PDA

View Full Version : Help with Problem X axis Stepper Please



howser37
20-12-2016, 06:11 PM
My Machine is a Heiz S1400 and until today it has been fairly trouble free, however during a cut today the x axis racked and stalled.

I have twin steppers on the X axis and at first i thought i had narrowed it down to one of these steppers, however on switching over the control cables and seing the propblem move from one x axis stepper to the other i then realised it was an output problem frm the control box.

I have taken the control box apart and the first thing i have noticed is what seems to be a circular crack in one of the driver chips TA8435HQ.

Is this likely to be the culprit ?, all 3 other chips seem intact

and if so does this need replacing with exactly the same chip as the others are TA8435HQ too.

i am assuming this will be a relatively simple soldering job, just desolder the old one and replace?

I will upload a picture of the chip in a minute

howser37
20-12-2016, 06:22 PM
2005820059

Here are the pictures

Clive S
20-12-2016, 06:32 PM
does this need replacing with exactly the same chip as the others are TA8435HQ too.

Can't answer that but you can buy them here:-
http://www.ebay.co.uk/itm/1PCS-TA8435HQ-TA8435H-TA8435-NEW-TOSHIBA-ZIP-/251057536704

I would also check the current rating you have set.

GND
20-12-2016, 08:31 PM
I'd say it's definitely the first thing to try - that driver chip is clearly dead! Looks from your photos like they are socketed though, so no soldering required. Just remove the old one and replace like-for-like, being careful not to bend any of the pins.

Also, check to see if the chip needs to be electrically isolated from the heatsink. Some do, some don't - take your lead from what is currently there, and if the back of the chip is metal. If it needs to be isolated, then there will be evidence of an insulating pad or washer used between the chip and the heatsink, and then you need to take a lot of care when you fit it all together, testing it for isolation with a meter. But maybe it is just bolted together. Looks like there is at least some heatsink compound used, but what lurks in between is not visible in your photos. Try and salvage as much of that as you can for the new chip - ideally you'd buy a small tube off eBay and smear it on to help conduct heat from the new chip.

Hope that helps!
Graeme

howser37
20-12-2016, 08:43 PM
Thanks GND and Clive S. Yeah that will be handy if the chip is socketed, im crap at soldering to be fair. I think they are just bolted together with small bolts with thermal compund between the heatsink. I have a tube of MX4 Processor compound from building PC's so im thinking that would work.

That chip definately is the one leading to the X axis output with the problem so i will replace it and see if that fixes it. There is one on ebay available from the uk but it is missing the Q designation at the end so not sure if it would be ok TA8435H rather than TA8435HQ. I will order one of the ones Clive suggested but might be a little wait for that one to arrive.

regards
Adam

GND
20-12-2016, 09:00 PM
I had a quick look at this device's datasheet, and the only difference between the H and the HQ versions seems to be that the HQ is tinned with lead based solder, so it's not "lead free" as is required these days for commercial use. But for your purposes it really makes no odds - go for either and you'll be fine!

It also said that the rear of the chip is metal and connected internally, so you need to take precautions. However depending on what the heatsink is connected to, then it may have been designed to be non insulated, as that part of the chip appeared to be grounded. But just be aware of the options here!

howser37
20-12-2016, 09:11 PM
Thanks GND, i have ordered the untinned one and will try that. I will look at the other other chips for guidance on how its connected to the heatsink.

GND
20-12-2016, 10:10 PM
Sounds like a plan! Let us know how it goes....

Cheers
Graeme

john swift
21-12-2016, 12:44 PM
either the TA8435HQ stepper driver IC needs to be electricaly isolated from the heatsink or the heatsink needs to be grounded
otherwise a faulty IC will damage other ICs mounted on the same heatsink !

take a close look at the discoloured contacts
20061

check the socket for heat damage that may of softened the spring contact
so it will no longer grip the new IC's pins as tight as it should


John

GND
21-12-2016, 01:06 PM
Good clarification, John - I assumed as much but I was in a bit of a rush and didn't explain myself very well last night! The back of the chip is internally connected to ground, so if the designer of the equipment didn't insulate it from the common heatsink, then the heatsink must be grounded as you say. I'd argue it's not ideal practice, and that insulating the chips individually is perhaps preferable. But best just reproduce what the manufacturer did, as it seems to work. Unless that was the cause of the failure of course....!

