Spindle Error (PCNC 1100 Series 3)

[popspipes]

I appreciate your sympathy. It really drives home how much my life plans are reliant on a working mill when it suddenly goes down. As I can't go back in time and buy the extended warranty, I should start saving up an emergency repair fund.

Definitely worth re-checking every part of the electrical system. I would absolutely love to find a stray chip or frayed wire somewhere!

I don't really know what an 'intermittent ground' is, although I am now researching this. Can it mean that the mains supply to the building is imperfect? I'm very keen to figure out how this fault happened in the first place and remedy it. So is it possible that an intermittent ground gradually caused damage to the mill's electronics over time? I've been at this property for just over a year.

those can cause big problems too (power lines, grounds etc.), I had one last year, loosing one leg of the input power intermittently or partially took a bit of doing to find that one!

What I was referring to is a ground circuit in your machine or on a circuit board, their can be several, and they arent easy to find unless its obvious (burnt connection, discoloration etc.)

Vibration can be a big cause of this, loosening lead connections etc.

I had one on the X stepper motor driver recently, it was relatively easy to find, it was missing steps and odd noises occasionally, luckily it got bad enough that I found a burned contact in a connector, found it visually.

[Zetopan]

Just checked this:
Auto mode: 0V regardless of spindle command from PathPilot or if the spindle is actually turning.
Manual mode: 4.98V regardless of spindle on/off/speed.
I was able to remove and replace the socketed microcontroller, sadly with no effect. There was no evidence of misalignment or bent pins etc.

That suggests that the front panel switch and possibly the microcontroller are not at fault. The latter is still somewhat problematic since you have not tested all of the microcontroller functionality, but I would not really worry about a defective microcontroller at this point since the counter may be at fault.

This diagnosis is the only thing I'll be using an oscilloscope for ...

That is a completely unwarranted assumption on your part. Once you learn how to use a highly portable oscilloscope it will become your instrument of choice. A DMM is fine for accurate DC and low frequency sine wave AC measurements, but it cannot tell you about slow rise and fall times, noise spikes and a host of other things involving digital or analog signals that vary significantly with time. A narrow pulse with a low duty cycle may not even register on a DMM, but be very visible with a scope. Likewise with a double pulse where a single pulse should be located (or the inverse) which are generally undetectable when using a DMM, etc.

I did locate an abbreviated user manual for the DSO 094 if you want to examine it:
http://www.jyetech.com/Products/LcdS...Manual_094.pdf

The standalone unit seems like the better choice, but the long delivery time is unappealing.

Here is a potential temporary workaround. If you are sure that you want to replace the CD4024 (which appears to be the culprit at this point) then ground yourself and carefully solder a short wire from the pin 12 pad (at the control board surface) to the pin 11 pad. Now carefully clip pin 12 at the DIP package so that pin 12 of the down counter no longer connects to the control board. At this point the output from the down counter first divisor stage (divide by 2) to the microcontroller input has been disconnected and replaced with the output from the second divisor stage (divide by 4). If the down counter pin 12 output driver was the problem then you should see about 2.5 V on your DMM on pin 11 when in AUTO and commanding a spindle speed with PathPilot. Likewise, using PathPilot to command a spindle speed should vary the actual spindle speed, but it will be slow by a factor of 2X (half speed). If you want 2000 RPM you would need to command 4000 RPM, etc.

If all of that works you can use the mill up to about 2,600 actual RPM by commanding up to 5,200 RPM. Path Pilot may prevent you from commanding anything higher than 5,200 spindle RPM, hence the temporary actual speed limitation. An alternative workaround would be to jumper the down counter pin 12 pad to pin 1 before clipping pin 12. In this case the commanded speed would be 2X higher rather than half speed, but it is also much more problematic since the microcontroller may not handle a 2X higher input frequency correctly, while I would expect it to handle a half speed input frequency since the latter falls well within its normal operating range.

