Twelve DL-205 stands, each driving a starter motor, engine ignition, and a fuel valve for durability testing, need one CPU decision made before any wiring: the 250-1 or the 260. The rest of the job is DirectSoft programming, a multidrop link on CPU port 2, and Lookout as the monitoring layer. Build it in that order and prove each step before moving on.
The source material dates from 2002. Treat catalog status, pricing, and software versions as unverified and check the current AutomationDirect catalog and selection guide before ordering.
Choose the 250-1 or the 260 before you order
The quick fix is to buy the cheapest CPU that runs the sequence. It fails on a durability rig because the sequence is trivial (starter, ignition, fuel valve) and the pressure comes later: a request for a serial device, a Modbus link to another instrument, or ASCII strings to a printer or scanner. The 250-1 has no ASCII and Modbus function blocks. The 260 has them.
| Item | 250 / 250-1 | 260 |
|---|---|---|
| Comm capability | Improved over earlier CPUs; port 2 is the networking port | Same class of comm, plus more functions |
| ASCII and Modbus function blocks | Not in the 250-1 | Included |
| Memory | Less | More |
| Local expansion | Added in the 250-1 | Verify in the selection guide |
| Price gap (2002) | Baseline | About 50 USD more; recheck current pricing |
| Product status | The 250 family was stated to be continuing; the 250-1 was developed to add functionality | Newer CPU |
Decision path: if all 12 stands only run the local sequence and Lookout polls them as slaves, the 250-1 is enough. If any stand may need a PLC-initiated serial exchange (ASCII in/out, or the PLC as Modbus master), buy the 260. At 12 units, the price difference is small against a mid-startup swap of 12 CPUs. Standardize on one CPU model across all stands so one program and one spare covers the fleet.
Check before moving on: the part numbers on the purchase order match the CPU you decided on, and the catalog selection guide lists the comm and function features you need.
Install DirectSoft and build one master program
DirectSoft is the programming software for these CPUs. Write the program once, on one stand, then download the identical project to the other 11. Wrong fix: editing each stand's logic separately during commissioning. That produces 12 slightly different programs and a maintenance problem the first time someone changes a timer.
- Create a project for the chosen CPU type in DirectSoft.
- Document I/O with nicknames per point: starter, ignition, fuel valve, plus the stop, run, and fault inputs.
- Set a unique station number per CPU now, before the link exists. Keep the assignment in a table taped inside each panel.
- Save the master project and archive it. Download to each CPU and record the date.
Check before moving on: a project compare in DirectSoft against each CPU shows no differences, except the station number if you keep that as the only per-stand setting.
Write the sequence so the PLC owns the safe state
The PLC, not Lookout, runs the start sequence. Lookout monitors. A dropped comm link must never leave a starter engaged or a fuel valve open.
- Fuel valve: drive it from an output that de-energizes to closed. Use a fail-closed valve so power loss or CPU stop closes it.
- Starter: limit crank time with a timer in the PLC. Add a lockout so the starter output cannot energize while the engine is already running (use a running-indication input such as speed or oil pressure, per your rig).
- Ignition: enable it only when the run sequence is active and the stop and fault inputs are healthy.
- Hardwire the emergency stop through the power path to the starter and fuel valve. Do not rely on a PLC output alone for the stop function.
- Do not write control bits from Lookout unless you add an explicit permissive in the ladder; a remote write should request, not force.
Check before moving on: with the engine disabled or the load side disconnected, run the sequence and confirm crank timeout, stop, and CPU-to-stop transition all leave the starter, ignition, and fuel valve outputs off.
Wire the outputs for starter, ignition, and valve noise
Do not connect the starter motor directly to a PLC output. Switch it with a relay or contactor rated for the starter load. Ignition systems radiate noise and solenoid valves produce inductive kickback; both cause random output faults and comm errors that look like a bad CPU.
- Suppress every coil the PLC drives (flyback diode on DC coils, RC or MOV on AC coils), per the output module and coil specs.
- Keep ignition high-voltage leads and starter cables away from the comm cable and PLC wiring; cross at right angles when unavoidable.
- Ground the shield of the comm cable at one end only.
- Read the output module's per-point and per-common current limits in its datasheet and size relays to stay below them.
Check before moving on: crank and fire the engine with the comm cable disconnected. The CPU stays in Run, no outputs chatter, and no error indicator appears on the CPU.