Well spotted regarding the contacts - they do look a bit iffy! Those are "turned pin" type sockets - high quality ones, which is a good start. But definitely worth following John's suggestion to check them carefully, or even replace them. That of course involves soldering :-)

Graeme

howser37
21-12-2016, 02:09 PM
Thanks John and GND. I will check the pins for damage

howser37
26-12-2016, 04:34 PM
Just in case anyone else is searching the forums about a similar issue, i replaced the damaged driver chip with a new one and the axis is now working fine again. There was a small silicone pad on the back of the old driver chip and thermal grease so i used this pad on the new chip and applied MX4 Thermal non electrically conductive paste to the back of the chip.

I did check that all the socket pins were working correctly and were not damaged prior to installing the new chip. The only slightly difficult thing was getting the pins into the socket without bending them as they are quite delicate. I did several hours of cutting last night with no issues

Clive S
26-12-2016, 06:40 PM
Its always satisfying to get something that was broke working again:yahoo:

GND
27-12-2016, 10:54 AM
Nice work! Hopefully a one-time glitch that won't ever happen again!

Cheers
Graeme

howser37
01-02-2017, 07:47 PM
Oh well, bad news its gone pop again, machine was not even cutting at the time just idling then pop and burning smell, opened up the controller and the same chip has blown again (the new one).

Any ideas what the most likely culprit could be ?

john swift
01-02-2017, 08:33 PM
if the replacement TA8435HQ was not a fake / factory reject

then it could be an intermittent break in the motor cables or connectors

damage to the current sense resistors and the PCB tracks
if the reference voltage to the IC is adjustable
the preset potentiometer may of become intermittent

plan B could be to unplug the stepper IC and connect a single axis stepper driver board to replace the faulty axis drive

a TBH6064 board will be a better choice than the infamous TB6560

John

PS
http://sound.whsites.net/fake/counterfeit-p1.htmS

howser37
01-02-2017, 09:33 PM
Thanks John, thanks for the warning about the fakes, could be i guess, just seems a bit too uncanny that it would blow so soon after the original non fake on that was in it as that was the original chip.

Im seriously considering building a new box from the ground up, if theres one thing i dont like its worrying whether the machines gonna blow up during a cut, especially as i cut expensive material usually so it can cost me big if it craps out.

Ive alwasy been a bit dubious about this control box to be honest

GND
02-02-2017, 11:56 AM
Can't really add much over John's useful suggestions. It seems clear now that there is a fault beyond the chip - this second failure suggests the first chip didn't just expire for no reason. So checking for decent connectivity to each physical pin using a multimeter could be worthwhile. Ideally right to the motor connections, and to the other connected components on the board. Maybe an intermittent fault though, like a cracked PCB track, which may measure fine most of the time, but open up as things get warm.

Maybe it is time to get new electronics as you say? Particularly if the risk of failure is costly to you. I hate to waste perfectly serviceable kit, but sometimes pragmatism has to take over!

Cheers
Graeme

howser37
02-02-2017, 02:16 PM
Hi Graeme, im pretty much resigned to the fact im going to build a new control box now, i do intend to build a new CNC myself at some point so it makes sense to get learning now and also i may well overspec it so that i can use it with my new machine when i get it built.

