PDMX 126 with AM822 drivers and relays.

Hi guys,

I am upgrading the electronics on my G540 driven CNC/Router, and need a little assistance with interface between the AM822 error signal, and the PMDX126 error/e-stop connectors.

On the PMDX 126 BoB, it says use connector J13 as below:
1 GND
2 'K' ( input or machine hold )
3 GND
4 E-Stop ( disables the outputs and relays on the PMDX board )
5 GND
6 Fault ( Fault Input ( Active Low) )
7 ERR- ( External ( Gecko, etc ) drive error input negative
8 ERR+( External ( Gecko, etc ) drive error input positive
9 COM ( Common Terminal for Restart Output )
10 RST ( restart output)

Now the Alarm output of the AM822 is:
Alarm signal: OC output signal, activated when over-voltage, over current,
under-voltage, phase-error and stall-error protection. Output polarity can be
programmable. 30V@100mA

So - I had a look at AM822 provided picture for Open Collector and PNP signal, and it makes no sense to me, they have pulse +, Dir +, and ENA + coming from BoB, then they daisy together Pul-, Dir-, ENA-, to the ALM-, going to a controller ground, with resistors fitted where VCC = 12 or 24, but no resistors if VCC = 5 Volts. Given I am not planning on using the ENA+ signal, I would just skip pluggin in the ENA on both accounts, but Do I then Run the ALM -, and daisy chain them together for the 5 different drivers into one line, and then go to the ERR-, or fault, terminal? Is this Common Cathode connection - anything to do with Active high signal, or is that setup seperately to having a common cathode or anode?

The confusing thing is that the AM822 page has the ALM+ signal, going to the ALARM on the Controller, but also the ALM- going to a ground on the controller. SO should all the ALM+ be in series or paralell and go to the ERR+, with the ALM- link together to say pin 5 GND?

Should the PDMX 126 be setup as ( PNP SIGNAL) or Anode? for the motor connectors, as it would seem either can be done?

Confused as hell!
Nick

We'll start with the basics.
Open collector outputs essentially work the same as a switch. As it's a switch done electronically, the output is polarity sensitive, hence the + and -.
The PNP example given for such output, simply means you are switching the high/positive side, as you would if using a PNP transistor to do the switching.
I'm guessing there is a NPN example as well, which simply means you are switching the low/0V side, as you would if using a NPN transistor.

In terms of drive inputs (Pulse, Dir, and Enable), think of each input + & -, as having a LED/bulb/anything that needs power to light up attached (this is what an opto is!). For it to work, you need to complete that circuit, and makes sure power is getting into that circuit.
As you have access to both the + & - independently for each input, it gives lots of flexibility compared to something such as a Gecko drive, which internally connect all + or - together (whether it's + or - varies between drives). What you have to do, is ensure each input + gets supplied with a suitable voltage, and each input - gets connected to 0V.
Ignoring the fact these inputs can be driven with a full differential signal, that leaves three options. NPN, PNP and OC.
NPN - connect the + to the voltage source, and - to the controller NPN output.
PNP - connect the - to 0V, and the + to a controller PNP output (would typically be called a TTL output, if the controller is providing the voltage).
OC - there are two options for this. Connect voltage source to the controller OC output +, then the controller OC - output to the drive input +, then the drive input - to 0V. This method switches the positive voltage going to the drive, and the alternative is you swap the controller output, and drive input around in the circuit, so you are now switching the negative side of the drive input.

Now in terms of the outputs (aka the Alarm), the same principles apply, just the controller and drive are swapped around in the above examples.
However, you need to decide if you're going to use active on alarm, or inactive on alarm. Using active on alarm, you need to wire all drive alarm signals in parallel, so any single drive faulting will connect the + and -, completing the alarm circuit.
If you use inactive on alarm, everything has to be wired in series, so any single drive faulting will break the alarm circuit.

Thanks, but it's still not clear in my mind regarding physically hooking up the ALARM+ and _. Just looking at the ALM+/_ as a single issue.

