TOS-KURIM CNC Upgrade

Just had a play with the transformer, with 240Vac through the 380V terminals it gave secondary reading of 95V on each and 190V across both... could be usable?

Thanks for the link to the caps, they look a bit more substantial!
would something like this bit a better solution for the bridge rectifier https://www.mouser.co.uk/ProductDeta...n%2Fkm7zBfw%3D

Thanks for the photos, very handy to be able to visualise the whole set up.

I have no idea on the photos, all of mine were rotated through 90 degrees as well! I tried everything I could think of but they just revert back to side ways!

That is a bit of a mother and IXYS is a "premium" brand. That first one you linked to from CPC would be absolutely fine. These diode bridges are generally very robust.

A better solution would now be to connect the 2 primaries in series and the 2 secondaries in series (as they are already).

I take it that the 380V windings are the brown and blue. Each contains 2 wires - one from each limb. You could separate them and connect them in series so that there is a (floating) centre tap formed by one brown and one blue. Then the voltage on each primary would be 120V and the voltage across the ends of the series connected secondaries would be your 95Vac. Then you would simply feed the 95Vac into a bridge rectumfrier and you would get your 130Vdc or so.

Apart from reterminating the wires, you are almost done. It's a rather large beast but there is no extra expense and alongside your 2 tonnes of cast iron, it's all relative!

ok I'll stick with that first one I found, save a few quid as well!

I manage to un cover some terminal labelling hiding under the grime! pictures attached of which wires do what..

Yes, as you say, the 380 windings are terminals 1.3 (Brown) and 1.1 (Blue). So leave the secondary as is, and separate the pairs of blue and brown and have a single blue, blue and brown together, single brown..? Hopefully I'm not being dumb and that's what you mean!

At this stage I think the money saving out weighs keeping the size and weight down, at least for now. So I'll persevere with this lump seeing as its quite close to where I need it to be. your not wrong though, it must weight somewhere between 50-60kg easily.

.Yes, single blue, blue and brown together, single brown would work.

You can always swap the transformer out later when things are running and you are feeling flush. Keep your beer tokens for other stuff.

My original transformer is even bigger, as it's a 3-phase jobby. Which reminds me, I must get it down the scrappie when it reopens....

Awesome, I'll get them swapped over today.

Yeah I probably will end up changing it long term to be honest, its not the nicest thing to look at beside the machine..

Yeah I would! probably fetch a fair amount in scrap for something that size!



Have you got any recommendations for relays, I've attached a wiring diagram for the controller I've bought which calls for a 24Vdc relay to switch the power for the drivers ( I think!).

Would a solid state work alright for this? or go for a non-latching type?

You could use an SSR but TBH, they can be a bit expensive. If you decide to go that route, make sure you buy a DC one, not an AC one. If you try to switch a DC load with an AC SSR, it will go on but you won't be able to turn it off. CPC have a very limited range but Farnell and RS have a lot more to choose from.

That schematic shows all of the axis drives passing through the relay. That would make it a BIG relay and switching DC is more tricky for the contacts than switching AC. If you go this route, you'll be needing a big clonker of a relay, possibly more of a contactor and make sure it's actually rated for the DC current and voltage you would switching in a fault condition. I haven't gone that route myself. I use the drive enable signal into the drivers to inhibit them if I need to use the e-stop or the controller / VFD get upset. If the worst comes to the worst, I could always kill the mains using the front panel isolator. Certainly, if you switch the high power DC loads, you probably won't want to be using an SSR unless your bank account is very healthy.

I have SSRs for the flood and air (300VAC type) and for the spindle gear solenoids (60VDC). My mains is switched with a front panel isolator and I have a time delay relay after that to soft start the 100VAC transformer / diode bridge / electrolytic circuit. The e-stop kills the drives and the Acorn and the VFD via the low voltage inhibit circuit. As I said, if that were to fail I'd kill the mains. This isn't an industrial safety critical system, just like the other systems here, so there's no point kidding yourself that there is one approach you must take.

If you go the route I went, there's no need for heavy relays in the power circuit. If not there are probably contactors out there that are in common usage for this kind of application that somebody here could recommend.

