Haas Brushed Servo AMP repair

[autotechsteve]

Well I have looked all over and I have not come up with any solid info on repairing the servo drive boards for the older Haas brush style servo motors. I have 2 bad drive cards and it kills me to look at them and know that all of the components on the boards are easily replaced. If I could figure out what went bad, I could have a good working servo amp again for a fraction of a new or rebuilt card would cost me. So I thought I would start a thread on testing and repair of the Haas Servo AMP card. Try to gather collective information in an effort to find a fix for all of us. I am competent enough to test and solder on new components but the theory behind how the cards actually work is where I am in the dark.

So far on both my boards I have not burned or discolored areas. No burned smell or any other indication of irreparable damage. Both cards set servo overload alarms but the axis themselves were not overloaded and moved freely. Long story short, new card, axis works.

My gut reaction was to point the finger at the MOSFET transistors but they "look" OK and are difficult to test. The original style of mount transistor does not exist any more. But new ones do in the same electrical configuration so if they were bad, then new mounts and heat sinks would need to be made. Part# IXFH50N20-ND. Best I can tell, the controls command movement, the amp commands the servo on and looks for the encoder to show movement. When the encoder does not show movement, the alarm sets.

I have also come across a mention of the opto-isolator going out on these boards, but I am not sure how to test these.

Here are a few pictures of the boards I have.



Steve

[haastec]

I believe the Mosfet is the most common component to go, but I to am weak at troubleshooting a lot of the individual IC components.

I believe there are several people on the forum with excellent skills for trouble shooting these or similar circuits so hopefully they will chime in before to long.

I am sure many people will be interested in what you find out here.

Good luck!

[underthetire]

The Mosfets should be quite easy to test. Should just have to remove the screws and pull them out. Most all those i've seen are in sockets, even though it may not look like it at first.

http://search.digikey.com/scripts/Dk...&keywords=ixfm

Check against the data sheet
http://ixdev.ixys.com/DataSheet/91522.pdf

[cygnus x-1]

Usually when MOSFETs go bad they short out. So to test them you can use an ohmmeter to check the resistance between the pins and also between each pin and the case (assuming the case is electrically connected). If you get zero resistance (dead short) between any two points the part is dead. Replacing any dead MOSFETs may bring it back to life, but then maybe not as there could be other components that went bad and killed the MOSFETs.

C|

[autotechsteve]

Most all those i've seen are in sockets, even though it may not look like it at first.

http://search.digikey.com/scripts/Dk...&keywords=ixfm

Check against the data sheet
http://ixdev.ixys.com/DataSheet/91522.pdf

I have not tried to remove the MOSFETs yet but that makes more sense on the mounting. That will make them a lot easier to test and replace.
The link to DigiKey is for the correct mount but the amperage and voltage are different. 24A vs 50A and 500V vs 200V. Are MOSFETs like capacitors where they do not have to be exactly the same to work? I tried to find a comparable MOSFET but no luck. I am probably not looking in the right place.
Here are the 2 data sheets so everyone can compare;

This is for the original part number;
Manufacturer: IXYS
Product Category: MOSFET
RoHS: RoHS Compliant Details
Transistor Polarity: N-Channel
Drain-Source Breakdown Voltage: 200 V
Gate-Source Breakdown Voltage: +/- 20 V
Continuous Drain Current: 50 A
Resistance Drain-Source RDS (on): 0.045 Ohms
Configuration: Single
Maximum Operating Temperature: + 150 C
Mounting Style: Through Hole
Package / Case: TO-204AE
Packaging: Tube
Fall Time: 16 ns
Forward Transconductance gFS (Max / Min): 32 s
Minimum Operating Temperature: - 55 C
Power Dissipation: 300 W
Rise Time: 15 ns
Factory Pack Quantity: 20
Typical Turn-Off Delay Time: 72 ns

This is the part number referenced above;
Family FETs - Single
Series HiPerFET™
FET Type MOSFET N-Channel, Metal Oxide
FET Feature Standard
Drain to Source Voltage (Vdss) 500V
Current - Continuous Drain (Id) @ 25° C 24A
Rds On (Max) @ Id, Vgs 230 mOhm @ 500mA, 10V
Vgs(th) (Max) @ Id 4V @ 4mA
Gate Charge (Qg) @ Vgs 160nC @ 10V
Input Capacitance (Ciss) @ Vds 4200pF @ 25V
Power - Max 300W
Mounting Type Chassis Mount
Package / Case TO-204AA, TO-3
Supplier Device Package TO-204AA
Packaging Tube

The Mosfets should be quite easy to test. Should just have to remove the screws and pull them out.