On the PDMX126 there is a Fault-Active LOW pin amongst the e-stop pin and a few GND's. So I will setup all the AM882's to go fault Active low in the AM882 software. Then I connect all the ALM+, ALM- , ALM +ALM- together in series, from the last ALM- goto the GND pin on controller that's next to fault pin, and from Fault pin goto ALM+ for the first drive. Where is the signal coming from because the Fault pin is an INPUT active low? Does the AM882 make it's own fault signal, or does it just ground a signal that needs to already be there?

If I look at the AM882 schematic in both cases they have the PUL+signal connected to the ALM+ signal. Is this the source for the voltage? what happens between the pulses ie - can't the pulse naturally drop to zero which would trigger the ALM active low, just because the pulse is low?

If I have to connect a PUL+ to the ALM+, do I only do that before the first driver, as it is now effectively powering the series chain of all the ALM+/_'s. Can't this create an issue if Pul+ gets grounded to another driver when the other driver faults?

Then also - If I am daisy chaining the PUL-/DIR- to GND, should that be done seperately for each drive, or can I daisy chain all of them together also and then plu into any GND on the PDMX 126?
The AM882 schematic has all the PIL'DIR/ENA/ALM -, all connected together, then all going to earth - but on the PDMX I have earths in the J13 ( e-stop area ) and I also have individual PC GNDfor each of the driver setp/dir signals, so I figured that the DIR/SteP- should go together to their respective VCC GND, and the ALM- with be seriesed and then run to a GND in the Fault area ( J13 )

These are the bits I need explained. as far as using the fault pin, I could use the ERR+, and ERR-, but really have no idea on what the difference between that is, and using the Fualt pin.

Nick

I've just looked at the PMDX-126 manual.
The fault input should be left floating (aka unconnected) so everything gets enabled. To trigger the fault input/disable outputs, it has to be connected to 0V/Gnd.
For that to work with your drives, they need to be left in the default alarm output (active on alarm i.e. the + & - get connected on alarm), and then wired in parallel to the Fault input (all ALM+ to Fault input, and all ALM- to 0V/Gnd).

Regarding daisy chaining the Pul/Dir, I would personally do that at the BOB end, and use twisted pair/shielded wiring to minimise the risk of interference. The Enable signal is not as sensitive to noise, so I wouldn't worry about simply daisy chaining the Enable wiring across all the drives, if that was easier for wiring.

From what I understand from the manual, the Fault input once pulled low, disables everything, including the E-stop signal to the PC. The Error input doesn't disable anything, but triggers the E-stop input to the PC for 1 second, so is reliant on the PC disabling things.
The benefit of using the Error input, is if the drives allow re-enabling after a fault without needing their power supply cycled (i.e. turned off until the drives fully power down, and turned back on again), then resetting the E-stop via Mach and re-activating the drive enable signals should mean everything powers up again without the need for a power cycle. I know EM drives have the option to re-enable simply by disabling/enabling the enable input without a power cycle, but it has to be set via software.
The disadvantage being it is reliant on the PC handling the E-Srop.

Ok - So I'm going to connect all ALM+ in Parallel to the Fault pin, and then all ALM- in Parallel to the GND pin next to Fault.

By my reading of the manual - the ENA + and - , is in the Enabled state by default as long as it is disconnected. ( Ie - I'm going to leave these pins not connected to anything = drive enabled )

For the Fault signal, my understanding is that Fault signal functions the same as the E-stop signal once it is active - as you mentioned. In that PDMX will disable all of it's own outputs, ( servo drivers, Relays, as well as signal to PC on PIN10 that the e-stop has been tripped). That then halts the operation on the PC and places my UCCNC software in reset active mode - which turns off the watchdog timer, which in turn backs up the PDMX in getting it to turn off outputs in the case the PDMX internal E-stop circuit fails to turn outputs off. ( In this state the AM882 drive signals should be unpowered - including the ALM+ and _ , and the PDMX has returned to Off due to watchdog timer not active? )

I then click on the software reset, which should effectively turn on the PDMX when watchdog becomes active again. Then I Home the axis to get rid of missed steps, turn on the spindle / coolant, and Run from a few lines before the fault freeze. - That's my plan. I was wanting to run the spindle power through the PDMX larger relay so that the spindle turns off, but I don't know of that's the right thing to do as it will cycle power to the inverter when the spindle is a full speed. - Maybe there is a better way like an e-stop that I can wire into the VFD. I did want to use the smaller PDMX relay to drive my spindle water pump, and also the dust extractor so that if a fault is detected, effectively everything turns off.