In the mean time i am wondering if it would be as simple as possibly utilising the spare driver in this control box as the second X axis driver, as there is a 4th axis driver in the box that is not being utilised as i do not have a rotary axis, not quite sure what would need to be changed in terms of settings as the other 4 drivers are on one board and the rotary is on a seperate board

howser37
02-02-2017, 02:47 PM
Think i have found the likely culprit, looks like a cracked track on the back of the pcb, i will post a picture up

GND
02-02-2017, 02:52 PM
Sounds promising!!

howser37
02-02-2017, 03:11 PM
20567

Here is the picture of the cracked track, is this repairable?. Definitely seems to be the culprit as it is on the correct chip and even leading to the correct area of the chip that has popped both times.

GND
02-02-2017, 03:18 PM
Definitely repairable! I'd suggest you cut out that offending length of burnt track so it's not flapping around in the breeze, and then re-make the connection with a short length of thin solid core insulated wire. You're lucky in that it essentially connects two of the chip pins together, so you can use the PCB pads for those pins as end points on which to solder your little wire. Pull the chip out and d the work. Given a few minutes with the chip datasheet we could also work out which pin has become disconnected, and hence what the possible issue might be.

Now of course it could be that the track got burned out as the chip failed, and hence wasn't the cause of the failure - but more the result. Or indeed, has the chip actually failed? Could the board have failed whilst the chip is still OK? Looks like that track took a bit of an overload, but why is that?

Datasheet, datasheet....

howser37
02-02-2017, 03:27 PM
Thanks Graeme, i will see if i can find the data sheet online.

I would say both chips have definately failed, the black plastic has popped in both cases and in the case of this particular chip i opened up the contol box fairly quickly and the place where it had popped was still melted and gooey. But like you say whether this failure resulted in blowing the track or wwhether the track blew as a result is a different matter. I will admit i was not particularly thorough last time i replaced the chip in checking the back of the pcb, i just tested the socket pins were gripping ok and only quickly scanned the back so it is possible that this crack was there prior to replacing the chip.

This failure did take out an 8 amp slow blow fuse on the power supply too

GND
02-02-2017, 03:27 PM
OK, so a quick look at the datasheet suggests that the "disconnected" pin is pin 15, and that it is normally connected to pin 24. This makes sense and these are the two main voltage supply inputs for the motor drive stages. Hence if one of the motor windings was taking too much current, that thin track could fail. Looks a bit thin to me for a motor supply line? Designed to take an amp or two? Hence maybe it simply burned out and effectively disconnected the motor? Worth a repair and a retest?

GND
02-02-2017, 03:33 PM
OK, just seen your latest email - scrub the retest! New chip required, but the question is - why did it fail? The motor must be taking too much current at some stage thus blowing the chip and in this case taking the voltage rail track out too, a bit like a fuse. I suspect it must have been intact last time or your repair would not have worked. Hence I think this is a symptom of the latest failure rather than its cause.

I think I'd be checking my motor wiring for issues, because if there is a problem there, you'd potentially end up blowing whatever you connect - new drivers for example - until that is fixed!

howser37
02-02-2017, 03:34 PM
Thanks Graeme, i will find a magnifying glass and attempt the repair this evening and will test it tomorrow and see what happens, if it blows up it just means i get my new control box more urgently lol

I will also take a look at the motor wiring too

Although the track is clearly damaged and fractured it was making contact when i was looking at it, so is it possible that if it was still making some kind of contact that it would have worked with the new chip for a short period of time?

GND
02-02-2017, 03:37 PM
You'll still need a new chip though? Or can you steal one from the unused 4th axis driver?

Either way, definitely check the motor wiring for shorts first!!

howser37
02-02-2017, 03:40 PM
I assume i can steal the chip from the 4th axis without it causing a problem being absent. I also have a new one on the way that i ordered last night online

GND
02-02-2017, 03:44 PM
I'd say removing a driver should be fine - nothing connected to its outputs anyway, so a useful source for another chip.

But I will emphasise again - there must be something causing these to draw too much current, so without finding the root cause, it will likely just keep happening....

howser37
02-02-2017, 03:44 PM
I wish i could remember if i switched the x axis ports around last time i replaced the chip as both x axis motors are interchangeable, i think i did which would help elimenate the motor wiring as it would probably blow a different chip, but i just cant be sure if did infact switch them.