If the above jumper between pin 12 and pin 11 does not work as expected - i.e. the voltage is not near 2.5 VDC on your multimeter and the commanded speed does nothing to vary the spindle speed - then the microcontroller pin input buffer is likely blown and you would need a new microcontroller. The alternative is that pin 12 of the down counter may connect to additional loads unrelated to the microcontroller and one of those is overloading the counter output.

the model is Hantek 6022BE

That is not a standalone unit, it needs a computer connection to operate. Hence it is much more of a bench solution rather than a field solution. You still seem to be assuming that whatever instrument you buy will only be used this one time rather than being of any use in the future. For an inexpensive one time use, your original single channel scope kit would work for examining the signal on pin 12 of this specific binary down counter. But it also represents a throwaway solution since that scope is also very limited in capability. It is adequate for single channel audio frequency applications but not a general purpose solution for higher frequency / voltage, or much of anything digital.

[TurboStep]

...snip off the legs of that chip and solder the new chip to those legs...

I usually snip off the legs and then remove them from the board one by one. Removing a complete IC in one go without the right tools is almost impossible without damaging the board.

Zetopan wrote "If you are sure that you want to replace the CD4024 [...] then..." well I'd just replace it. If you cut the legs you'll have to replace it anyway.
An alternative might be to pull out the microcontroller and plug it into a 28 pin socket. Cut off pin 11 from this socket and re-insert it into the board. That way you'd be able to check the output of the counter in isolation - just in case the microcontroller is defective. Use a socket with flat pins, not turned pins, to prevent stressing the socket on the board. As I mentioned earlier "As Pin 12 connects to the microcontroller it's theoretically possible that the input pin on the microcontroller is pulling pin 12 down. If that were the case then the additional impedance of the DVM is very unlikely to affect the signal." so I don't expect the microcontroller to be the case, but on the other hand... One last point is still bothering me. Assuming that the counter is defect. The counter has 7 outputs so why does this pin appear to have a problem. It could just as easily have happened to one of the other pins and you wouldn't even have noticed. Perhaps the microcontroller is not as innocent as it appears!

The IC you've chosen appears to be correct.

Step

[Eli_Ben]

That is a completely unwarranted assumption on your part. Once you learn how to use a highly portable oscilloscope it will become your instrument of choice.

Very true, you make a good point. I've no idea what other problems I'll need to deal with in the future. I'm really just being cheap and wanting a quick fix.

I did locate an abbreviated user manual for the DSO 094 if you want to examine it:
http://www.jyetech.com/Products/LcdS...Manual_094.pdf

Thank you, it does look easy to use. It's nice to see that it appears to be well supported too.

Here is a potential temporary workaround. If you are sure that you want to replace the CD4024 (which appears to be the culprit at this point) then ground yourself and carefully solder a short wire from the pin 12 pad (at the control board surface) to the pin 11 pad. Now carefully clip pin 12 at the DIP package so that pin 12 of the down counter no longer connects to the control board. At this point the output from the down counter first divisor stage (divide by 2) to the microcontroller input has been disconnected and replaced with the output from the second divisor stage (divide by 4). If the down counter pin 12 output driver was the problem then you should see about 2.5 V on your DMM on pin 11 when in AUTO and commanding a spindle speed with PathPilot. Likewise, using PathPilot to command a spindle speed should vary the actual spindle speed, but it will be slow by a factor of 2X (half speed). If you want 2000 RPM you would need to command 4000 RPM, etc.

This works! The spindle turns on when commanded to in PathPilot. The speed is definitely lower than commanded (can't check until the tachometer arrives). The voltage on the output pin 11 is 2.48V. So I'll order a new CD4024 counter today, which should arrive by Thursday.

The alternative is that pin 12 of the down counter may connect to additional loads unrelated to the microcontroller and one of those is overloading the counter output.

From examining the board, it appears that the only thing connected to pin 11 of the micro controller is pin 12 of the counter.

I usually snip off the legs and then remove them from the board one by one. Removing a complete IC in one go without the right tools is almost impossible without damaging the board.

Sounds like a good method, better than fiddling around with de-soldering wick.

An alternative might be to pull out the microcontroller and plug it into a 28 pin socket. ... to check the output of the counter in isolation

Since I already snipped pin 12 of the counter, I was able to measure it in isolation. I got 0V. Whereas when it was connected to the micro controller it was 0.46V.

