The 5 VDC you are hunting lives on the USB side of the adapter and never reaches the serial pins, which carry bipolar RS-232 voltages instead. A -9.91 V reading is a valid idle mark level, so the adapter is healthy. The failed path through the RJ12 breakout is a cable-mapping fault: two reversing cables in series cancel each other, so the link needs one straight-through section.
You are measuring the wrong thing, and the missing 5 VDC is normal. A USB-to-RS-232 adapter takes the 5 V the USB port supplies and spends it on its own converter; that rail stops at the USB side. The serial side carries bipolar RS-232 signals, where a logic 1, called mark, is -3 V to -15 V and a logic 0, called space, is +3 V to +15 V, both measured to signal common. The -9.91 V you read between two terminals is a valid idle line, so the adapter is alive. The fault is in the pin map through the breakout module: two reversing RJ12 cables in series cancel each other, and the panel never receives the conductor swap it is wired to expect.
What the adapter does with the USB 5 V
The 5 V the USB port supplies powers the converter inside the adapter, and it stops there. The serial side of the adapter carries RS-232 signaling and has no pin that reproduces the USB rail, so a meter across any two serial terminals reads a signal level, or zero, and 5 VDC is the wrong thing to hunt for. This is a close cousin of expecting an operator panel to draw operating power from a controller's serial port, a trap covered in HMI panel not powering from a PLC serial port cable. The health check for your adapter is a signal measurement, and you already have half of it.
The voltages RS-232 puts on the wire
RS-232 defines its signals with respect to one signal common. A mark, logic 1, is any level from -3 V to -15 V. A space, logic 0, is any level from +3 V to +15 V. Anything between -3 V and +3 V is an invalid level, and the standard sets the maximum open-circuit voltage at 25 V. An idle driver output sits at the mark level, negative, which is why your reading between a signal terminal and its reference came out near -10 V: that is a healthy idle line. Measure each signal against the module's documented common terminal. A reading taken between two signal pins names neither of them.
Symptoms and what each points to
- No 5 VDC anywhere on the terminal block, adapter plugged in alone. Normal. The rail stays on the USB side; the serial side carries bipolar signals only. Move on to the signal-level check.
- The -9.91 V reading between two terminals. A live driver idling at the mark level. Identify which documented signal it is by measuring to common.
- Adapter works direct to the panel, fails through the breakout. The adapter and the panel are healthy; the end-to-end pin map through the breakout is wrong.
- Two reversing cables in series through adjacent jacks. The two swaps cancel, the path becomes straight-through, and the crossing that put one transmitter onto the other side's receiver is gone.
- Two breakout modules wired terminal block to terminal block, tried both ways. An attempt to build a two-drop bus. One of those modules is specified for RS-485, and RS-232 will not run through it as a bus.
- A hand-wired terminal block to DB9 pigtail communicates. The adapter, its driver and its cable pass RS-232. Keep that path as the known-good reference while you re-map the module route.
Two other faults look like this one from a distance: an adapter that vanishes when the controller goes online, covered in USB serial adapter disconnects after going online to the PLC, and a COM number your programming software refuses to list, handled in remap a USB serial adapter COM port.
Example: two reversing RJ12 cables between the adapter and the panel
The inputs: the reversing cable's documented map, pin 1 to 6, pin 2 to 5, pin 3 to 4, taken from the cable's documentation; a second cable of the same type; a breakout module whose jacks meet at its terminal block; and the fact that the direct adapter-to-panel connection works over one such reversing cable. Your meter reads about -9.91 V on the adapter's idle output, inside the -3 V to -15 V mark range.
The rule: two mappings in series compose. Whatever position the first cable produces, the second cable maps that position again.
The arithmetic: pin 1 goes to 6 through the first cable, and 6 goes back to 1 through the second, so 1 → 6 → 1. Pin 2 → 5 → 2. Pin 3 → 4 → 3. The remaining positions mirror the same way. End to end, the two cables behave as one straight-through cable, 1-1, 2-2, 3-3 and so on.
The result: the working direct path swaps every pair, and the through-path swaps none. Both sides now transmit on the conductor pair that carries transmissions, and the link stays silent at the mark level, even though the adapter itself is fine.
The decision: the module path talks only when its end-to-end map equals the single reversing cable's map, pin 1 to 6, 2 to 5, 3 to 4. Two reversing sections give straight-through wiring and will never match. Replace one section with a straight-through RJ12 cable, or cross the terminal-block jumpers between the two jacks with that same swap, which turns three swaps into one.
Checks before you change a wire
- Power everything down and unplug the adapter. Continuity-map each RJ12 cable through all six positions and write the map down. Continuity names the joined contacts; it says nothing about which contact should carry which signal.
- Write the end-to-end map from the adapter's plug to the panel's plug, counting the terminal-block jumpers between the two jacks.
- Lay that map beside the direct cable's map. They must agree pin for pin.
- Power up with only the adapter and one cable attached, and measure a signal terminal against the documented common. An idle output reads a mark level between -3 V and -15 V; the reading you already captured is that measurement.
- Read the label on every module in the path. A module whose label names RS-485 comes out of an RS-232 link.
The fix, and the two-module trap
Two repairs give the panel the map it already talks on. Replace one of the two reversing cables with a straight-through section, or keep both cables and cross the terminal-block jumpers between the jacks by the same pin swap the cable uses; three swaps in a row equal one. Verify the finished map against the direct cable, then download with the same settings the direct connection used.
The two-module arrangement is the trap. RS-232 is point to point: one transmitter, one receiver, one signal common. Wiring two breakout modules together was an attempt to build a two-drop bus, and one module in that path is specified for RS-485 besides. Why that part drops the link lives in its schematic; the productive move is not another permutation of wires. Run this connection point to point through parts specified for RS-232, and when the physical run outgrows point-to-point RS-232, the interface choices and distances are laid out in RS-232 to RS-422 cable distance for an HMI panel link and HMI to PLC serial cable maximum distance by interface.
Safety note: never tie two RS-232 driver outputs onto one conductor. Two transmitters driving against each other can damage both drivers.
What to do
Recommendation: Treat the idle mark reading as the adapter's health check and close the 5 VDC question. Where the breakout module sits in the path, put one straight-through RJ12 section in place of a reversing cable, confirm the end-to-end map equals the direct cable's swap pin for pin, and download with the settings that already worked direct. If a module label in the path names RS-485, pull that module from this link and wire the panel point to point; save bus wiring for an interface built for a bus.
Questions
Can I power a small device from the serial pins of a USB RS-232 adapter?
No. The USB rail stays on the USB side of the adapter and never reaches the serial pins. The serial pins carry bipolar RS-232 signal voltages, and the driver circuits behind them are built to signal, not to supply a load. Anything in the cabinet that needs 5 V must take it from a USB port or its own power supply.
How do I identify the TX terminal without a wiring sheet?
Put the meter on a signal terminal with respect to the documented common and watch it while characters are sent. The terminal that sits at a negative mark level while idle and moves toward positive when characters go out is the adapter's transmit output. A terminal that floats or holds a dead level until the far device answers is a receive input.
What does a 0 V reading on an RS-232 line mean?
Zero volts sits in the dead zone between -3 V and +3 V, where the level is neither mark nor space. It usually means the driver behind that terminal is off, the terminal is unloaded, or the reference point you used is the wrong common. Retest against the documented common terminal before condemning the adapter.
Planning information, not legal or engineering advice. Rules differ by place: check codes and permits with your local authority and your licensed trades.