I do know whci axis motor failed last time though so i will switch ports this time after checking it so that it would blow a different driver

GND
02-02-2017, 03:46 PM
Can't help you there!!

john swift
02-02-2017, 03:52 PM
after looking at the latest photo
in addition to adding a wire link to replace the damaged track
I would resolder the current sense resistor as well

20568

John

PS
if you switch the two X motors now
then if it fails again you will know if the fault goes with the motor

howser37
02-02-2017, 03:55 PM
Ok thanks John, will do

howser37
07-02-2017, 01:26 AM
Ok im gonna have a go at this repair tomorrow but just want to check a few things first with you guys as my electronics knowledge is minimal lol,so a few questions.

1. For the track jumper repair i can only find/source solid core wire with 0.6mm core unless i use something ridiculously thick like a core from lighting twin and earth, however im slightly worried by the max amp rating of 1.8 amp for the 0.6 solid core wire as i am not sure how much current goes through this track. Would i not be better using say 3 amp stranded wire?.

2. I would like to ground the heatsinks to avoid the need for the micah washers to isolate the driver chips agsinst the heatsink. However is this simply a case of grounding the mounting bolt of each heatsink to the case via maybe one of the standoffs on the corner of the pcb?. Or ground it to where the incoming earth is grounded to the case? Or is it more complicated than that?. Grounding seems to be a bit of a dark art reading about it on some threads.

3. I want to check the wiring from the db9 to the stepper motor as suggested for a short or breaks etc, is there a method for doing this, i have a multimeter but ive not much idea how to use it other than the continuity test.

Thanks
Adam

GND
07-02-2017, 11:42 AM
Hi Adam,

To answer your questions;

1. I had suggested solid core wire simply because it is easier to work with on the back of a pcb, but your concerns about ratings are legitimate, and so I'd say try the stranded wire and see how you get on. Strip the ends, twist the strands together, and then tin them before you try soldering to the board. Sometimes stripping a bit overlength is good, so that once tinned you can trim them right back to something that will neatly sit on the PCB pad. All you need is a couple of mm exposed at each end ideally for the actual repair.

2. Grounding is a whole subject in itself! In simple terms, the issue is that wires and connections aren't perfect, and don't have zero resistance. Hence although you might connect things together to form a common ground, there are always tiny differences in voltage across them. This is especially the case when large currents are involved, of if there are multiple paths between the grounded items. And since ground is your reference for everything, then this can cause issues in operation. Hence care has to be taken as to how you handle grounds, and that is why it's a big deal. In your case, I know the mica washers are a pain, but the original designers specified them for a reason. I'd be reluctant to change things and ground the heatsinks, even if it looked like it should be possible. With a system that is already a bit unreliable, I think such a change would be very unwise.

3. I guess we're looking for short and open circuits on the interconnections to the motor. So, to check for open circuits, put the meter into "Ohms" mode, and select a low resistance range (if ranges are selectable!). Disconnect the motor, and probe each end of the cable assembly on matching pins, and check that the reading is low - ie. there is a connection. Should be one or two ohms max. I'd also flex the cable when doing this - although that might need three hands! - to make sure there isn't an intermittent fault. Try and test as much of the connection as you can, so start on the PCB at one end, and finish at the motor connections themselves if at all possible. Access to connections may however be limited, so you may need to compromise this, but the closer to the PCB and motor you can test, the more of the system you are actually checking.

Open circuits - ie. broken connections - are one thing, but since you are blowing your chips, the short circuits are the more likely culprit, as these are more likely to cause large currents to flow. So you have to try and find where your motor wires are shorting to. Here you should again disconnect the motor (to avoid its low resistance coils giving you false readings) and then pick a motor drive pin and apply one of your meter probes. Then touch the other probe onto each of the other motor wires in turn. You should not see very low resistance readings, otherwise it indicates a short circuit. Again flex the cable when doing this, especially near the connectors, as the fault is likely intermittent. If all are OK, move onto another pin, checking if it is shorting to any of the others. If the cable is still connected to the PCB, you may get erroneous readings due to the driver components, which sit across the motor connections - especially protection diodes - so you may have to just check the cable itself, totally disconnected. See how it goes....

