Resolving CX-Programmer Set Value Range on Unnamed DM Symbols

Problem Overview

When commissioning or troubleshooting an Omron CJ1, CS1, CS1H, CP1, or CP1E program in CX-Programmer, an engineer often needs to Set Value (Force / Online Edit → Set Value) to a DM area word such as D100, D200, or D3000 to verify scaling, deadbands, setpoints, or floating-point constants. The expected workflow is: select the word → Online → Set Value → enter a decimal or hex constant → confirm.

The defect symptom is that the Set Value dialog only accepts values inside one of two windows:

• &0 to &65535 (unsigned decimal)
• #0000 to #FFFF (unsigned hexadecimal)

Negative numbers, signed 16-bit values below -1, 32-bit double-word values, BCD, or IEEE-754 REAL values cannot be entered. The dialog appears to grey out the sign or simply rejects the value and reverts to the unsigned range. When the same address is right-clicked and changed to display as Integer or Real, the full -32,768 to +32,767 (INT) or -3.4E+38 to +3.4E+38 (REAL) range suddenly becomes accessible — but on the next re-open of the project the setting is lost and the original restriction returns.

Root Cause: Unnamed Symbols Default to CHANNEL

CX-Programmer's symbol table uses two distinct mechanisms for labelling an address:

1. Symbol (named global / local): A registered entry in the global or local symbol table with a Name, Data Type, and Address. The data type can be BOOL, CHANNEL, INT, DINT, UINT, UDINT, WORD, DWORD, BCD, REAL, LREAL, STRING, etc.
2. Comment only (unnamed global symbol): A legacy mechanism in which the user types a comment directly above an instruction operand without registering a symbol. CX-Programmer still tracks the address internally as an Unnamed Global Symbol — a holdover from earlier SYSWIN and CPT support when symbol names were limited to 6–8 characters.

Unnamed Global Symbols carry no explicit data type. To stay backward-compatible, CX-Programmer restricts them to two implicit types only:

Because no data type is registered, CX-Programmer cannot render the value as INT (signed), BCD, or REAL — even though the bit pattern in the PLC is identical. When the dialog's right-click context menu is used to switch the view to Integer or Real, CX-Programmer applies an ad-hoc display override for that session only. The override is not saved with the project and reverts the next time the symbol is opened.

Editing the symbol via Edit → Symbol or right-clicking the address and choosing Edit Symbol reveals the underlying limitation: the Data Type dropdown is greyed out to BOOL or CHANNEL only. CX-Programmer explicitly disallows promotion of an unnamed symbol to REAL or DINT.

Why Some Addresses Show Integer or Real Automatically

If the engineer has previously registered a named symbol pointing at the same DM address — for example, PID_Error as REAL, or Counter_PV as INT — CX-Programmer stores the data type in the symbol table of the .cxp project file. From that point onward, every reference to that address inside the project, including bare operand entries in ladder, inherits the named symbol's data type. The Set Value dialog then opens with the full numeric range, sign, and decimal-point support appropriate to the data type.

Common patterns that produce a working symbol without realising it:

• The address is referenced inside a Function Block (FB) instance whose internal variables map onto the DM area.
• CX-Programmer auto-creates a symbol when the engineer types a name into the Symbol column of a rung operand (instead of the Comment column).
• The project is imported from CX-Integrator / Sysmac Studio round-trip conversion, which re-emits symbols with explicit data types.

Solution 1 — Register a Named Symbol with the Correct Data Type

This is the canonical, long-term fix and applies to CJ1, CS1, CS1H, CP1, CP1E, and CP1L PLCs supported by CX-Programmer V9.x and earlier.

Prerequisites

• CX-Programmer V7.0 or later (recommend V9.76 or later for full CP1 support — see the Omron Industrial Automation portal for current version notes).
• Project is offline or in Program mode; symbol changes are not accepted while in Run mode and the dialog will refuse the edit.
• Right to modify symbols: Protection → Release Password if the project is write-protected (UM protection or task-level protection).

Step-by-Step

1. Open the project .cxp / .cxm file in CX-Programmer.
2. Press F8 or choose View → Symbol Table → Global Symbols (or Local Symbols inside an FB).
3. In the first empty row, type the desired name in the Name column, e.g. PID_SP.
4. Type the address in the Address/Value column, e.g. D200.
5. Click the Data Type cell and select the appropriate type:

6. Click outside the row to commit. CX-Programmer validates the type against the address family and rejects the entry if the address is incompatible (for example, attempting to assign REAL to a single bit W0.00).
7. Repeat for every DM address that will receive negative or floating-point data.
8. Save the project (Ctrl+S). The .cxp file now contains the data-type metadata; reload the project to verify persistence.
9. Go online, place the PLC in Program mode, then return to Run. Right-click the symbol in the watch window or on a rung operand → Set Value. The dialog now opens with the full INT or REAL entry range and a sign field.