On a side note - is there any benefit to setting up the AM882 to run using a common PC+5V vs, the other way with the common PC-gnd? On the PDMX manual it has a pic with typical connection for leadshine, showing the Common +5V PC as the setup.Which one should I choose?

I'm assuming disabling the outputs will also cut power to the stepper drives?
If it doesn't, the drives will remain in fault until the power is disabled.

VFD wise, cutting power is fine, provided the power is cut cleanly, however the spindle could still continue to run for a few seconds afterwards depending on how quickly the internal capacitors discharge to the point the VFD faults out due to under voltage. The easiest option is to have the forward/reverse/run disconnect instantly, which they should do if they're wired to PMDX outputs, but also have the input power cut at the same time. That way the VFD should have things stopped, before it runs out of power.

From a practical point, there should be no noticeable difference between using a common +5V or common 0V/Gnd.
The main thing is to use suitable wiring to eliminate/minimise noise, which could cause lost/gained steps.

When you say "cut power" to the drives, I was going to wire the motor power supply direct to the stepper motors meaning only the step/dir inputs from the PDMX would be cut. The main stepper motor power would still be connected. Is there a problem leaving the drivers in fault mode? or are people removing the power to stepper motors also?

Think you should knock up schematic which will help you and us understand better exactly what you have in mind.
Like MC says if you don't clear the fault the PMDX will never come out of fault so you'll be stuck in a loop. This means you'll need to Kill power to the drives or use the Enable/reset option on the drives.but this may not suit your plans or the PMDX but without seeing schematic it's not so easy to see what your intending regards E-stop etc.

I read in the PDMX manual, if the charge pump signal is disabled, then it will disable the PMDX-126 and all motor drivers. If the drivers are disabled, then they cannot be in fault mode right?

The UCCNC software, can be set to de-activate the charge pump signal, whenever the software is in reset - ( which occurs when passed an E-stop or Fault signal for example )

So in summary Am882 Drivers detect fault. PDMX detects the 882 fault signal and passes to UCCNC. UCCNC goes into reset, which disables charge pump. PDMX now disabled ( powered down ) as charge pump off.

Click on UCCNC reset button now powers up PDMX, which powers up motors ( fault signal effectively reset ), Continue job.

Am I missing something regarding the fault loop? or does this sound like it will work?

Attached, schematic of rough config
Attachment 21408

Quote:

---

Attached, schematic of rough config
Attachment 21408

---

This is showing an invalid attachment at my end.

Quote:

---

Originally Posted by dachopper

I read in the PDMX manual, if the charge pump signal is disabled, then it will disable the PMDX-126 and all motor drivers. If the drivers are disabled, then they cannot be in fault mode, right?


Am I missing something regarding the fault loop? or does this sound like it will work?

---

Yes, you are missing the fact the Drives will still be OUTPUT a fault signal unless the fault is cleared inside the drives. The PMDX only disables its OWN OUTPUTS it cannot control the Drives OUTPUTS and because you're watching these OUTPUTS with the PMDX INPUTS then the fault signal going to the PMDX will Never be cleared.

Also, think your misunderstanding the Charge Pump. The function of the Charge pump is to disable the PMDX outputs if the connection to the PC is lost. Shouldn't be used for E-stop or to control drives etc.

The drives could be reset using the Enable signal but not a good solution IMO because it's relying on software.

I see, I just read this from 882 manual "Reset the drive by repowering it to make it function properly after removing above problems" so, if I have one of the pdmx relays controlling the driver voltage, so it cuts driver voltage on fault, this should accomplish the job, or I can manually cycle 882 drive power to clear ?

If I'm using the relay, should it be cutting power on the DC side, or AC side of my power supply?

Nick.