Connect port 2 as a multidrop link
Networking these CPUs to Lookout goes through port 2. Confirm in the CPU manual which electrical standards port 2 supports and its pinout; do not assume from a similar port on another CPU.
- Set port 2 protocol, baud rate, parity, and station address on every CPU. Every CPU on the same cable uses the same baud and parity and a different station number. Set this in DirectSoft's CPU port setup or, on the 250-1/260 generation, through the port setup memory locations documented in the CPU manual.
- Run one daisy-chain trunk, not stars. Terminate the two physical ends only, per the cable and port specs.
- Keep the PC-side and CPU-side signal polarity consistent; swapped A/B lines cause total silence, not partial errors.
Check before moving on: connect DirectSoft to the trunk and establish a link to stand 1, then stand 12. If the far end fails and the near end works, suspect termination, polarity, or a duplicate station number before touching the software.
Set up Lookout to poll all 12 stands
Lookout communicates through the PC comm port. Create one driver object for the serial link, matching port, baud, parity, and protocol to the CPUs, then add one PLC object per station address. Use a protocol the CPU port supports as a slave; confirm this in the CPU's comm port table.
- Point the driver at the same PC comm port used by the converter, and make sure DirectSoft is closed. Two programs cannot share one port.
- Map tags to the same nicknames and addresses documented in DirectSoft: run state, fault, crank timer, and cycle count.
- Keep the poll list short. Twelve stands share one serial trunk, so update rate falls as the tag count rises. Poll status tags fast and counters slow.
- The 260's ASCII and Modbus function blocks are for PLC-initiated communication. Lookout polling as master does not require them.
Check before moving on: toggle a harmless bit on one stand and see it change in Lookout; repeat for stand 12. Then confirm every stand shows current data, not stale values.
Prove the whole system end to end
| Symptom | Likely cause | Fix |
|---|---|---|
| Only some stands answer | Duplicate station number, bad termination, or a broken trunk segment | Verify station table; disconnect stands from the far end until the near ones answer |
| No stand answers | Wrong baud/parity, swapped polarity, or wrong PC port | Match settings on every CPU and driver; swap A/B; recheck port assignment |
| Errors only when the engine cranks or fires | Ignition or starter noise on the comm cable or unsuppressed coils | Reroute and shield; add suppression |
| Lookout works, then DirectSoft cannot connect | Lookout holds the comm port | Stop the Lookout driver first |
| Missing ASCII or Modbus block in the programming software | CPU type is a 250-1 | Change to the 260, or drop the PLC-initiated serial requirement |
Final acceptance, in order:
- All 12 CPUs hold the archived program (project compare is clean).
- Every stand runs crank, fire, and stop with fail-safe outputs verified.
- Pull the comm cable mid-run on one stand: the stand continues its sequence or stops safely, and Lookout flags the lost node.
- Trip the emergency stop: starter, ignition, and fuel valve de-energize on all affected stands regardless of PLC state.
- Run all 12 stands at once for a full poll cycle and confirm no stale tags.
Why does the 260 cost more than the 250-1?
The 260 includes the ASCII and Modbus function blocks and more memory, which the 250-1 lacks. The 250-1 added local expansion instead. In 2002 the gap was about 50 USD; check current catalog pricing.
Why does Lookout only reach some of my DL-205 stands?
The usual causes are a duplicate station number, mismatched baud or parity on one CPU, or a missing or extra termination on the trunk. Connect DirectSoft to the failing stand directly to separate a CPU setting fault from a wiring fault.
Why does port 2 matter for networking the CPUs?
Port 2 is the CPU port used for the multidrop link to Lookout on these CPUs. Confirm its electrical standards and pinout in the CPU manual before buying converters or cable.
Why does the comm link fail only when the engine cranks or fires?
Starter current and ignition noise couple into the comm cable or upset unsuppressed coils. Reroute the cable away from ignition and starter leads, ground the shield at one end, and add coil suppression.
Why stop troubleshooting and call AutomationDirect support?
Stop when a CPU fails to enter Run with a clean program, the comm port hardware shows no response on a bench test, or the CPU manual does not list the protocol or port you need. Contact AutomationDirect technical support with the CPU part number, firmware level, and the DirectSoft error message. Do not run an engine on a stand whose safe-state behavior you have not verified.