use. A coleague of mine had a tumour the size of a tenis ball taken out of his head some years ago and it was on the side that he had a phone clamped to for half of the day, but there's no evidence that was the cause.

Ionising radiation is the dangerous stuff but that's way above even the highest frequencies used for communications.

cropwell
07-04-2020, 02:29 AM
Going back to the original topic, Kit, you don't have any transmitter masts near you ?

Kitwn
07-04-2020, 02:44 AM
I get noise when my stepper motors enable. It affects my FM radio.
Now.. I have no switches and all circuits etc in a metal control box. The noise is from my estop (un-screened cable in use). :concern::confusion:

Have a look at that video Cube3 posted on page 1. Ferrite rings should help you there. You can wind the cable through the ring to increase it's effect.
27740

Kitwn
12-04-2020, 11:52 AM
Going back to the original topic, Kit, you don't have any transmitter masts near you ?
Sorry, I missed this comment first time round:
Actually, yes. As far as I'm aware it's the largest wireless transmitting antenna in the southern hemisphere and is about 14Km away from the CNC machine.

cropwell
12-04-2020, 12:13 PM
Sorry, I missed this comment first time round:
Actually, yes. As far as I'm aware it's the largest wireless transmitting antenna in the southern hemisphere and is about 14Km away from the CNC machine.

It transmits VLF signals for the US Navy Submarine Fleet. I thought this could be giving your sensor LEDs a buzz. The sensors themselves generate an ac em field, but I don't know the frequency, but there might be some heterodyning.

Just a thought, but it might have no merit.

Kitwn
12-04-2020, 01:57 PM
Embarrassing but I must own up: I realised during my tests that for all these years I've left the 36V supply for my stepper motors floating. Neither rail was grounded so the whole supply was bouncing up and down by 8 Volts at several KHz. It's a miracle the machine worked at all! Any way that's now fixed and some of the bits from RS have arrived, though not the screened cable for re-wiring the sensors yet.
-
Fortunately I have access to the highly sophisticated electronics test laboratory shown below and have been able to make some very revealing, noise-free measurements on the spare cheap Chinese breakout board. Earlier reported voltage measurements were made in the presence of noise and are not reliable.
-
With an input fed to the board header rather than the screw terminals, a falling voltage must drop below 0.6V to set the output of the board. A rising voltage must go above 1.7V to reset it. This gap is the hysteresis we expect from a Scmitt trigger input chip and provides some noise immunity.
-
When using normally closed (NC) microswitches as sensors the inputs are a good solid short circuit for most of the time and only open briefly during homing. Excellent noise immunity is obtained in this configuration as proved by the fact that my machine has worked reliably for a few years!
-
Using proximity sensors presents a problem, even with opto-isolators or other electronic interfacing, since the saturated collector-emitter voltage at the output of the proximity detector or from an isolator when triggered is about 0.6V. No guarantee of reliable triggering here even without noise. The use of a potential divider circuit as described earlier is an option but I've decided to use a more direct and low-impedance option... reed relays. These will be fed by an NPN emitter-follower transistor which serves to boost the current available from the proximity sensor.
-
Some will scoff no doubt, but it actually makes sense for a few reasons:
I know from past experience that a solid short circuit on the inputs works reliably, even in the presence of high noise.
Low-pass filtering is excellent and requires no additional components.
When wired as I plan to have them, with the 12V from the sensor holding the relay contacts closed as the normal state (this mimics the previously known-good NC microswitches), the circuit is proof against wiring faults as any breaks will trip the limit input.
-
Regarding positional accuracy, a key point of using proximity sensors is improved repeatability over microswitches. The reed relays I'm going to use have a specified response time of 0.5mS. Allowing for, say, 20% variation that gives a repeatability of within of 0.1mS. I can't remember exactly what my second approach speed is set at for homing in LinuxCNC but if it's even as high as 1m/min that equates to a variation of around 2 microns. Which equals nothing.

The 4 relays are $20 the lot from RS, the transistors and a wee bit of veroboard are existing stock worth much less. The whole circuit has been tested on the bench ( I have 2 relays in stock already) and works a treat. So I might just have a fully working machine again well before Mark McGowan let's me back in the pub!

27796

Doddy
12-04-2020, 02:04 PM
No scoffing from this direction.

I'm somewhat surprised with the Vce(sat) for the opto, but if that's what you've measured then that's what it is. Personally at that stage I'd be using change-over contacts for single-switched inputs and using NC=gnd, NO=5V just to yank that signal line to one rail or the other (nothing left floating).

