A couple of years ago, I sold my Grizzly G0463 milling machine and upgraded to the larger Grizzly G0720R (SIEG SX4). To this point, various SIEG machines I've owned have been pretty reliable but my luck finally ran out a day before Christmas. In the middle of power-taping a hole at low speed, the spindle abruptly stopped with a loud clunk and would not restart.
While troubleshooting the mill, Grizzly support was not immensely helpful, narrowing it down to "there is something wrong with your motor or the controller". The information on the internet was pretty sparse as well. Nevertheless, it appears that this sort of failure is not uncommon on the SIEG SX4 and other SIEG machines that use brushless DC motors. Rather than dropping over $1000 on a new motor and controller, as suggested by Grizzly, I decided to attempt repairing the board. At the end the fix was relatively straightforward and set me back around $60 in total. Hoping that this will be helpful for others in this situation, in this post I will go over my troubleshooting process and provide some tips for the repair.
Disclaimers and Warnings
If you decide to follow this article, you will have ample opportunities to seriously hurt yourself or even die. For that reason, I can't be held liable for anything you do, so proceed at your own risk. That said, the following is especially important to keep in mind:
- The board is connected to mains power. Depending on where you live, this can be 110V or 220V, either of which can kill you. Whenever you access the electrical parts of the machine, make sure to unplug it from the power.
- There are pretty large capacitor bank on the motor controller board that can hold charge for hours or even days after the board is powered off. If you accidentally touch the terminals, in the best case scenario it will hurt a lot. In the worst case it can stop your heart. Before doing anything to the board, look up how to safely discharge the capacitors and don't forget to do it after each time the board is powered up.
- Finally, the controller board and the motor combined cost over $1000 from Grizzly in the USA, excluding shipping. If you do something careless you can damage a working part, potentially making the repair bill even larger.
I hope you get the point: what I'm about to describe is dangerous. It can seriously hurt you, kill you or, at a minimum, cost you some serious additional money. If you don't have any prior experience with electronics, this is not the project to learn on.
Initial Troubleshooting
As mentioned above, my mill failed while tapping a hole at a relatively slow speed. After power-cycling the machine, the spindle would not run and the LCD would briefly flash "ERR". On one or two occasions, after I turned the spindle by hand a bit and tried to start the mill, the spindle jerked a few degrees and stopped again.
When turning the spindle by hand with the machine powered off, I felt intermittent resistance as if the spindle had detents it was jumping over. I quickly eliminated the possibility of mechanical failure in the spindle by removing the drive belt and turning the spindle and the motor by hand. Normally, the motor should have a bit of resistance due to a small amount of residual motor braking when the power is off. It should still spin smoothly without any noticeable bumps. In my case the resistance was pretty lumpy, with a noticeable increase for about 1/3 of each revolution.
After eliminating the possibility of loose connection, tripped circuit breaker, or blown fuses, I determined that the problem had to be with the motor or the main controller board.
Jumping ahead, the problem with my mill ended up being a blown output transistor on the motor controller board. Since there are a few different failure modes, let's look at how I arrived at this conclusion. If you intend to play along at home, you will need only a few basic tools at this point:
- #2 Phillips head screwdriver
- A multimeter that can measure DC voltage and resitance
- Optionally, a pair of needlenose pliers or sturdy tweezers (to remove and insert wires into terminals)
BLDC Motor Control Theory
Before going further, it is helpful to get a bit of theory out of the way. As mentioned earlier, my Grizzly mill uses a brushless DC motor (BLDC in short). Fundamentally BLDC motor is not that different from a stepper motor. Stepper motors usually are driven by two windings (or, more correctly, phases) that are 90 degrees apart, and BLDC motors have three phases 120 degrees apart. As such, both require controllers that apply current to the appropriate winding for a precise amount of time. The timing is handled a few different ways depending on the controller implementation. In this particular case, the motor has three Hall effect sensors positioned at 120 degrees to each other; the controller can detect the position of the rotor based on the sensor state and send the current to the appropriate winding.