Hope that helps!
Graeme

Robin Hewitt
07-02-2017, 12:37 PM
Looks to me like that track had a problem, someone tried to fix it by gobbing solder over it, found the solder wouldn't bridge the gap until some of the tin had boiled out of it, so he delayed and destroyed the glue holding the track down to the PCB.

After 30+ years soldering for a living I could fix that in a jiffy. OTOH I can remember what it was like when I started so I feel your pain :crushed:

howser37
07-02-2017, 01:09 PM
Thanks Graeme. Much apreciated. I read a few things about the mica washers not being as effective at getting the heat into the sink, hence the idea of just having the thermal compound. But you are right i will not start messing with things like the grounding in a system that i already dont trust lol. Thanks for the info on testing the wires, im pretty good with mechanical things but electronics baffles me at the moment.

Hi Robin, thanks for the input. You know now that you say it i dont know why this didnt occur to me that this looked like a previous repair. But it does look like a soldered over track and the solder gas blown, but again i guess its back to what caused the first failure.

Anyway i will report back what happens

howser37
08-02-2017, 03:17 PM
Hi guys just a quick update and a query, i have cut away the broken track as suggested and eventually after a bit of trial and error have managed to solder a jumper nicely to the right terminals on the back of the pcb, quite a task as the pins are very close together :-o. replaced the IC chip and also upgraded the cooling by having a much beefier fan suspended from the top of the contol box onto the Driver chips and heatsink while i was at it.

I am now checking the wiring on the suspect X axis motor, i have attached a picture to show what i am dealing with, i want access to these termianals but they are burried inside these 3M terminal blocks, i popped one out and the brown driver wire literally just fell off whereas the other two needed a bit of removal force, so i guess even that brown wire could be a potential short. To be honest im not too impressed with these 3M terminals, they seem a bit fiddly and precarious, what kind of terminal are you guys using for this, wouldnt something i can get a tester on at a later date without having to rip the thing apart be more suitable ?. Id like to remove them all to test the cable and replace them with something a bit less permanent and more robust if possible

20703

GND
08-02-2017, 03:23 PM
Not sure the picture got attached. Interested to see what you're dealing with before offering suggestions.....

howser37
08-02-2017, 03:38 PM
Hi Graeme i have attached it now, i believe they are grease filled scotchlock terminals, maybe they are good they just seem a bit awkward and open to failure to me and difficult to get apart without potentialy damaging the connections.

GND
08-02-2017, 03:48 PM
They look like some kind of one-time crimped connections, which are highly protected against shorts and therefore also restricting access to the signals for testing. Probably a good choice for the job, but definitely not very serviceable! I'm sure someone else on the forum will have a proper name for them....

I use good old fashioned chocolate blocks for my connections. They have their flaws, but they are cheap and easy to use - and probe-able for testing! You could always start with them and see how t goes, at least to get back up and running?

Cheers
Graeme

howser37
08-02-2017, 04:03 PM
Hi Graeme, i have some chock block connections here so i will give them a try for now and see how i get on. I can see how the current connections are highly protected against a short but im wondering about a break as i highly suspect that brown motor wire was not connected well, and i know that these drivers cannot be hot unplugged without blowing the drivers as there is no protection, and ive heard many stories of people unplugging the cables whils energised and blowing the drivers, i wonder if a poor or nearly broken connection here could intermittently simulate a similar event.

GND
08-02-2017, 04:09 PM
Well, it's certainly one possibility! Crimped connections are great when they work, but can fail without it being obvious to look at. If that brown wire came out without much effort, then it suggests it wasn't right.