For the nay-sayers, what Kitwn's post is providing is belted-and-braced robust and deterministic behaviour. There's too many problems in the world to save pennies on uncertainty on our machines.

Perhaps I'm feeling humbled after a day's worth of trying to get my control box switching automatically between two machines... and making some stupid assumptions... right, where's that damned scope gone...

Kitwn
12-04-2020, 02:17 PM
Not measured on a specific opto, but it is the measured output voltage of the proximity sensor when triggered. Opto outputs will vary depending on device and load current but with only 0.6V to play with I'm not chancing it. I was surprised at the low trigger voltage, but it's what I measured on the board input and the chip spec says it can be as low as 0.7V
-
This is likely to be a problem with any board that only uses 5V logic on it's inputs. I assume the more expensive boards are more sophisticated. They'd better be for the difference in price.
-
I'm happy that the 10K pull-up on the board will work OK with the single pole relays. The relay will isolate the actual board input from any noise on the wiring from the sensors and the overall noise level is going to be significantly reduced by my improvements to the wiring overall.

Doddy
12-04-2020, 02:41 PM
I'll stick with my self-centred view of having BoBs and similar with opto-isolated inputs. I've never been a fan of distributing a high speed logic input signal into the big, bad world.

Now... where to buy a replacement for a badly crimped D25-26w IDC cable (Bob2->UC300) that's given me gip for the last day, on Easter Sunday?...

JAZZCNC
12-04-2020, 03:14 PM
Throw it in the bin I say..:joker:

Kitwn
12-04-2020, 03:30 PM
I'll stick with my self-centred view of having BoBs and similar with opto-isolated inputs. I've never been a fan of distributing a high speed logic input signal into the big, bad world.

I'm in complete agreement with you on that one, but the logic following it probably needs to be slightly different from the input circuit to my cheapo card. Something with a supply voltage well above 5V would be a good start.

EDIT: Been doing a bit more looking at opto-couplers and there certainly is a big range of options. The Darlington output types have a higher saturated output (up to 1V) compared to the conventional type (as low as 0.1 - 0.2V at low current) so you really must choose your device, but there are devices that would interface successfully to the cheap card. Since this application is at a guaranteed very low frequency I'm happy to let old technology do the job this time.

Kitwn
12-04-2020, 03:33 PM
Throw it in the bin I say..:joker:

I promise I will do exactly that. As soon as I receive your cheque for the $400 AUD or so that I'll need to replace it with something better. Now THAT'S how you give a Yorkshireman a heart attack :hysterical:

JAZZCNC
12-04-2020, 05:44 PM
I promise I will do exactly that. As soon as I receive your cheque for the $400 AUD or so that I'll need to replace it with something better. Now THAT'S how you give a Yorkshireman a heart attack :hysterical:

If you are prepared to use Mach3/4 then I've probably got something I could send your way.

Voicecoil
12-04-2020, 10:03 PM
Not measured on a specific opto, but it is the measured output voltage of the proximity sensor when triggered. Opto outputs will vary depending on device and load current but with only 0.6V to play with I'm not chancing it. I was surprised at the low trigger voltage, but it's what I measured on the board input and the chip spec says it can be as low as 0.7V
-
This is likely to be a problem with any board that only uses 5V logic on it's inputs. I assume the more expensive boards are more sophisticated. They'd better be for the difference in price.
-
I'm happy that the 10K pull-up on the board will work OK with the single pole relays. The relay will isolate the actual board input from any noise on the wiring from the sensors and the overall noise level is going to be significantly reduced by my improvements to the wiring overall.

It's a shame they've used a 74HCT14 for the buffer on your BOB, the old (and now probably uncommon) 40106 hex Schmitt buffer was a better chip for such purposes IMHO, as the threshold voltages are more or less equispaced around 1/2 the supply voltage, something like 2.3V & 2.7V.

Kitwn
13-04-2020, 03:27 AM
If you are prepared to use Mach3/4 then I've probably got something I could send your way.

PM sent.

Kitwn
13-04-2020, 05:00 AM
Just been perusing the HDBB2 breakout board from CNC Drive. This appears to be nothing more than an up-market version of the cheap Chinese parallel port breakout board with opto-coupled inputs and a number of improved features. The company advertise this board as suitable for use with LinuxCNC direct from an old PC's parallel port as well as with CNC Drive's own products. Since this board is a necessary add-on to the UC400ETH Ethernet controller anyway, it might make a good starting buy for a budget build with the option to upgrade at a later date.
Has anybody used one of these in this way?

https://www.benl.ebay.be/itm/HDBB2-breakout-board-CNC-Mach3-LinuxCNC-/121287188690

dazp1976
16-04-2020, 01:25 AM
And do you think you'll need more than 10 inputs and 24 outputs.? I'm also not sure if some of those Outputs on the second port can be configured as inputs in software.? Only if you plan on having an ATC will you probably need more than 10 inputs.