I wouldn't personally use an SSR for anything deemed critical, as a common SSR failure mode is to fail short circuit.
A suitably sized relay, or contactor is a more robust option.

That diagram you posted, it would appear the AXBB uses an output to power up the drives. Personally, I'd use that output as an enable signal to the drives, and have power controlled via a suitable e-stop system.

I'd agree it's better to run the enable signal to the drives rather than switch the DC. I don't see any role for an SSR here either, other than coolant etc.

As you say, it's simpler to run an AC side contactor on an e-stop circuit if you can be bothered, rather than switch the lower voltage DC side. Personally, I'd rather not throw everything out each time I hit the e-stop but that's a matter of preference.

Great info, cheers!

So it certainly sounds like a better option to tweak that layout somewhat, guess I need to have a think of what I want out of the set up. worryingly I hadn't even thought about an E-stop system yet!! I would at a minimum hope to have an e stop to kill the motors and spindle, whilst keeping the system powered still.

We had a machine at work that had floor mounted E-stops and god knows why but they literally killed everything, PC included, so you lost all your datums and tool offsets. Most annoying when you accidentally kicked them!! I don't want that...

Personally I wouldn't want that. On my machines, the e-stop stops the spindle turning and the table moving. The most likely cause would be senior moment / finger trouble causing the tool to dive into something unauthorised, or the tool clogging up and breaking, or realising the speeds and feeds are somewhat ambitious. Mistakes have been made!

If you have a crash, there's no need to kill the whole machine. And if the whole thing really goes native, the isolation switch will stop any arguments. Unless you are planning to sell your equipment commercially, it's your call what you do. There's no right or wrong answer although as there are humans involved, I'm sure there are strong opinions in favour of all manner of different solutions.

I’ve been having a look at what I have from the old system and if I could use any of it for the E stop circuit and how to go about it. Just wanted your opinions to see if this would work and not cause any damage to anything… or if its just stupid!

I’ve got at the moment a phase and neutral from the 3 phase supply coming into the cabinet that go into fuses, which then go into a large contactor. I was wondering if I could use that contactor set up with a momentary switch, E stop, and relay (which I have from the old setup which are all rated to 600Vac). Then on one of the switched terminals from the contactor could I then feed the 230Vac supply to the transformer which will power the drivers, and on another the supply that goes up to the spindle start button.

With the spindle I should mention it is currently still set up as it would have been when in manual operation. So it is not linked to the CNC side of the machine other than it used to run through a relay so you could only turn the spindle on when the pendant was powered up.

Just reading through what I have written, I guess I could just run the spindle feed through the same relay as the e stop and just switch the 230 through the contactor…

Hopefully someone can make sense of my awful explanation!

Officially all user buttons should be low voltage (IIRC in this context, it's sub 60V), which is why most control systems use 24VDC.

In an industrial setting, you would normally use a dedicated safety relay (essentially a relay with multiple contacts, that all get monitored, and should any fail, the relay won't activate).
However what happens when an E-stop gets pressed, is down to risk assessment. An E-stop should bring the machine to a stop in a safe and controlled manner. With modern servos, you would normally do that by triggering the servo drives fast stop input, then a few seconds later, remove drive power. This means the drive should bring the servo to a fast stop, then power down. If you simply cut power, the servo can keep moving then freewheel to an uncontrolled stop.

However, as it looks like you're using older DC servos, then you have to consider a servo runaway failure.
DC servos is one setup where I would seriously recommend where limit switches are part of the e-stop circuit, and the E-stop kills the drives as quickly as possible, as you really don't want a servo running past the limit switch under full power.

Makes sense, hadn't thought of it like that! Maybe the E stop button I have was rated to 60V and not 600V then, need my eyes checking!!

Thanks for the info, better get back to the drawing board on this one then!

Yo JimJam - wondering how you are getting on - I expect you will have been cutting chips by now!

Let us know what you are up to on this!

Murray

Hi Murray,

Still making some slow progress! Back at work now so I haven't had too much time to play with it.

Waiting on a few last electricary bits to arrive then I should have everything I need to get it going in some basic form! would be nice to see a bit of movement from it soon!