One last point is still bothering me. Assuming that the counter is defect. The counter has 7 outputs so why does this pin appear to have a problem. It could just as easily have happened to one of the other pins and you wouldn't even have noticed. Perhaps the microcontroller is not as innocent as it appears!

Interesting point. I'd like to explore it further to hopefully reduce the chances of more problems in the future. Are there any other tests I can do to check for problems with the micro controller?

The IC you've chosen appears to be correct.

Thanks for confirming! Ordering now.

those can cause big problems too (power lines, grounds etc.), I had one last year, loosing one leg of the input power intermittently or partially took a bit of doing to find that one!

Can intermittent ground errors or other problems with the mains supply cause damage to the mill's electronics? Is it possible for me to check the mains supply for problems, perhaps with an oscilloscope or some other equipment? Or would an electrician be able to do this for me? Sorry for all the questions, I just want to do everything I can to reduce the risk of more downtime.

[Zetopan]

The easiest and most common method of replacing a soldered in DIP package is to clip all of the leads, then remove the body. Then you can reheat and remove each pin one at a time. Use some solder-wick or a solder sucker to remove the solder from the pad through hole. Here is a video showing the latter if you are unfamiliar with it:
Note that some solder suckers create static charges, which can damage static sensitive electronic devices, so make sure to get/use one that does not create static charges. Also note that if you can support the EC board vertically, the soldering iron can be used to heat the pad through hole on one side of the board while the solder sucker can be applied to the other side of the EC board. This can work very quickly but it requires good EC board support and good two handed coordination (I'm ambidextrous so that is a distinct advantage).

Can intermittent ground errors or other problems with the mains supply cause damage to the mill's electronics?

This is not usually a problem. An intermittent ground more commonly causes a lot of things to behave strangely rather than just one thing, and in any case your failing CMOS binary counter is highly unlikely to have been caused by an intermittent ground. When working correctly that part will operate off of a 5 to 15 VDC supply and in your case it was operating from a 5 VDC supply. Hence, I suggest that it was simply due to an infant failure of an output driver or the driver bond wire to the lead frame. See the common failure rate bathtub curve: https://en.wikipedia.org/wiki/Bathtub_curve

Something that you can do is to examine the grounding block inside the electrical control cabinet. In my now quite old late Series I PCNC 1100 this is located near the lower right, and you should see a lot of wires attached to it with machine screws. Make sure that each mechanical connection is tight and the wires are well attached to each lug. Other than that, replacing your binary counter should get to back to where you were before the failure.

[Eli_Ben]

Sorry for the long delay.

I've replaced the counter, which appears to have solved the problem! After several hours of use, the spindle is still performing normally. I stocked up on counters just in case there's another failure in the future.

I'm absolutely in awe of this community's ability and expertise to diagnose and reverse engineer. TurboStep and Zetopan, your help was invaluable. I can't thank you enough.

This episode has highlighted Tormach's incredibly disappointing approach to customer service though. Tormach refused to do anything other than tell me to replace the entire control board for $470 (the fix cost less than 1$). Everything on the board seems to be off the shelf apart from the micro controller (which Tormach refused to sell me individually). But I'm wondering if this board is actually proprietary, or if it's just Tormach branded. Considering their low cost approach I'd be surprised if it wasn't available under a different name or unbranded from China. If anyone has info on this, I'd love to hear it.

[TurboStep]

Tormach are unable to offer more support than telling me to replace the board. In their words, they don't have the 'technical aptitude' to offer further advice. Apparently the person who designed the board was their founder Greg Jackson who sadly passed away in 2015.

I was under the impression that this is indeed a Tormach design and judging by some of the details it was carefully designed by someone who did not want to cut corners. That said, the price price for a replacement is a little high. It would be nice to think they were in a position to offer a repair service, but that doesn't appear to be the case. Based on their statement I wouldn't expect them to be in a position to make any future additions/improvements - I hope I'm wrong!
Your good descriptions made the diagnosis fairly easy. Glad to hear you you got it up and running again.
Step

[lens42]

After reading this thread, it seems like there is an opportunity in a third party Tormach controller repair service. Though maybe not enough boards fail to justify the effort.


