Centroid Community CNC Support Forum

I was seeing the failure in less than 20 minutes. After rewiring, I went for 30 minutes with no failure. This doesn't constitute proof that it is fixed but is certainly a good sign.

Here are a few pictures of what I have now. They show:
* EStop run in two-conductor shielded cable with the shield connected only at the enclosure case (IN8)
* All home sensors are now in parallel with three-conductor shielded cable with the shield connected only at the enclosure case (IN1).
* ProbeTripped run in two-conductor shielded cable with the shield connected only at the enclosure case (IN7)
* On each motor driver, the PULSE + and - wires are twisted together. Same for the DIR+ and -

One deviation from the schematic is where the Pulse and Direction current flows. As shown in the schematic, I can trace the current path starting at Pulse +. The current flows through the opto inside the driver and exits Pulse -. From there, it flows through the DB-25 cable, through a saturated NPN, and into COM. From COM it flows into the negative terminal of the 5V supply. Out of this supply, it flows OUTSIDE of the DB-25 cable to the Pulse+ terminal. That wire from the 5V supply to the drivers is unshielded and unpaired to the rest of the signal path.

I'm sure this is fine if the drive is in the same box as the Acorn. But in my configuration, they are in separate boxes.

Instead, down at my driver box, I first connected each of the COM terminals of the DB-25 connector to a common point. The common point connects to the negative terminal of my floating 5V supply. The positive terminal connects to a common point that fans out to all of the + terminals. OK, it is a short daisy chain.

With this configuration, the current flowing into COM in the Acorn flows back on the many COM wires in the DB-25. In this way, the return path is paired with the signal and is inside the shield. Down at the drivers, the current flows out of the COM wires, into the floating 5V supply's negative terminal, out the positive terminal, and into the Pulse + and Dir+ terminals. By connecting to all of the COM wires, its inductance is minimized.

Even if the problem has only been reduced, I believe I'm on the right track now. Thanks to all that provided guidance.

Rick

Rick, the DB25 COM terminals are tied back to common via the Acorn board. No need for an external wire or terminal block.

This terminal block connects the four Pulse+ and four Dir+ wires to the positive side of my floating 5V supply. It also connects the negative side of my floating 5V supply to the 8 COM leads out of the DB-25 cable.

In the Acorn, all eight of the COM leads of the DB-25 are connected together. At the other end of the DB-25 cable, all 8 COMs are available. From a DC standpoint, I can use any one of them to complete the circuit. But from an AC standpoint, paralleling all eight wires cuts the inductance down by a factor of eight. So it is well worth the effort to run these eight wires to a single node.

If my explanation of the current path is confusing, I can draw it up.

Rick

Floating power supply? (Separate 5vdc power supply)
I would hardwire the COMs together from your 24vdc and 5vdc power supplies. In essence you are doing this by connecting the 5vdc com to the Acorn DB25 COM.

Just a note, Centroid now ships a dual voltage (5/24vdc) power supply with Acorn.

Marty,

Maybe these schematics will help explain my concern and remedy. You can see in the current design that the current path between motor driver and Acorn box is over two separate cables. The way I wired it, the path is entirely within the DB-25 so should have better noise immunity.

A floating 5V is no big deal - just a USB wall wart with the 5V plug cut off.

Rick

I can appreciate your schematics Rick

Simply connect a wire from 5VDC Common to 24VDC Common on your power supplies.

New power supplies shipped with Acorn are dual voltage and they are connected.

Marty

The key is to have all current flow inside the DB-25 cable. This is only possible if the 5 volt supply is at the motor driver end of the cable.

Maybe the confusion here is that if you look at the circuit from a DC standpoint, it doesn't matter where you put the 5V supply. It is fine to put it at the Acorn end and run a separate wire over to the motor drivers. The current does flow correctly.

But if you look at it from a high-frequency AC (inducted noise) standpoint, it is critical that the path from motor driver to Acorn be closely paired with the return path from Acorn to motor driver and that they are both inside the same shield.

This is identical to how you recommended I run my EStop connections - two conductors inside the same shield. I'm sure you would have an issue with me running one EStope wire inside the shield and having a separate wire run outside of the shield.

Rick

I ran error free for 1.5 hours so am now confident in saying that my noise problem has been solved.

Thanks to all!

Rick

RGSparber wrote: ↑Mon Dec 27, 2021 4:42 pm I ran error free for 1.5 hours so am now confident in saying that my noise problem has been solved.

Thanks to all!

Rick

Great to hear Rick.

You should be able to connect your VFD to Acorn next.
Marty

Not now. My spindle got too hot to touch after running for that 1.5 hours. Gotta deal with that next.

I did just get my Software Defined EEF to work with Acorn. I’m seeing repeatable accuracy of +0/-1 tenth.

Rick