Regards the Cheap BOB then yes, of course, it will make a difference. It's like having 500HP engine and fitting bald tyres.? Pointless.!

Any machines only as good as the weakest link so fitting a good card with a cheap BOB that probably can't keep up doesn't make sense to me. This whole thread proves this.
I fit good electronics, which yes cost more, but I never have any of this shit to deal with and the extra cost is made back by the fact I'm not running around fixing machines or put another way the owner doesn't have any downtime which costs money.!! . . . . You get what you pay for.!

If your going to use a cheap BOB then save your money don't buy ESS or ETH400 and use Linux CNC. You have half a chance of good machine and won't cost anything other than it's a steep learning curve.

Looks like I can get it ok so I'm leaning towards the UC300eth-5lpt and opting for a UCSB breakout board. Get some cables and NPN-NO sensors too and use it for axis, limit & stop control- http://cncdrive.com/UCSB.html.
Then I'll use my cheap board for spindle control and any relays.
What do you reckon?
Futureproofs it in case I ever get to build a 3 axis ATC or similar at a later date. (likely £3000 ish with tooling....ouch!!!! :concern: Don't tell the wife!)

I'd be comfortable with the ESS but like you say. It gives me the option of going away from Mach3.
It's a tricky decision.

Doddy
16-04-2020, 08:11 AM
That UCSB board, on paper, looks to be a pretty good BoB for the extra ports on (at least) the UC300ETH. It's pricey, though, but looks to be built like the proverbial brick-outhouse. Probably a bit of an overkill for most machines (how many layers of optical isolation do you really need?) but if you're happy with the price. I'd be thinking that if I was ever to need more than the 12 outputs (bearing in mind you'll need 8-9 for a 4-axis) then I'd relegate the axis stepper signals to a cheaper BoB on the port 2/3 on the 300 and retain the highly flexible I/O of the UCSB for other purposes on the machine.

Kitwn
22-04-2020, 12:47 PM
At last!! A nice man in a white van turned up today with a box from RS containing the cable for re-wiring my new proximity sensors. I managed to get the old wiring ripped out of the drag chains and the new cables in place for the Y and Z sensors before wine 'o' clock.

Tomorrow I should have time to get the two X axis sensors wiring in place and all the soldering done to connect everything up to the control box. Sadly the two remaining miniature relays for coupling the sensors into the BOB are a separate order which is still on it's way from somewhere on planet Earth, but I should be able to test all the wiring and confirm that everything else works as expected. I already have a list of outstanding projects from Mrs W-N that are waiting for the machine to be back in action so I'm hoping the relays arrive soon!

Kitwn
23-04-2020, 12:52 PM
I was just getting to the end of all the new wiring when the same nice man in the same white van turned up with another parcel from RS containing the relays. Apart from a slight problem mentioned elsewhere I now have the 4 new proximity sensors wired up with screened cables back to the controller and the relay/transistor interface into the BOB finished and tested.
Tomorrow I will tidy up the impressive mess and get everything put back together and ready for testing and alignment. Auto-squaring will need to be re-calibrated and the finished aluminium (rather than the old plywood prototype) Z axis lined up as well. Might be able to make something soon!

cropwell
23-04-2020, 01:53 PM
Kit,

If you find reed relays in the homing circuit don't give you a consistent zero point, you could try DC-DC SSR's

Kitwn
23-04-2020, 02:26 PM
Kit,

If you find reed relays in the homing circuit don't give you a consistent zero point, you could try DC-DC SSR's

To be honest I'm hoping I don't have to try anything else for quite a while. I've spent too much time messing about with the machine rather than making anything with it. I'm reasonably confident that the relays will work OK. The response time (including bounce) is specified at 0.5mS. Variation in response time is going to be a lot less than that, especially as they are safe in the control box and not being vibrated and bashed about like the old microswitches were. If the dimensional errors from all the other weaknesses in this machine prove to be small enough for the relays to become significant I shall be very happy!

Kitwn
24-04-2020, 09:11 AM
Well I never thought we'd end up with 8 pages of guff on this over three weeks, with an assortment of new/abandoned circuit designs and quite so much head-scratching!