Being back at work means I might be able to knock up the adaptor bits I need to mount the encoders to the motors as well.

Did you have much trouble getting your motors and drivers set up and working properly? I forgot to order the USB programming stick for the drives, so it would be lovely if they just worked out the box. I have a sneaking suspicion that couldn't be further from reality though....

They may just work out of the box after a fashion (I don't recall) but it's unlikely the PID (tuning) settings will be magically anywhere near optimum.

Might be sensible to order a USB thing now, as you are going to need it at some point. It's their own interface as far as I can see so you can't just reuse something else.

Yeah it would be silly not to get one for the price, think postage will be more than the costs of it! Just waiting for them to get the payment info to me and It'll be on the way, been about 10 days since I ordered it and not heard from them. hopefully they're doing alright over there!

Started getting everything mounted up in the cabinet now, changed the layout god knows how many times. The base plate's like swiss cheese now!!

Got the mounts for the encoders machined up to bolt to the motors, they add a bit more length than I would have liked but I didn't want to start modifying and shortening the motor shafts really so I'll just have to live with it. Once (if) it's up and running I'll make a cover to protect them a bit more.

Also finished the new pulley mount for the Z axis ball screw. (and yes, I did reassemble the ball screw the wrong way after machining it...that was rather annoying to assemble twice!)

Starting to get somewhere now though!

Attachment 28296
Attachment 28293
Attachment 28294
Attachment 28295

I enquired about buying something small from them once - think it might have been a line driver board - but never got a reply. You might need to give them another nudge until you get a response or you may not see it. Larger orders possibly attract their attention more easily.

Dunno if you are aware that there is a cncdrive forum - you may be able to contact them through that http://www.forum.cncdrive.com/

Those housings look the part. Good to have access to decent machines when doing this sort of work.

Looks as if you actually replaced the Z axis ballscrew / ballnut? Did you enjoy machining the hardened steel? Sparks everywhere?

Yeah that did cross my mind, the initial order was quite a chunk of money and I had a reply within a few hours.. I'll give the forum a try, thanks for the link.

Thanks, it certainly makes things easier! Worth taking the extra effort to make them properly, hopefully less head aches with setting them up knowing they are machined properly so the alignment between them should be good.

I replaced them both in the end, the old screw was cracked and it was an odd size. Also needed new bearings as well so just replaced the lot. Can't beat a bit of hard machining, sparks were disappointing though! It was actually worryingly not that hard if I'm honest... looked to be case hardened around 3mm per side, the rest was really soft...

I guessed it wasn't going to be amazing quality based on the price, think its was £30 for the screw and nut... we'll see how it does.

Starting to get somewhere with wiring it all up now, just a quick query before I do something stupid...

so the wiring instructions for the main plug from the drive to AXBB board are:

The pinout of the Main connector is listed below:

1) Step signal input
2) Direction signal input
3) Ground for Step and Direction signals (NOT for power)
4) 5V power input. (if this input is fed then the 7.pin should not be connected!)
5) Reset (input) and Stop input)
6) Error (output) and Stop input
7) DC power + (12V DC input for digital processor)
8) DC power – (ground for the 12V DC power supply)

Step 3 says NOT for power, but the wiring diagram shows it connected to the 0V of the 5VDC power supply...

I think all it means is that you shouldn't connect the power supply 0V to it as the main connection to the board - that needs to go in to 8). The tracks from 3) to the rest of the circuits are probably very thin and come via the processor. May get damaged if you try to run the power supply connection in through them. But when you run the step/dir signals to the stepper / servo drive, you should take the associated 0V from 3) to give you the three wires you need. Better noise immunity.

The digital / logic circuits on the board are powered from the 5V either directly from 4) or by an internal step-down regulator fed from 7). Which ever you choose, the associated 0V to the power supply should connect to 8), not 3).

Hope that make sense....

Ahh yes, that makes perfect sense now you've said it!

I've got the one axis wired up and started to try and get it going, then it dawned on me I haven't got a clue what I'm doing!!

When its powered up the motor is quiet until you try to move the axis (by hand or by jogging on control software) then I just get a humming noise from the motor and that is all...