Usually when MOSFETs go bad they short out. So to test them you can use an ohmmeter to check the resistance between the pins and also between each pin and the case (assuming the case is electrically connected). If you get zero resistance (dead short) between any two points the part is dead. Replacing any dead MOSFETs may bring it back to life, but then maybe not as there could be other components that went bad and killed the MOSFETs.

C|

Thanks for the tips on checking the MOSFET. My meter is at work and I am off today so I will give it a shot tomorrow and post the results.
What could cause the MOSFETs to go bad? I know that shorted or bad cables and motor problems can kill the drive but is there any thing on the board itself that can cause it?

Thanks for the input guys, hopefully we can figure this out and save us Haas guys some dough!

Steve

[cygnus x-1]

I'm not an EE, but I have with some success replaced dead MOSFETs with not exactly equivalent parts. The trick is to find a similar part with equal or better specs. Probably the best thing to do in this case would be to contact Ixys and see if they can recommend a closest replacement.

Again, not an EE, but from my understanding MOSFETs are usually killed by overloads of some sort, too much current, voltage spikes, etc. If something failed in the gate drive circuit that could cause a failure. A machine crash is also a possible cause due to the sudden huge load on the motors.

You can use the ohmmeter trick to check all the other smaller transistors as well. A dead short between any two pins of a transistor indicates that either it's dead or there is a short somewhere else nearby.

Resistors should read their marked resistance or lower. If it's significantly higher something is wrong.

And of course check for the obvious, like open fuses, burned traces, etc. I notice there is a fuse missing in your picture.

C|

[Al_The_Man]

Being a DC drive this is going to be an H-Bridge configuration, If one side is on or firing and the other side turns on, it creates a dead short across the supply through two adjacent devices, sometimes this can occur if one side of the drive is on and a power supply is switch off for any reason, this can turn the opposite side on as the HV decays and blow the HEXFETS, one protection for this can be to fit high sweep rectifier fuses in the H bridge feed.
This used to be the norm in SCR drive packages to protect very expensive drive components.
Al.

[autotechsteve]

I'm not an EE, but I have with some success replaced dead MOSFETs with not exactly equivalent parts. The trick is to find a similar part with equal or better specs. Probably the best thing to do in this case would be to contact Ixys and see if they can recommend a closest replacement.

Again, not an EE, but from my understanding MOSFETs are usually killed by overloads of some sort, too much current, voltage spikes, etc. If something failed in the gate drive circuit that could cause a failure. A machine crash is also a possible cause due to the sudden huge load on the motors.

You can use the ohmmeter trick to check all the other smaller transistors as well. A dead short between any two pins of a transistor indicates that either it's dead or there is a short somewhere else nearby.

Resistors should read their marked resistance or lower. If it's significantly higher something is wrong.

And of course check for the obvious, like open fuses, burned traces, etc. I notice there is a fuse missing in your picture.

C|

Thanks for the tips. I started to check over everything and I tested the transistors but I came up with nothing. Then I saw a bad solder connection for the main 160VDC power feed to one of the boards. This is the board that would alarm instantly when I powered on the servos. I am assuming that the 160VDC was not even making it to the rest of the board and hence setting the alarm. I am going to have to de-solder it, clean it and re-solder it back on and test it. I tried to just re-solder it but it has oxidized too bad to take the solder. I will post the results when I have them.

Being a DC drive this is going to be an H-Bridge configuration, If one side is on or firing and the other side turns on, it creates a dead short across the supply through two adjacent devices, sometimes this can occur if one side of the drive is on and a power supply is switch off for any reason, this can turn the opposite side on as the HV decays and blow the HEXFETS, one protection for this can be to fit high sweep rectifier fuses in the H bridge feed.
This used to be the norm in SCR drive packages to protect very expensive drive components.
Al.

Thanks for the info on the drives. I am still trying to learn how they work. It makes diagnostics a lot easier if you know how the component operates. This is why I started this thread, to get a collective of information for the good of all. Thanks

Steve

[DaOne]

Get yourself some good liquid flux. It does wonders on improving the solder connection.