Possible Failure Modes
Between the motor and the controller there are a few different possible failure modes. First of all, the motor winding can simply burn out or short internally. Second, one of the hall effect sensors could fail, thus breaking the feedback loop. Third, there could be a problem with the power output stage of the controller. Finally, there could be a failure on low voltage digital side. The last failure will be very difficult to troubleshoot without proper schematic, but the first three are relatively easy to test with pretty basic tools. Let's look at each of them individually.
Don't Forget The Basic Stuff
I can't count how many times, after having spent hours going through complicated troubleshooting steps with fancy test equipment, I eventually found that a wire was loose, a switch was faulty, a fuse was not seated properly, or some other silly problem was present.
To avoid wasting time chasing "red herrings", find (or download) a copy of user manual for the machine and study the "Wiring Overview", "Wiring Diagram", and "Electrical Components" sections. Then check the fuse on the main board, main circuit breaker, terminal connections and e-Stop switch function. Of course do this with the machine powered down to avoid electric shock.
Internal Motor Failure
As you probably know, motors can burn out and fail. A winding in the motor can simply break, either from mechanical stress or due to severe overheating. Alternatively, the insulation can fail creating a short and effectively reducing the inductance (and thus impedance) of a winding. Generally, the second failure mode leads to overheating and ultimately broken motor winding. Fortunately, this is really easy to test on in brushless DC motor using simple multimeter:
- Unplug the machine from the mains power
- Loosen screw terminals W, V, and U (on the right side of the board)
- Make sure the wires are not touching each other or any other metal
- Check that the motor is spinning freely
- Using a multimeter, check the resistance across each winding by check every combination of two wires (on my mill they are Red, Brown and Blue). The resistance should be on the order of 1 Ohm
- Carefully shorten any two of the three motor leads and gently try to slowly spin the motor by hand. You should feel moderate resistance.
- Repeat for the other two combinations of leads
If all of the above steps check out, your motor windings are good; otherwise, you motor is likely at fault and will need to be replaced.
Motor Encoder Failure
Motor position encoder is an integral part of the circuit a failure in the encoder will render the motor controller inoperable. Fortunately, this particular machine uses three Hall effect sensors that are very easy to test. Since the encoder will need to be tested under power, this operation is inherently more dangerous, so be extremely careful not to touch any exposed wires or traces.
The process is as follows:
- Leave the motor windings disconnected from the board
- Plug in the machine and turn on the main power
- Locate a large green 8-pin connector in the lower-right corner of the board
- Using a voltmeter check the voltage between +5V and Ground terminals (Red and Black wires respectively). You should see 5V or so(+/- a few tenths)
- Check voltage between Ground terminal (Black wire) and SA, SB, and SC terminals (Yellow, Orange and Brown wires respectively). One of them should be at 5V and others should be close to 0V.
- Leaving the voltmeter connected to the terminal that was at 5V slowly rotate the spindle. At some point the
voltage should drop to 0V.
If for some reason you have less than three hands, it's a good idea to get a helper for this step. - Find the next terminal that is at 5V and repeat steps 5 and 6.
- Repeat steps 5 and 6 for the third terminal.
If all of the above steps are successful, the encoder is in good shape and you can proceed to the next steps. On the other hand, if you don't get 5V in step 4, the 5V power supply rail on the board is faulty and will need to be repaired, or you will need a new board.
If one of the terminals doesn't reach 5V or doesn't drop to 0V, there are two possible causes: either the Hall effect sensor is faulty, or there is a problem with the board. To isolate the cause, disconnect the SA, SB, and SC leads and check the voltage between Ground terminal and each of the three terminals - they should be at 0V. If that checks out, repest steps 6-8 (checking voltage between Ground terminal and each of the encoder wires). This will tell you if the problem is with the board or the encoder. In the former case inspect the three ICs (optoisolators, presumably), capacitors and resistors immediately next to the terminal block. In the latter case one (or more) of the Hall effect sensorts is not working.