Anyway the machine is now back together and appears to be working every bit as well as it did before I decided to replace the microswitches with inductive proximity sensors. The real excitement is that I can now see how well the all-aluminium Z axis behaves compared to the old plywood prototype I've been using for the last year or more.

dazp1976
24-04-2020, 10:28 AM
Well I never thought we'd end up with 8 pages of guff on this over three weeks, with an assortment of new/abandoned circuit designs and quite so much head-scratching!.

Lockdown sending us all a bit loopy that is :rolleyes:

Just bought a load of prox sensors myself. Might get round to fitting them in the next year maybe.
Have fun with yer machine and hope it behaves for you.

Kitwn
24-04-2020, 12:16 PM
Lockdown sending us all a bit loopy that is :rolleyes:

Just bought a load of prox sensors myself. Might get round to fitting them in the next year maybe.
Have fun with yer machine and hope it behaves for you.

Actually the lockdown isn't making too much difference in the W-N household. Except that now we have to call it "Social Distancing" instead of "Anti-Social Distancing".:glee:
Thanks for the good wishes, I'm looking forward to getting on with the list of projects I have outstanding. Funny how everything else seems to rely on the CNC machine as soon as you have one.

Kit

ericks
24-04-2020, 01:51 PM
The AXBB-E is not a bad option...also easy to connect your inductive proximity switches directly to the inputs.

Kitwn
25-04-2020, 10:08 AM
The AXBB-E is not a bad option...also easy to connect your inductive proximity switches directly to the inputs.

If/when I ever decide to build another machine from scratch and move away from the dirt-cheap LinuxCNC/Chinese BoB controller, that looks like a good alternative.

cropwell
28-04-2020, 04:16 PM
Embarrassing but I must own up: I realised during my tests that for all these years I've left the 36V supply for my stepper motors floating. Neither rail was grounded so the whole supply was bouncing up and down by 8 Volts at several KHz. It's a miracle the machine worked at all! Any way that's now fixed and some of the bits from RS have arrived, though not the screened cable for re-wiring the sensors yet.
-
27796

I am just rewiring my control panel and realised that my 48V to the drivers was also floating.

Muzzer
28-04-2020, 05:26 PM
I found that those clip-on ferrite clamp filters are pretty good at dealing with some of the noise spikes you get in installations like these. With switched mode power supplies, VFDs and stepper / servo drivers all doing their little thing, there are lots of noise sources thrashing away. Just a few of those spikes getting in to the step / direction etc lines can cause lost or gained steps or gradual creep of the table.

You can get them from ebay, CPC etc for peanuts and they are very simple to fit, without needing to disconnect anything. The actual specs probably aren't massively important as long as they actually fit over your cables(!). Worth having a few handy or even just a good idea to fit some by default.

I got some of these randomly chosen Chinesium Mystery Brand filters that work out about a quid a pop for the 9mm version. Other brands are available.....
https://www.ebay.co.uk/itm/402221769539

Neale
28-04-2020, 05:27 PM
I'm pretty sure my 68V DC driver supply is also floating, and has been for the last 3 years or so. Why not? So, it bounces around a bit, but considering that you are generating 4A+ spikes of current up the power leads generating lots of potential EMI, what does a potential difference between the wires and some arbitrary ground point mean?

One good point about leaving it floating is that if there is an accidental short to earth of a motor lead at some time, you are less likely to make a driver go pop.

I have my "naive" hat on here but isn't there an argument that says that leaving it floating is less likely rather than more to inject random noise into the environment?

Muzzer
28-04-2020, 05:32 PM
Yes, a lot of the big industrial gear does that. They use a ground fault detection circuit to look for a fault to ground and flag up a warning if one is seen, rather than bombing out immediately. That allows the system to continue running until it can be brought safely to a halt. This can be handy if you happen to be operating eg an elevator lift or a paper / steel mill at the time.

You can always connect the floating node to ground via a Y cap to stop it flying about so much. This gives the noise a path back to ground without the danger of a DC path popping your drives or controller.

Kitwn
29-04-2020, 12:51 AM
Muzzer, Neale,

My original issue with this was in regard to noise getting into the BoB supply which appeared to have been a major part of my problems. It seemed obvious at the time, but I take the points you both make. As long as the BoB supply and the motor supply are completely separated and the cable screens effectively grounded it should not be an issue. I have my machine working now and am keen to start using it to make things but I thing the next time I decide on a significant upgrade it will have to include a complete rebuild of a new controller in a better enclosure with a far more comprehensive approach to noise screening.

The clamp-on ferrites look like an excelent idea and I'll order a bagful.

Rob,
As you'll see from the above replies, you shouldn't be changing that anytime soon. My reasoning lacked the rigourous engineering logic it should have had.