Sent from my iPhone using Tapatalk

[popspipes]

After reading this thread, it seems like there is an opportunity in a third party Tormach controller repair service. Though maybe not enough boards fail to justify the effort.


Sent from my iPhone using Tapatalk

There are board repair services available, I have never used one of them but imagine they are rather expensive.

I had a problem with the PDB controller board a few years back, it had standard components, as well as a programmable IC, the IC could be purchased and replaced but it needed to be programmed and this was beyond my capabilities so I went a different route to get the machine operable then bought a replacement board which I havent installed yet.

[Eli_Ben]

I was under the impression that this is indeed a Tormach design and judging by some of the details it was carefully designed by someone who did not want to cut corners. That said, the price price for a replacement is a little high. It would be nice to think they were in a position to offer a repair service, but that doesn't appear to be the case. Based on their statement I wouldn't expect them to be in a position to make any future additions/improvements - I hope I'm wrong!
Your good descriptions made the diagnosis fairly easy. Glad to hear you you got it up and running again.
Step

Tormach CS said that they only have one electronics engineer, and they didn't want him reverse engineering the board to help me, because it would take time away from his 'other projects'. So clearly zero priorty for them. They also tried to justify the high price by saying that a Mori Seiki board is about $2000. Which isn't a logical argument as it's actually a much lower percentage of the total price of the machine, compared to the 1100. Plus Mori Seiki are high end production quality machines designed for a totally different market!

But of course, with my machine being out of warranty Tormach aren't obliged to do anything at all. So I have to limit my outrage slightly.

After reading this thread, it seems like there is an opportunity in a third party Tormach controller repair service. Though maybe not enough boards fail to justify the effort.

It would be quite easy for Tormach to put together a list of third-party experts who can help diagnose issues. There's clearly considerable knowledge and expertise in their customer base, so it would be wonderful if they could tap into it. And it would be in keeping with their general DIY, bootstrapper approach.

I had a problem with the PDB controller board a few years back, it had standard components, as well as a programmable IC, the IC could be purchased and replaced but it needed to be programmed and this was beyond my capabilities so I went a different route to get the machine operable then bought a replacement board which I havent installed yet.

PDB?

I've used an Arduino to program Atmel ICs in the past. It's quite easy because there are some excellent guides available online. Took me a while to get it working just because I hadn't done anything like it before, but it wasn't difficult or expensive in the end. Not sure if other manufacturer's ICs are just as easy, but if you have a need to do this in future I would look into using an Arduino.

[Contract_Pilot]

I just use the right tools..

[CastorTroy]

Resurrecting an old thread here. It looks like I am having almost the exact same issue.

Details:



1. The voltage from J1-1 to J1-2 is .36 volts.
2. When I try and run the spindle the activity LED flashes so signal is making it to the board.
3. After grounding myself I took out the 28 pin microprocessor and re-seated it but it didn't make a difference.
4. Finally, I measure the counter chip (U17) and I only get a voltage read out from pin 14. Everything else is dead (see below). I measured with a multi meter but I can measure with a scope once my father in law returns from vacation
5. I have an additional problem in that the machine will not come out of reset for about 10 minutes after I turn it on. If I let it "warm up" it finally allows be to reset it. Everything else works as far as I can tell (i.e. I can jog the axis with the shuttle controller and I can turn the coolant pump on and off with the touch screen display).


6. With that said I am leaning towards replacing the counter chip like the fellow in the UK did. It seems to me the IC has failed even worse than Ben's did in that his chip only pin 12 failed. Mine has failed almost completely.


Thoughts? Any advice would be greatly appreciated!




Pin 1: 0V
Pin 2: 0V
Pin 3: 0V
Pin 4: 0V
Pin 5: 0V
Pin 6: 0V
Pin 7: GND
Pin 8: 0V
Pin 9: 0V
Pin 10: 0V
Pin 11: 0V
Pin 12: 0V Probing turns spindle on/off sometimes (see below)
Pin 13: 0V
Pin 14: 4.97V