Looks to me like you have those motors wires in parallel, as there are four connections with two motor wires to each. Disconnecting one winding would have reduced the current on that phase, but would also imbalance the load on the chip. Quite what effect that would have is beyond my experience, but who knows. Certainly something to fix though!

howser37
08-02-2017, 08:16 PM
Thanks Graeme, well i am back up and running again. I removed the x axis stepper cable completely and it tested fine, i then rewired it using different connectors and fired it up, all seems to be running fine at the moment, but it was running fine last time i replaced the chip for a good few hours too, so we will see what happens. I have switched the over the x axis stepper cables so that if it is a problem from the motor/ cable it should blow a different driver next time and atleast i will have narrowed it down a bit more, if the same driver blows i can probably assume its more something inside the control box.

Thanks for all the help so far, fingers crossed :-o

GND
08-02-2017, 10:32 PM
Sounds good! Let us know how it goes.....!

Neale
08-02-2017, 11:03 PM
I've used choc block in exactly this situation and it seems to work OK but I have wondered about these (http://www.screwfix.com/p/3-way-lever-connector-222-series-pack-of-50/76776) instead. Spring-loaded so more vibration-proof and even include access for a test probe. They come in 2, 3, and 5-way versions which could be useful in some of the junction boxes on my machine where, say, multiple limit switches with common ground connections all meet. I find it fiddly trying to use choc block for multiple connections as you end up with two wires in the same hole, which never seems satisfactory.

GND
08-02-2017, 11:14 PM
They look really quite smart, and certainly seem to get the thumbs up from the electricians, judging by the reviews on there. They don't look like they have a way to fix them down though, in the way you might screw a choc block to a panel. So worth bearing in mind for some applications. Don't some forum members use something similar on DIN rails in their control cabinets? I seem to have seen something along those lines in pics on the build logs.....

cropwell
09-02-2017, 10:07 AM
I bought this kit from screwfix http://www.screwfix.com/p/wago-basic-installer-box-75pcs/48808#_=p I haven't used them on anything CNC yet, but for solid core domestic wiring they are very good and quick. For multi strand wires I crimped an uninsulated ferrule on to get the wire in to the block. That was fiddly, but better than choc blocks.

Rob

Neale
09-02-2017, 12:47 PM
I had a couple left behind by an electrician after some work in the house which I have used on my own router and like you, I crimped on ferrules to stranded wires. They seem very effective, slightly larger than the equivalent choc block but easier to wire. I just left mine floating in one of the plastic boxes I use as junction boxes (which I was buying from Maplin but now 3d print).

cropwell
09-02-2017, 02:17 PM
I had a couple left behind by an electrician after some work in the house which I have used on my own router and like you, I crimped on ferrules to stranded wires. They seem very effective, slightly larger than the equivalent choc block but easier to wire. I just left mine floating in one of the plastic boxes I use as junction boxes (which I was buying from Maplin but now 3d print).

Wago do junction boxes, you might like to adapt your 3D prints accordingly http://www.screwfix.com/p/wago-abox-lever-connector-222-series-junction-box/9390G?kpid=9390G&cm_mmc=Google-_-Product%20Listing%20Ads-_-Sales%20Tracking-_-sales%20tracking%20url&gclid=CNbdvpGag9ICFa237QodaegHvA

Neale
09-02-2017, 05:01 PM
Custom printed boxes including guides for the Wago connectors, which could be tweaked to suit whatever combination of connector sizes needed. Sounds like a neat idea - thanks for the pointer, Rob.

cropwell
09-02-2017, 05:27 PM
Custom printed boxes including guides for the Wago connectors, which could be tweaked to suit whatever combination of connector sizes needed. Sounds like a neat idea - thanks for the pointer, Rob.

You could also print these https://www.tlc-direct.co.uk/Products/WA332.html

Neale
09-02-2017, 05:31 PM
Now those do look like a good idea, and cheap as well. Look much easier to use than screwdown DIN terminals which are a pain to use in a crowded control box.