Attachment 21413 try again

Quote:

---

Originally Posted by dachopper

I see, I just read this from 882 manual "Reset the drive by repowering it to make it function properly after removing above problems" so, if I have one of the pdmx relays controlling the driver voltage, so it cuts driver voltage on fault, this should accomplish the job, or I can manually cycle 882 drive power to clear ?

If I'm using the relay, should it be cutting power on the DC side, or AC side of my power supply?

Nick.

---

Always cut the ac side. Cutting the dc side risks blowing up the drives.

OK - I'm goint to try knock up a more complete schematic, and see what you think

Attachment 21420again#2

Don't like this setup because you don't have latching safety circuit. How you have it now soon as you release the E-stop the system becomes active again regardless if the fault is cleared or not.

Personally, I would run the E-Stop thru Master relay wired so forms latch. This then controls the Killing of power to high powered devices like Drives, Vfd etc. Then informs PMDX E-stop occurred which then does the Rest and turns off its Outputs. The Master Relay can only be Reset by pressing Momentary Button and will only allow the Relay to reset if all faults are clear.
The E-stop circuit can then be run using 24V giving better noise immunity.

Ok, But latching or not, the software that is running the machine will go into reset condition the second it receives an E-stop or limit switch touched, until someone physically pushes the reset button in the software, and then clicks on start again in the software. So regardless of weather the E-stop remains latched or not, The PDMX will not activate anything again as the software is in reset active mode? A latching circuit is adding another layer of buttons I have to press to get the machine started again.

What do you think about the load I was placing on the K2 relay? All i did was look at the rated Watts, or Amps per motor and figure out what that would be at 240 Volts ( 100% conversion) to guess the load at max drain for everything on? It says 10A max in the manual

A key rule is never to trust software!

What size of machine is this?
Personally, I'd ensure the e-stop circuit removed all power to anything that moves directly, which if I remember correctly from the PMDX manual, the BOB will kill all outputs as soon as the E-stop signal is lost. However, I'd add an extra relay (the onboard relay is probably not ideally rated for handling PSU start-up surges) controlled directly by the E-stop circuit (adding a reset/latch is reasonably easy - they are a legal requirement for commercial use, but hobbyist use you can do whatever you want as long as it's only you that's going to risk getting injured), which controls power to the drives.
If the E-stop directly controls the main power relay, then you can use an extra set of contacts in the relay to switch the BOB E-stop input. Unless you're using more than about one KW/KVA of power supply, I'd use a common DIN rail mounted DPDT 10A relay. RS part numbers are 488-2610 for the relay, and 803-376 for the socket to give you an idea of the type - other sources are available.

Having the main drive power supply active without the BOB active is not a problem, as the BOB/software should prevent any motion from occurring. The big problem arises when you rely on the software to kill things, and the reason you want to kill things is because the software isn't doing what it should, so you really want to make sure the E-stop kills everything, even if the computer/software is doing it's very best to not stop things.

Quote:

---

Originally Posted by dachopper

Ok, But latching or not, the software that is running the machine will go into reset condition the second it receives an E-stop or limit switch touched, until someone physically pushes the reset button in the software, and then clicks on start again in the software. So regardless of weather the E-stop remains latched or not, The PDMX will not activate anything again as the software is in reset active mode? A latching circuit is adding another layer of buttons I have to press to get the machine started again.

What do you think about the load I was placing on the K2 relay? All i did was look at the rated Watts, or Amps per motor and figure out what that would be at 240 Volts ( 100% conversion) to guess the load at max drain for everything on? It says 10A max in the manual

---

Was going to answer this then noticed Moray did the job for me.
Everyone does it there own way just some do it better than others, it's down to you to decide if think yours is better or not.!

JAZZ, most of my faults have been detected by controllers etc, but regarding the 882 error loop..... Do you know, or does anyone know 100% that it is the main motor DC power that needs to be cycled in order to clear the fault?

How are guys doing this? just Cutting the Active line before the ac/dc converter using a low voltage DC press button through a relay?

