• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

CNC motion control servo amplifier help? Missing a phase!

Not open for further replies.


I am helping trouble shoot a machine that has not run in years. It has an AC servo system with resolvers, not encoders. We suspected there was an issue with an axis due to an alarm when trying to bring the machine online. We invoked a diagnostic test called a phase test in which each phase of the servo is sequentially powered and held at position for a few seconds. It basically jolts the ballscrew with a small bit of motion, then holds it firm. You can feel and hear it. One phase of that test is completely MIA, and you can hear the servo motor completely let go.

The cables were all swapped with another axis to troubleshoot if this was a motor/resolver issue, or the amp/feedback card.

We found that the problems follow the electronics but no one isolated the amp from the axis card. The card functions as a ADC converter from the resolver and sends that data to the control unit as well as takes digital motion command from the control and relays that to the amplifier.

I have the amp opened up now and there is no simple indication of failure. Usually when they fail, they let the smoke out, and the IGBT pack is blown. In this case, it seems to test good. As well, I simply cannot get my head around how only ONE phase of the test can be missing? If you drop out one phase of a 3 phase system, you should lose two of the phases!! I think that detail may be key in understanding what is going on here.

The system basically works with a logic control board inside the amp that further sends command to the 3 phase IGBT module. There are also current monitors in the board and one of them has corroded traces. I am not sure if the feedback from this device would cause complete shutdown of a phase or not but it seems to at least have continuity in the board for now.


Well-Known Member
Most Helpful Member
Is this a commercial machine or a home built? if the former, what is the make of the system, who's servo's and drives are they? Fanuc? Siemens?


This is a commercial servo amp by kollmorgen. I have now triple verified this amp IS the culprit and I have combed the circuit for hours and cannot find an issue with it, at least with my skill level. As I recall, IGBTs like to fail catastrophically and take out the gate drivers. Because this is 3 phase output and part of the circuit is good, I am simply testing and comparing meter readings with all legs. They all test the same and the gate drivers appear to me good.

The control system has a "phase test" in which it will energize each phase separately. You can both hear and feel the servo jump to position in two phases, but one phase, the servo completely shuts off like it has no power at all. Nothing is blowing fuses or pulling excess current.

There is a "drive card" that is separate from the amp that is responsible for DAC conversion and sending signal to the amp. That card was swapped with another and even tested the servo with another identical amp next to it. The servo functions perfectly.

I don't have any documentation or diagram so it is a difficult repair.

The outputs of the IGBT in all phases test in the megaohms, and the gate drive resistors and optos seem to check out. I really cannot suspect anything on the DC supply side because the amp is fed with DC from a common PSU and other phase seem to work fine.

There are two current feedback loops on the amp board on two of the phases that work to give feedback to the control. I really don't think those are of issue though as I believe these are only designed to provide current data to the control so the control can stop motion in the event of the current limit being reached.

Because there is a common DC PSU that provides power to all amps, there has to be some type of of fuse in the circuit but there are no simple buss fuses or anything and I would think they would just have a common fuse on the DC buss side, disconnecting the whole amp from power.

Not open for further replies.

Latest threads

EE World Online Articles