So I fired up the PID tuning software and its all nonsense to me, I cant even begin to understand how I would calculate the figures I need to enter...

I've attached a photo of what the settings are and the default values that were there.

The other thing is I'm not 100% on the wiring of the encoders, so it could also be that... or it could be the settings in UCCNC... could be any number of things with me involved!Attachment 28472

Stroke of luck managed to get the one axis working, when I looked into what all the parameters actually were I was able to get them down to somewhere useable, probably no where near perfect but at least it's a starting point!

Did you see this? http://www.cncdrive.com/downloads/help.pdf

Looks half sensible. I'll also see if I noted my own values somewhere for you to compare against.....

EDIT - that's annoying. I seem to have logged almost everything else I ever did on this machine but there's no trace of the PID settings in the servo tuning. Sorry about that. I seem to recall that you could plot the response in the software as you played with the numbers.

You start with I and D turned off (to zero) then turn up the P term until it gets unstable then back off. Then dial in some D term until it gets a bit nervous then back that off a bit. Play with turning P and D up together until it seems happy and the response is good. Finally dial up the I term until it starts hunting again. Then dial it back a bit and play with all 3 together until you are sick of it. Note down some of the settings that seem to be getting there because you will screw it up and forget what looked promising.

The I term doesn't make a lot of difference to the speed of response but you need it to eliminate steady state error. It's a bit of a black art and there's loads of bollocks talked about it but the above is the guts of it. There are autotuners in some of these programs but in my experience they are usually pretty crap. I've done several courses on digital control systems over the years and I still end up doing a lot of it by trial and error.

Nope didn't see that one, thank you! I'll give that a go today.

I spent a couple of hours trying to look into understanding it and how it all works but it went straight over my head!

Had another strange issues yesterday.. when I connect the 5V DC power to all three drive units none of them work, but when I unplug one of them the other two work. it doesn't matter which one I unplug the others will then fire up. power supply issues? I am only using one unit to feed the three drives and the controller board....

I have got a 12V DC unit somewhere which I'm pretty sure can be used for the drive units, might give that a go see if that makes a difference.

Yes, sounds as if the 5V supply can't supply enough current to get more than 2 of them started up. There may be an inrush current surge at turn-on that is upsetting the PSU. As you say, feeding 12V into pin 7 is an alternative. It may also be safer if there's an internal regulator in each of the drives to drop it to 5V by the sounds of it. Pretty sure that's what I did for my 3 drives.

Finishing up my degree has taken up the last couple of months, so progress has been pretty slow up to now.

I had the machine roughly wired up to test everything before spending the time making it look tidy, the machine was working great and ran through some programs with no problems. It needed a few tweaks to the settings to get the dimensional positions a little closer but I didn't want to spend too much time on that until it was together for the final time.

Since re wiring it I'm having a nightmare! I managed to get the transformer wiring wrong on the primary coil, so when I fired that up to check the DC voltage output the cabinet filled with some lovely black smoke from the braking circuit. luckily I had unplugged everything else at this point so I didn't cause too much damage (hopefully).
I didn't stick around to check the actual voltage that was being output from it but could this have potential caused damage to the bridge rectifier and capacitor?

I re wired the transformer and checked the voltage output and its back to what I was expecting, but since wiring the drivers and motors back up I'm having power issues with all three drivers when I try to tune them. They all seem to get an voltage error and go into error mode..
I don't know at this point if it's power supply to the drives that's the issue or if its referring to the power to the motors.. need to do a bit of reading on the error codes.


I also managed to snap the silly little pins the USB connection boards plugs into on one of the drives so that's going to need replacing/repairing!

So all in all I've taken a few steps back at the moment! :thumbdown:

If you have blown your braking module and the DC voltage is close to the max limit, you may be getting an overvoltage condition when you decelerate your axes. Slowing down / stopping movement will generate voltage on the bus. Perhaps you could try reducing the feed rates in the controller to start with and see if that helps.

What DC voltage have you got? You need a good 30-40V below the overvoltage threshold. That threshold is probably shown in the manual.