Redone Schematic, can't decide where to put the E-stop. In the pic, it overrides everything, and turns the pdmx off, but keeps the motion controller on as I don't want to cut that off.

Option 2, is rely on the pdmx safety E-stop circuit, as it will cut power to spindle / motion commands / dust extractor and coolant. then I can on / off the motors if needed to clear fault separately?

I'm normally no where near this machine when it's running, as the jobs are hours long, and I rely on it turning itself off when it faults, not sitting there watching it non stop - as such the e-stop rarely gets used as I'm not there ( No-one is )

Quote:

---

Originally Posted by dachopper

How are guys doing this? just Cutting the Active line before the ac/dc converter using a low voltage DC press button through a relay?

---

My machine uses dual motors/ballscrews on X axis, driven by EM806 (very similar to 882), so I'm a bit paranoid about stopping everything if one of the motors stalls to avoid tearing the gantry apart. To do this, I've wired the fault outputs of drivers in series with e-stop switches (all three of them) so hitting e-stop or fault signal from any driver do the same thing. I also use a safety relay with latching function. Relay is energised by momentary action switch and will drop out on e-stop or fault. Safety relay contacts (a) cut power to second relay which removes a.c. power from driver power supply, (b) cut enable signal to drivers, (c) send signal to motion controller e-stop input (I'm using CSMIO IP/M). So on fault, driver loses power so resetting safety relay (front panel push button) will reapply power and reset driver. So far, all driver faults have been motor stall, when I was trying to drive X a bit too fast. Dropped max rapid speed for X and it's been fine since.

It all depends on your paranoia quotient, but with a largely unattended machine I would probably do something similar - safety relay does the basic "stop everything" actions while the motion controller does other bits and pieces in the background, but I don't rely on it for primary safety. Downside, maybe, is that you have to hit front panel button as well as reset button in Mach3 to restart but that's not a big deal.

One man's view - I'm sure others will do it differently.

Quote:

---

Originally Posted by dachopper

JAZZ, most of my faults have been detected by controllers etc, but regarding the 882 error loop..... Do you know, or does anyone know 100% that it is the main motor DC power that needs to be cycled in order to clear the fault?

---

You have two choices to clear the Fault. Recycle power to the drives or use the Enable input.
To use the Enable input for a reset you'll need to set it up in the software. Then setup a Drive Reset button which cycles the Enable input.

Like has been suggested most use latching circuit because it's safest approach and kill power so the drives get reset by default so not an issue.

Take your pick.!

Using the enable input tickles my fancy. So I could run a series wire to a single pin, and send an enable active high/low signal to the drives. Then if the software goes into reset, the enable signal should stop all drives at the same time. Then if I click software reset, that will re-enable all drives up again as it turns enable back on.

I'm sure that would work but...

If a drive fault happens, you've lost position, or don't know if you have lost position which is almost as bad. So you will probably have to rehome, rewind the gcode to an appropriate point, and then restart. Pushing an extra button on the control panel is the least of the issues. But I don't know exactly how you plan to use the machine so maybe that's not relevant.

Again you are relying on software which isn't good or safe.

Not sure why your resistant to using Master Latching relay system because this could handle all you need and more. When fault or E-stop occurs it's single Reset button when a fault is cleared. If you want the Software to be reset at the same time that could also be done but wouldn't advise it.

Like Neale says any Fault or even software Reset which drops power to the drives or releases Motor hold will result in Re-homing so compromising your safety to save one button press just isn't sensible IMO.

Ok, so by a master latching relay, how can I mechanize it? I assume your talking about the AC power that powers the motors, spindle, and coolant being physically cut by a relay, when e-stop or fault conditions are detected and staying latched, but how can I achieve this using fault signal - like you say without software?

Neale - the relays on my PDMX perform all the functions of your latching safety relay, except they are not latched ( Unless activated by pressing the latching E-stop switch )

Quote:

---

Originally Posted by dachopper

Ok, so by a master latching relay, how can I mechanize it? I assume your talking about the AC power that powers the motors, spindle, and coolant being physically cut by a relay, when e-stop or fault conditions are detected and staying latched, but how can I achieve this using fault signal - like you say without software?

