An entire option-slot output card going dead after a routine download is almost never the card. A one-rung edit that bypasses a sensor cannot kill sixteen points at once, so the edit is not the cause — the download event is. What actually changed is the CPU's I/O map, and once you clear memory you lose the only copy of it that was keeping the machine alive.
Stop Swapping Parts and Prove Where the Break Is
The reflex on night shift is reseat, swap the ZIPLink, swap the module. Those moves are cheap, and on this failure mode they occasionally look like they worked — a lane comes to life after a few power cycles, you get output voltage "a couple of times," then it dies again. That intermittent recovery is not a loose connection. It is the CPU re-reading installed hardware on power-up and briefly agreeing with a stale, retentive I/O map that a later download or memory clear then overwrites.
Before you pull anything else off the shelf, split the fault at the module status LED. Go online with the programming software, open a data view on the coil the ladder drives, and watch three things at once: the image-table bit, the module LED, and the voltage at the field terminal.
| What you see | Where the break is | Next check |
|---|---|---|
| Y bit ON in image table, module LED off | CPU-to-module path: I/O configuration, slot seating, module fault | Open the I/O configuration online; read the CPU error history |
| Module LED on, 0 V at the field terminal | Field side: supply, common, ZIPLink pinout, failed driver | Measure the field supply at the card's own supply terminals, then across the load |
| Y bit never turns ON in the image table | Ladder logic or addressing | Check for a duplicate coil, a disabled rung, or a coil left stranded by the sensor-bypass edit |
| Nothing anywhere, CPU not in RUN | CPU-level fault | Read the error code through the programming software or handheld before clearing it |
Check: disconnect the field devices, set the Y bit, and look at the LED. If the LEDs stay dark with a known-good card, the problem is upstream of every part you were about to swap. Put the spares down.
Confirm the Card, the ZIPLink, and the Field Supply Match
Read the part number stamped on the card itself, not off the drawing. D0-16T2 is not a valid AutomationDirect module number; the nearest real part is D0-16TD2, and that card does not carry a terminal labeled V+. So either the card is something else, or the V+ and V0 you are probing are silk-screened on the ZIPLink board, not the module.
That distinction matters. A ZIPLink module is a passive pass-through — it switches nothing. Its terminal labels are printed for one specific module family, and if the ZIPLink module or cable is not the pair listed for your output card, field power lands on terminals the card never drives while the switched points come out somewhere you are not measuring.
- Record the module part number from the case, the ZIPLink module part number, and the cable part number.
- Verify that trio against the AutomationDirect ZIPLink selection table for that module. Mismatch here explains a perfect-looking supply reading with dead outputs.
- Confirm sinking versus sourcing from the module datasheet wiring diagram, and confirm which side of the load returns to common.
- With the cable seated, measure the field supply at the card's power terminals. DC output cards do not source load current from the PLC's internal supply.
- Check the module fuse if the card has one, and the field supply fuse.
Check: supply voltage present at the card's supply terminals and continuity from the ZIPLink terminal to the intended card point. Voltage on the supply rails proves the field supply and the pass-through cable — it proves nothing about whether the CPU is updating that slot.
Read the I/O Configuration, Not Just the Ladder
Here is the mechanism that burns people on the DL06. The option slot map — which module sits in which slot and what X/Y addresses it starts at — is stored in a system configuration area, separate from program memory and separate from the ladder. Reading the program out of the CPU into a project file does not bring that map with it. A ladder download does not overwrite it either, which is why the machine ran fine for months and survived a small edit.
Clear the memory and the map goes with it. Now the download restores the ladder and whatever slot configuration the project file happens to carry — which on this machine did not match the real hardware. Coils write to Y addresses that map to nothing, no compile error is generated, and every point on the card stays at 0 V.
Check: go online, open the I/O configuration dialog, and compare the module type and starting address in each populated slot against the addresses the ladder actually writes. Also read the error history for an I/O configuration mismatch flagged on the last power-up.
Set the Expansion Addressing by Hand
Expansion slots are configured manually. Anything at 100 and above in octal belongs to an option slot; the DL06 base points occupy X0-X23 and Y0-Y17. On the machine in question the correct map was X100 and Y100 for the first populated slot and Y200 for the next.
- List every X and Y address the ladder uses above the base range. That list is your target map.
- Put the CPU in PROGRAM.
- In the I/O configuration dialog, select the actual module for each populated slot and set the starting address to match the ladder.
- Write the configuration to the CPU. This is a separate operation from the program download — doing one does not do the other.
- Power cycle and confirm the CPU comes up in RUN with no I/O configuration error.
Remember the addressing is octal: a 16-point card at Y100 occupies Y100 through Y117. There is no Y118.
Check: with the CPU in RUN, the image-table bit and the module LED now agree on every point you drive.
Verify End to End Before You Hand It Back
- Reconnect one lane at a time. Force each Y and confirm the field device actually moves, not just that the LED lights.
- Watch the field supply under full load for sag or nuisance tripping — the earlier symptom set included random glitches, and an overloaded or intermittently shorted field supply produces the same dead-output picture from a different direction.
- Run the production program and exercise the bypassed-sensor logic. Confirm the bypass edit did not leave a coil double-driven or a rung permanently false.
- Power cycle twice and confirm the outputs come back both times with no re-download. This is the test that proves the configuration is stored in the CPU rather than merely resident from your last write.
- Let it run a full cycle count before you leave it.
Back Up Program and Configuration Separately
Rebuild the backup so the next person does not repeat this. Pull the ladder, then pull the I/O configuration as its own step, then capture retentive V-memory ranges, comm port parameters, and any password setting. Tape a one-page slot map inside the panel door listing slot number, module part number, ZIPLink part numbers, and starting address. Verify the backup by restoring it offline and comparing the slot map field by field — a project file that opens cleanly can still carry the wrong expansion addressing, which is exactly how this started.
Stop if the configuration writes without error and the slot still refuses to update, or if the CPU throws an I/O error with a verified module correctly mapped. With one PLC on site and no spares, further swapping tells you nothing. Call AutomationDirect technical support with the CPU firmware revision, module and ZIPLink part numbers, the slot map, and the exact error code — and do not clear memory again before that call, because the current CPU state is the evidence.
FAQ
What happens if I clear PLC memory on a DL06 before backing up the I/O configuration?
You lose the option slot map. The ladder downloads and compiles normally, but writes to option-slot addresses such as Y100 or Y200 go nowhere and every point on that card stays at 0 V. You have to re-enter the module type and starting address for each populated slot by hand.
What happens if the ZIPLink module part number does not match the output card?
The terminal silk screen no longer corresponds to the card's points, so field supply reads correct at the terminal block while the switched outputs land on pins you are not measuring. Verify the module, ZIPLink module, and cable as a matched set against the AutomationDirect selection table before trusting any label on the connector board.
What happens if the expansion slot starting address does not match the ladder?
Nothing errors at compile time. Coils drive image-table bits that are not mapped to hardware, so the module LEDs stay dark and the outputs never energize. Depending on what else changed, the CPU may or may not flag an I/O configuration mismatch on power-up — read the error history rather than assuming.
Why does an output read 0 V when the field supply terminals show voltage?
Supply voltage at the terminals only proves the field supply and the ZIPLink pass-through cable are intact. Split the fault at the module status LED: LED off means the CPU is not updating that slot (configuration or addressing); LED on with 0 V across the load means field wiring, common, sink/source orientation, or a failed driver.