You may need to replace the braking module unless you want to see glacial feedrates. But first see if this is the issue by slowing things right down to reduce any overshoot. If you have a scope or DVM, you may be able to observe what is happening to the DC voltage when you decelerate.

I think I'm on the low side on the DC voltage already at 130V, the drives are rated at 160V 36A. I haven't actually checked the Amps on mine though, cant imagine it being higher than that though?

The braking circuit didn't actually get damaged in use, it started smoking the second I turned on the power when I had the wires mixed on the transformer... I will get a new module on the way, like you say it will probably cause me more issues leaving a potentially (definitely) damaged unit wired up.

I'll try reducing the feed rates as well and see if that helps, it does seem that they initially start to move the axis a rotation or so then goes into error state.

I also had a rather scary issue with one axis, when I hit the reset button the axis just went into rapid on its own! Managed to hit the E stop before it done any mechanical damage to the machine ( hadn't wired up the limit switches at this stage) So i may have some encoder issues as well on that motor.

You may have that motor connected up incorrectly, as you say. If your A and B encoder wires are swapped over, it will think the motor is going in the opposite direction. That causes the servo to run away. Or you might have 2 or 3 of the phase connections swapped around.

130vdc doesn't sound bad. I don't recall precisely but I think the voltage clamp / braking module starts clipping at around 170V or so, so you shouldn't be seeing a problem.

Shouldn't take a lot of current to jog the axes. 35A should be more than enough. You would be overcoming a large force to require that sort of power, possibly break something. I've jogged my table against a hard end stop a couple of times. Luckily on my system, the drive trips before anything snaps but that's down to the design of your machine. You could really f*ck something up, like a ballscrew / ballnut or its mount / yoke. Worth seeing what would take the load in that event and convincing yourself it would be OK come the crunch.

Going into error when starting to move doesn't sound like overvoltage, which would only happen if you try to bring an axis to a rapid stop. I wonder if it's an encoder / motor issue?

I'd disconnect 2 of the drives and remove the belt / coupling from the third. Then try to run the servo just using the tuning software. No point getting the controller involved yet. When that works, reconnect the ballscrew and try tuning the response with the software. Finally, connect up the controller and see how it plays together.

Right, so I've literally taken it right back to basics as you suggested disconnected all drivers except from the one. I was still having a weird issues where it would go into error state as soon as it was powered up, so I went back to the original 5Vdc PSU to power the drive rather than the 12Vdc and that solved that issue. Then I reset the firmware back to factory encase I had done something stupid along the way with the settings.

Then I connected the motor and the 130Vdc supply to the drive, and as soon as I hit the E-stop reset to power it up the motor continuously rotates until it reaches the max error in the settings.

At the moment I'm thinking either encoder is faulty (but all 3 do the same thing) or damaged drivers?

The only other thing I can think of that's different from previously is that I extended the encoder cables. I have double and triple checked the wiring is correct and swapping A and B has no effect. I extended it using CAT 5e cable, could this be causing some interference/noise giving a bad signal back to the drive?

I used shielded cable but the twisted pairs weren't shielded between each other, just the outside of the cable...

I also clipped the cable to the outside of the conduit supply the power to the motors, would this cause an issue?

Are you using the cncdrive Servoconfigurator3 software to test and debug the servos? https://www.cncdrive.com/downloads.html It sounds like ti from what you say.

The behaviour sounds like classic servo runaway which happens when the encoder signal is the wrong polarity relative to how the motor wants to spin. So in this case the drive tries to reduce the position error by going even further in the wrong direction. It doesn't know any better.

You should be able to talk to the drive without the motor power enabled ie provide the 5V but not the 130V. Then the s/w should see the drive and the encoder without it spinning off into the bushes. You should be able to turn the motor and check the operation and polarity of the encoder signal. It's a good idea to check the encoder signal makes some sort of sense before going any further.

With a DC brushed motor, you can't get the phases mixed up because there aren't any. But you can swap over the motor polarity by simply swapping the wires over. That's the 2 heavy wires from the drive to the motor. You'd think this should stop the runaway by changing the direction of the motor. Try that and see how you get on....