---

Creating latching relay is easy you just route the coil power thru one of the NO contacts before connecting to the coil. This breaks the circuit until the Momentary reset button is pressed. This Momentary button goes between power and the coil so when pressed powers the Relay coil which close the NO contact allowing power to flow to the coil, relay then remains latched until something breaks the latch. That could be E-stop, limits or fault signal.

You then use any remaining Contacts on the relays to operate or send signals to things that need them. What you control or inform is upto to you but no where in this does software control anything. It just gets informed whats happened so can stop code or flash messages etc.

Also don't confuse an E-stop which latches when pressed with a Latching Relay because they are not the same. The PMDX relays are just plain relays and soon as the E-stop is released they will become active again. Neils safety relay won't do that. Like wise if any of those contacts melt down closed they will Fail ON Neils Safety relay will sense this and won't allow restart.

You need to split the power into two distinct groups. One being control power, the other being motion power.

Control power you generally want to remain permanently on, and powers things like the controller, input sensors, buttons, etc. Basically anything that being still active isn't going to cause motion.

Motion power is anything that powers things that move i.e. stepper drives, spindles, solenoids. Motion power you want to kill as quickly as practical in the event of an E-stop.

If you don't want to go to the expense of using a higher voltage for controls, use a 5V version of the relay I linked to previously, and do what Jazz said to create a latch. That way you can use the contact that's providing the latch, to also control the E-stop input on the BOB.

...and to get the fault signal to trip the latching relay, you can wire the fault outputs from the drivers in series with the estop switches, and any fault output or estop operation will trip the relay. I have three driver fault outputs and three estop switches all wired in series and it works a treat. It's that easy. The set of "switches" are wired into the supply to the relay and you just have to watch the polarities of the fault outputs when you wire them in.

Thanks mate, that helps a lot.

I'm thinking then, that the PDMX and motion controllers will be always powered, unless turned off at the on/off switch.

I use the K2 relay on the PDMX, set to active if "outputs enabled", to open/close the DC power going through the latching relay setup.

Meaning, I have to turn on the PDMX first, and have the motion controller and PC up and running, then I can momentary reset to power the motor drivers?

I'll add an attachment, but my last 2 attachments have failed to post up for some reason.

Would something like this be ok for the relay http://www.ebay.com/itm/Single-Phase...YAAOSwdI9Y8ERd

Attachment 21470

Attachment 21526

Got the new power supply, a few more wires and I can fire up the beast.

I would not power that up, as by the looks of things, the drivers will release lots of smoke!

Do you have a link to the power supply?
By the looks of it, you've wired the drives up to what should be the AC input to the toroid. The two pair of wires will get connected in either parallel or series, depending on if you want 110VAC or 240VAC.
The main DC output I'd say is from the 4 terminal block at the bottom right, with the other two outputs to the left being for lower voltage/current outputs. 5 and 12VDC?

:) yes after a few beers, wiring not such a good idea.

Attachment 21528

Take 2.

Still awaiting the AC power relay, and the 882 USB to Rj cord. Does anyone know if this is an easy DIY to make up the USB to rJ connection ?

Was thinking of getting the 2 cords, splicing in the middle and connecting.

Attachment 21528

You need both a USB to serial converter, which is more than just a cable, and requires a driver, and the serial to RJ-11 cable. Here's the pinout for the RJ-11 cable.
Attachment 21551

I just ordered one from the US Leadshine distributor.

More progress,

Control box almost complete now with off latching e-stop and reset switch

Black is AC power in
White is relay controlled AC power out for Spindle / H20 pump / Chip extractor
Small black is always on DC power for the UC400 controller + feed to the PDMX BoB
Attachment 21641

Internals....
68V AC->DC converter on the left, with 12VDC supply off that holding a 40A/250AC Solid State relay at the back closed.

5x882's in the middle, ( still awaiting the tuning cable to arrive )

PMDX 126 on the right.

Still to come - install + connect the UC400ETH, secondary E-stop circuit to the E-stop button from PDMX, limit switch wires & stepper motor connectors on case, cooling fan

Attachment 21673