Solution 2 — Loading Floating-Point Constants Without a REAL Symbol

For legacy programs where renaming an address to a named REAL symbol would break documentation or HMI tag mapping, there is a well-known workaround that pre-dates the symbol-table fix. It uses only the +F (Floating-Point Add) instruction and exploits the fact that the destination operand of +F and MOVL may be any 32-bit-aligned word pair, even one without a registered symbol.

Step-by-Step (loading 50.0 into D3000/D3001 as REAL)

1. Insert a rung with the +F (Floating Add) instruction:

[ +F    +5.0E+1    +0.0E+0    D3000 ]
   S1       S2         D

2. Source 1 (S1) is the hard-coded floating constant +5.0E+1 — CX-Programmer accepts scientific-notation floats in the operand entry even when the destination is unnamed.
3. Source 2 (S2) is +0.0E+0 to neutralise any garbage previously in D3000/D3001. Alternatively, point S2 at a known-zero REAL register such as D0 after first clearing it with a MOV #0000 D0 rung.
4. The destination D is D3000. The +F instruction interprets the destination as a 32-bit REAL word pair, placing the IEEE-754 single-precision encoding of 50.0 across D3000 (low word) and D3001 (high word).
5. To copy the value to a different location preserving REAL type, use MOVL:

[ MOVL    D3000    D17020 ]
       S       D

5. D17020 is treated as the start of a 32-bit REAL block by all subsequent floating-point instructions ( +F, -F, *F, /F, CMP with F modifier, SQRT, etc.) even though no symbol is registered there. The PLC is unaware of CX-Programmer's symbol table — the type lives in the project file, not in the controller.

This technique works for CJ, CS1, CS1H, and CP1H/CP1L CPUs. The CP1E E10/E14 type does not support REAL instructions at all and will flag the +F rung as an instruction error at download time.

IEEE-754 Single Precision on Omron REAL

Once a value is stored as REAL on CJ/CS1 hardware, the bit layout is identical to any other IEEE-754 32-bit single-precision float. This is what allows the bit pattern to round-trip through HMI scripts, OPC tags, and Python helpers without manual re-encoding.

For example, +5.0E+1 (50.0) encodes as 0x42480000: sign=0, exponent=10000100₂ (132 − 127 = 5), mantissa=10010000000000000000000₂ (0.5625). The reconstructed value is (1 + 0.5625) × 2⁵ = 1.5625 × 32 = 50.0. CX-Programmer renders this in the watch window as +5.00000E+1.

Verification

1. After registering the symbol and downloading in Program mode, return to Run. Right-click the address in the watch window → Set Value. Confirm the dialog now shows a sign toggle and accepts -32768 through +32767 (INT) or -3.4E+38 through +3.4E+38 (REAL).
2. Enter -1 (INT) or -1.0E+0 (REAL) and click OK. The watch window should display the value with a leading minus sign.
3. Power-cycle the PLC or perform a project save → close → reopen → connect cycle. Reload the project file and verify the symbol's Data Type column still reads INT / REAL and the address still maps to the same DM word.
4. From a second engineering workstation, copy the .cxp file and open it. The Set Value dialog should behave identically — this proves the data type is stored in the project, not in a per-workstation setting.

Common Instruction Reference

Troubleshooting Matrix

Best Practices

• Always name symbols in the global or local symbol table. Comments are a documentation convenience; only names carry data type. A consistent naming convention (e.g. FB1_PID_CV_REAL) makes the Set Value behaviour predictable and portable across engineering workstations.
• Document the data type next to the address in the rung comment even after the symbol is named. Engineers returning to the project after several years will see (REAL) in the comment and avoid the CHANNEL-only trap.
• Group REAL data on even-word boundaries. CJ/CS1 do not enforce alignment for MOVL or floating-point instructions, so a stray MOVL with an odd-word destination silently overlaps the next variable. Reserve even D-words (D0, D2, D4, …) for 32-bit and 64-bit values.
• Use the workbench View → Watch Window with the data type column visible (View → Show Data Type) to confirm at a glance what the controller is interpreting each address as.
• When porting to NJ/NX (Sysmac Studio), the same logic applies but the data type system is stricter: variables must declare INT, LREAL, etc. up front, and there is no "Unnamed Global Symbol" fallback. The legacy CX-Programmer symbol table effectively becomes a de-facto contract for the new project.

Frequently Asked Questions