The motor wires don't need to be screened to work properly but it's good practice. Many of the problems people encounter seem to be caused by electrical noise, usually because of poor wiring practice. You don't need to be an expert but the methods are well understood and explained fairly well in places. I'm sure you will be able to find some good pointers if you look for them and we can always help you with suggestions.

You must be getting pretty close now!

Quote:

---

The only other thing I can think of that's different from previously is that I extended the encoder cables. I have double and triple checked the wiring is correct and swapping A and B has no effect.

---

I would have thought that this SHOULD have some effect as have been mentioned way back in the post like runaway.

I don't know your drives, you mention cat5 do the drives have a RJ45 connector or hard wired ? I have seen many times that using the RJ45 with crappy plugs the pins don't always connect.

Keep at it you will get there.

Quote:

---

Originally Posted by Clive S

I would have thought that this SHOULD have some effect as have been mentioned way back in the post like runaway.

---

Indeed - so first check that the encoder is talking sense. Given that the position error stacks up and finally trips the drive, it sounds so. But as as you say, swapping encoder phasing should have done the trick. So start from the beginning, one step at a time......

Quote:

---

Originally Posted by Clive S

I would have thought that this SHOULD have some effect as have been mentioned way back in the post like runaway.

I don't know your drives, you mention cat5 do the drives have a RJ45 connector or hard wired ? I have seen many times that using the RJ45 with crappy plugs the pins don't always connect.

Keep at it you will get there.

---

Yes, two RJ45. One for the encoder connection and one to the motion controller. Just to add another potential issue, it has a differential line driver, which from what I understand converts single ended to differential signals. The encoder is hard wired to the line driver which then has an RJ45 plug on the other end which then goes to an RJ45 on the motor driver.

I also made up all of the cables with the RJ45 plugs on, I did test them all before installing them but I guess that might not be a bad place to start looking for bad connections as well.

The only major change I made from when it was running was the encoder wiring, so I'm hoping its something silly like a bad connection!

Thanks for the help guys, I'll get stuck into it again tonight and see what I can find.

Quote:

---

I also made up all of the cables with the RJ45 plugs on, I did test them all before installing them but I guess that might not be a bad place to start looking for bad connections as well.

---

OK. Did you by any chance use the type of RJ45 that has a slot at the front that lets you pass the wires through then crimp then trim off the excess.

If so I have used them and chucked them out. Recently I made a 20Mtr patch cable (not cnc related) checked it with a cat5 tester that showed the errors. I did a continuity test pin to pin and it was fine.

But putting it into place it would not work. It turned out that the pins would not mate into the sockets properly.

On mine, I used standard Ethernet network cables and just cut the connectors off one end. I then just wired them into my controller. I believe cncdrive used to sell interfaces to allow everything to use RJ45 connectors and there's a guy in the US trying to do the same. Seemed rather pointless and expensive to me - and hardly industrial practice.

Using the s/w to see the encoder should tell you if it's connecting and if it's sending sensible info.

I just reverted the one axis back to how I had it set up before, I cut the CAT5 extended cable from the encoder back to the factory wiring, and used a premade ethernet cable from line driver to drivers and its working again!

I also tried the premade cable with the extended encoder cable which didn't work. so it would appear it didn't like the cat5 cable I added on.... I checked continuity from the encoder to the end of the extended piece and everything seemed ok. That said I don't know enough about electronics or the workings of an encoder, could I have been getting a false reading due to internal components being linked? ( I didn't unsolder the terminations at the encoder when testing)

I soldered and heat shrinked the individual cables as well as around all of them, so I shouldn't have been getting any kind of short where I joined it....

That's progress at least. You can buy cable testers like this https://www.screwfix.com/p/philex-ne...e-tester/93219 although obviously Screwfix isn't the best place for low prices. They simply apply voltage to each wire in turn. A remote display shows you if the sequence is wrong.

Sounds as if you should be cutting chips this weekend, then! Well perhaps not quite but things are on the up.

CAT5 plugs can be wired in two different ways T568A and T568B. You will even find 'Null Modem' cables with different wiring at each end. This might contribute to the problem. See disgrams on the link.

https://incentre.net/ethernet-cable-...oding-diagram/