Gekko 320 & Z-Axis servo Madness need help :( ! 1.5 year battle

[Rich05]

My Z-Axis is not reliable.

I have sent Geckos back twice for repair, replaced the servo motor ($450), replaced the inverter, cleaned the optical position reader, replaced most all the wiring in the power box, removed any extra electronics like Cambell isolator boards, i have tried to tune the Gekkos in every concievable setting from limit 1/3-full, from damp and gain with large seperation, small seperation, both low, both mid, 7/8 position you name it.

Still the Z-Axis sometimes just stops, Mach 3 continues to show z moving but it is not. Some times the z servo just starts again, sometimes I have to turn the limit up and down, sometimes I have to press on the wires that go to the gecko, also resetting Mach, or turning off and on the power to the Drives will restart the z axis.

There is continuity in my wiring, this is really perplexing. Has anyone had one of thier servos just stop for no reason?

Should also mention it is much worse when the mill is cold, just turning it on in the morning.

At this point I am debating replacing the Gecko drive or and the servo again!! possibly these things may have been damaged yet again even though they are new. I hear it only takes a couple secconds to burn the servo motor up, while tuning the Gekko I had some problems where the z-axis crashed into the part.

Today. I had the Z servo about 17 minutes into an operation retract for no reason, I had to competly re-tune the Gecko all over again. It suddenly was faulting all over the place.

I ran the program again, it retracted again but then dove into the part.

Each new day I have re-tune the Z-Servo Gecko to get it right again. the Gecko seems to lose its settings!

The PC I am using is a pentium 3.4 with plenty of ram, nice robust Intel 875PBZ motherboard, SIGII pci parrell card, and no other programs running. Purpose built for this job. Also the parrell cable stretches to the power box about 15'.

what gives?

[MattTheNoob]

Ok, total CNC noob here, but I can usually finger my way through a debug...

It sounds like the Z axis is the only one that experiences this phenomemon. Do you have the same servo/motor on any of the other axes?
If you swap components, does the problem follow the motor, drive, cabling, or any other component, or does it stay on the Z?
Is there a particular sequence or condition that you know will make it fail?

Intermittant problems stink. I like things that stay broke or stay fixed.

[Adamj12b]

You say that you have replaced the motor and the gecko before, but have you tried replacing the encoder? That could be your problem. Maybe you could post the specs about the motor, encoder and power supply? -Adam

[Torchhead]

Servo motors typically don't just move on their own (unless the drive thinks the motor is in the wrong place). Most common causes are defective encoder or noise getting into the encoder lines causing the drive to see pulses that don't exist. The same issue could come from noise on a step line. Cabling should NOT be sensitive to movement or temperature. Do you have a nearby VFD on the Z?

Do you have adequate cooling on the Gecko's? The G300 series is not thermally protected and as heat builds up it can case the drive to do wookie things! If you are getting drive failures (especially the same drive) you need to look at external conditions. Shorted wires can destroy a drive. Open wires or intermittent connections can be distructors. Make sure the case of the motor is not shorted to the motor winding. Check with it disconnected from a motor lead to the case with an ohmmeter.

If you get drive faults they can be caused by several things but a faulted drive will reset and start running again if you don't have some form of central fault shutdown. Just tying the Reset/Fault pin to the ENC+ will cause a reset but by then it's already lost position.

TOM CAUDLE
www.CandCNC.com

[Doug Schwochert]

Hi

not being famillar with a SIGII pci parrallel card. is this your parrallel port interface card between the computer and your Geckos. or is this a driver in your computer. can you re configure your ports and pins and swap X to Z then switch the wiring from Xto Z and see if the problem also switches from X to Z or stays on X sounds like your parrallel port interface card may be going bad. I had a similar problem.
\

Doug S

[Mariss Freimanis]

You cannot use just any breakout board with the G320. The breakout board MUST have 74ACT logic series buffer ICs on it. Look to see if the part number on the ICs on your board contains that 5-character sequence printed on it. There may be other characters before and after '74ACT' sequence, example: SN74ACT240. Not good is 74LS, 74HC, 74HCT or other series will not work.

Tom Caudle (message #4) makes an excellent breakout board for use with our drives.

Mariss

[mikrokontrola]

Hi,
May I ask why the 74ACT logic is the only acceptable serie?
gregory

[Torchhead]

Actually I must disagree with my learned friend. The specs for a 74HCT541 are 50ma source and sink to the optos on the G320. Also since the LS series will sink 23ma per output on that chip it can be used on any logic that sinks current (like the 201 and 320 series.) The specs for normal logic chips (not bus drivers like the 540 and 540 series we use exclusively for the interface) have lower drive capability and do indeed need to be ACT series. So it depends WHICH chip you use in a series. Some HC, HCT and LS chips are designed to be drivers and work fine for driving opto loads.

TOM CAUDLE
www.CandCNC.com

[Mariss Freimanis]

deleted. See #10.

mariss

[Mariss Freimanis]

And I must regrettably disagree with my esteemed colleague.:-)

1) From the TI datasheet, +/-50mA may be listed but it is the absolute maximum Icc current. Absolute maximum sink or source current per output pin is listed at +/-25mA.

More important is the sink/source current while maintaining TTL voltage spec. This is a far more modest +/-6mA at 0.33V and 3.84V at a 4.5V Vdd. Using Ohm's Law, this gives a 55 Ohm sink resistance and a 110 Ohm source resistance. Scaling these impedances to 16mA required by the drives gives 0.88V for a logic '0' and 2.74V for a logic '1'. This is a less than sterling performance.

Now let's look at a TI 74ACT541 datasheet and compare. A 24mA sink current gives 0.44V on the output pin while a 24mA source current gives 3.8V with a 4.5V Vdd. Scaling gives 0.3V at 16mA sinking and 4V when sourcing 16mA. This is acceptable performance.

The 74ACT part will reliably drive a G201 or G320, the 74HCT part will not. FYI: I never use 'typical' 25C specs when I design; I always use the guaranteed minimum specs at temperature (-40C to +85C in this case). 'Typical' means "most parts will be this this good'" so I worry about the remainder that won't be.:-)

2) Symmetrical sink/source current drive capability is optimum for driving cables. Cables are primarily capacitive loads at step pulse frequencies and it's current that charges and discharges this capacitance. It does no good to have a 74LS part that can sink 23mA when it can't source more than 400uA. The 'weak sister' is the source current which will set a pathetically low maximum frequency because of its miserable ability to 'charge' the cable capacitance.

Use 74ACT parts. They are as cheap as 74HCT or 74LS parts so why not?

For your edification and amusement, please see the attached .gif picture. It shows an HEDS-9100 encoder with 74ACT14 driving a 100' shielded cable containing four 22-gage wires. The signal shown is at the G320 end which is running at 100K step pulses per second. The signal is clean as a whistle when it comes out of the cable.

Mariss

[mikrokontrola]

Now I understand, thank you for an explanation.

[BernardR]

Mariss can you please confirm that you are talking about the step and dir signals here and the setup up is say a standard six foot parallel cable into a BOB and the ACT74xxx driving 12 to 18 inches of wire to the G320s?

Are you also making any recommendations about encoder signal conditioning?

Thanks

BernardR