Digital Logic Failure
A failure on the logic side of the board will be the most difficult to troubleshoot and repair, since schematics for the controller are not available and its inner workings are unknown. Fortunately, digital side of the controller is unlikely to completely fail out of the blue, and in rare cases that it does, there is usually some visual signs. This can be charred parts, puffed capacitors, cracked ceramic chips, etc. Careful inspection under good light and magnification can usually provide enough clues.
Furthermore, failures in the digital logic would not exhibit any symptoms while the machine is powered off, so in my case, it was highly unlikely that the machine failed due to issues with the control logic.
Output Transistor Failure
Most motor controllers, be it BLDC speed controller, 3-phase VFD or a stepper driver, use some combination of bipolar junction transistors (BJT in short) or some type of field effect transistors (FET in short) in the output stage to control motor windings. These transistors get pushed pretty hard by high current, constant heat cycles and motor back EMF. Incidentally, output transistors are usually the most expensive part of the motor control circuit, and the manufacturers often cut corners to save cost. As a result, output transistor failures account for an overwhelming majority of the burned motor controllers and variable frequency drives. It's not always possible to test the transistors in-circuit, but often a difference in resistance between pairs of U, V, and W terminals might offer a clue:
- Unplug the machine from the mains power
- Unplug motor terminals from the board (terminals U, V, and W on the right side of the board)
- Measure the resistance between all three combinations of the terminals (U-to-V, V-to-W and W-to-U).
- With the board powered off, you should see a very large value, on the order of tens of Kiloohms or even Megaohms. If you see shorts or values that are not within a few percentage points, it's a sure sign that you are dealing with a transistor that failed in the "On" state.
Unfortunately, consistent resistance does not necessarily mean that all is good. A transistor can fail in the "Off" state. Since the other transistors are off as well (the board is powered off, after all), the resistance test will not show it. To really test the transistors you will need to desolder them from the board and use an inexpensive transistor tester to test each one individually. Since this post is getting pretty log, I will cover desoldering and testing in the next post that is coming in a few days.
Troubleshooting Results
I got somewhat lucky: simple resistance test described above quickly showed that I had a burned transistor. In my case the resistance reading between U and V, and W and V showed a few ohms of resistance, while U-to-W resistance was around 60 Megaohms.
At this point it was clear that I will have to either order a new controller board or attempt to repair this one. With not much to loose I took the board off the machine and proceeded to take it apart to find what type of transistors are used.
Since this post is getting to be pretty long, I will cover the repair process in the next post that will be up in a few days. Jumping a bit ahead, I ended up ordering six replacement ON Semiconductor FGL60N100BNTD transistors and three
Infineon IR2101STRPBF gate driver ICs. Even though the transistor appears to be a legacy part that is being (or has
been) discontinued, there are quite a few listings for it on eBay and DigiKey still had a few hundred in stock.
Since I wanted to avoid getting a counterfeit part from China, I opted for DigiKey. The total bill for six
transistors, tariff and Ground shipping ended up being $50.44. The ICs I ordered from Mouser (with a larger order to
parts for the next TouchDRO batch), which set me back another $4.86.
Replacing the transistors ended up a bit fiddly, since they are mounted to the heatsink on the underside of the board. All in all it took a bit over an hour but the mill is back in service and I'm back to turning perfectly good metal stock into scrap useful parts.
There are so many things wrong with this article that I don't know where to begin. Why would you want to invest time and money into fixing a disposable china machine. For what you paid for it you could've gotten a good used Jhead Bridgeport. It wouldn't have the the gimmicky variable speed control that you don't need anyway. I see that you got yourself an expensive Fluke 87 and learned how to check resistance and voltage. Now you think you are an electrical engineer. Like they say, if you only know how to use a hammer, every problem looks like a nail. You don't test BNDT using an Ohmmeter. Unless you can trance the signal in-circuit, you need to remove the part and test on a transistor tester. To test the encoder you need to hook it up to the scope and spin the motor at operating RPM and not measure the voltage. Gosh! you don't even know the difference between a stepper motor and BLDC motor. I laughed so hard when you suggested to short the leads on the motor together and spin it. That is the dumbest thing I've read in a long time. You are a completely joke. Do you even know how to read a schematic? Guess not or you would not be aimlessly poking around the board with a voltmeter. Here is an idea for you. Sell the fancy Fluke multimeter to someone who actually knows how to use it and spend that money to buy yourself a clue.
ReplyDeleteAnonymous doesn't recognise the intended audience. And is a dick.
Deletestupid troll has no Idea but to criticize. Keep up the good work Yuri
DeleteYuri, please ignore this fool. You seem to be the only person out there who has actually taken the time to explain anything about these Chinese electrical systems and I thank you for that. The original manufacturers didn't even bother! Perhaps "Anonymous" could point their amazing skills at the issue and actually help instead of being so utterly rude?
DeleteI'm currently rebuilding and repairing a drilling machine that contains the same driver board and your photo's alone have helped me beyond anything else. I salute you Sir.
Thanks Yuri, don't feed the troll..
ReplyDeleteThanks Yuri, I really enjoyed and learned a lot from this article, and I've been a control systems engineer for almost 40 years now.
ReplyDeleteInterestingly, this weekend I also watched AvE's youtube video on replacing his mill motor, and also (especially!) BlondiHacks youtube video on troubleshooting and fixing her budget mill, which turned out (in her case) to be the motor. But in her case a series of interesting travails followed, centered around the issues she had by following customer service's advice and swapping out the control board... then the display... then the speed pot. Ugh!
I appreciated the steps and issues that all of you went through, and it will help my students seeing your thought processes and steps.
Thanks for being open and sharing!!!
Thank you for the feedback. Glad you found this useful.
DeleteSo far the board seems to be working. I was able to tweak the encoder timing a bit by simply turning the housing a few degrees, which made the rotation more even (it was pretty "wobbly" before). I will be replacing the motor with a 3-phase inverter duty motor pretty soon. At this point I'm not too impressed with the BLDC controller (the motor is pretty good, but the controller seems to be put together by a hobbyist, not an EEE who knows how to do these things), plus I want to get wider RPM range. 1600RPM is pretty limiting for me since I use small end mills.
Not to take away anything from Yuriy, but to add for completeness, if anyone else has a mill or lathe go out and is investigating and repairing it, here are links to the additional videos I referred to above:
DeleteQuinn's the quintessential "amateur" machinist, in the true sense of the word. You can tell she loves it!
BlondiHacks "Mill Explosion and Repair" - Motor failure and replacement, with additional travails at https://www.youtube.com/watch?v=QsOt80ShiGA
This one is great if you need to deal with a used motor replacement, or just to understand bigger AC motors. AvE's language is delightfully appalling! But his work is absolutely spot-on!
AvE "MAZAK MILL AUDIT - 9 wire dual voltage electric motor wiring" at https://www.youtube.com/watch?v=YnOlZpyAhDc
DWS, I posted this stuff to help other people that run into this. You adding more info doesn't take anything away. This saved me the work of finding the videos.
DeleteRegards
Yuriy
Great work Yuriy.
ReplyDeleteThanks for sharing all this.
-brino (from HM)
Yuriy,
ReplyDeleteI’m confused by a discrepancy between the wiring diagram and the actual board. The wiring diagram shows the L1 terminal being at the top which connects to the main terminal at the top of the circuit breaker. The N1 terminal is the 2nd one down. But on my board (and even the photo in the instruction manual) the N1 terminal is at the top and the L1 is the second one down. Which one is supposed to connect to the top main terminal on the circuit breaker?
Yuriy,
ReplyDeleteOne thing that confuses me about Grizzly’s wiring schematic is that they have the L1 terminal at the top and the N1 terminal as the second one in. The L1 terminal then connects to the top of the main circuit breaker terminal. But on the my controller board (and even in the photo in their manual) it shows the N1 terminal being at the end and the L1 terminal as the second one. I’m not sure which terminal connects to the top of the main circuit breaker terminal.
Thanks,
Marshall