Digital Weight Transmitters

Three different instruments sit between a load cell and the rest of a system, and they are chosen for three different reasons.

A digital indicator is built for a person. It has a display, a keypad, tare, net and gross, a print key and, where it matters, an NTEP certificate. If a human has to read the weight, or the weight is the basis of a sale, that is the instrument. Those are on the digital indicators page.

A signal conditioner is built for a wire. It takes the few millivolts a load cell produces and turns them into something that survives a long run and lands on a standard input: 0–10 V, or a 4–20 mA loop. It does not know what a weight is, and it never applies a calibration. Those live under signal conditioners.

A digital weight transmitter is built for a controller. It digitises the bridge signals where they arrive, applies the calibration itself, and hands the PLC a finished weight value over a network the PLC already speaks. Nothing analog leaves the box, so there is no analog card to scale, no 4–20 mA range to map, and no drift between the box and the input card. The LCT-2 does one more thing the other two cannot: because each load cell has its own converter, the host can ask about the load cells one at a time.

Order the protocol the controller already runs. Nothing about the measurement changes between them: the same four converters, the same calibration and the same diagnostics sit behind every option. Only the interface board differs.

RS-485 with Modbus RTU and Modbus ASCII is standard and costs nothing extra. It is half duplex, it multi-drops, and it is the cheapest reliable way to get weight into almost any controller built in the last thirty years. RTU is the compact binary form and is what you want for speed; ASCII is readable on a terminal, which makes it the easier one to commission by hand. Note that on the EtherNet/IP build the RS-485 port is fixed at 19200 baud, 8 data bits, even parity, 1 stop bit, no flow control, because that port is also the configuration and calibration channel.

Modbus TCP/IP adds $200 and puts the same Modbus messages inside a TCP wrapper on an RJ-45 port. If your SCADA package or PLC already polls Modbus over Ethernet, this is the smallest change to the rest of the system.

EtherNet/IP adds $440 and is the usual answer for Rockwell and Allen-Bradley hardware. It presents as a generic module with a Class 1 implicit I/O connection, so the weight arrives in the controller tag database on every scan rather than being polled by ladder logic.

PROFIBUS, PROFINET, CANopen and DeviceNet are among the twenty industrial network protocols the platform supports and are quoted per order. If the plant standard is Siemens or an existing device-level bus, tell us the protocol and we will confirm the build before you order.

An indicator wires its load cells in parallel and the published count is really an excitation budget: add load cells, divide the bridge resistance, and eventually the supply runs out. The LCT-2 is not built that way. It has four separate terminal blocks and four separate A/D channels, so four is a hard limit and each load cell is measured on its own.

Each block takes a full six-wire connection: +EXC, −EXC, +SE, −SE, +SIG and −SIG. The sense pair is not optional decoration here. It lets the transmitter measure the excitation voltage as it actually arrives at the load cell rather than as it left the terminal, which is what cancels the voltage drop in a long cable run. Excitation is +3.3 V DC and the input range is ±3.125 mV/V, which covers any standard 2 mV/V or 3 mV/V load cell.

One installation detail catches people out often enough to be worth stating twice: if you connect fewer than four load cells, the +SIG and −SIG terminals on every unused input must be shunted with a jumper wire. An open input is not read as absent, it is read as a fault.

The 24 V DC supply is external and not included, drawing 0.5 A. There are two sets of power terminals rather than one, so several transmitters can be daisy chained onto a single 24 V bus instead of each needing its own supply.

Calibrating a large installation conventionally means getting certified test weights to it, or substituting material and weighing that instead. Both are slow, and on anything elevated or enclosed both are expensive.

The LCT-2 offers a calibration-free mode that scales the transmitter directly to the mV/V calibration data printed on each load cell's certificate. Because every load cell is on its own channel, each one can be scaled with its own figure rather than with an average across the set. For a great many process installations that removes test weights from the commissioning plan entirely.

When weights are available, the transmitter also supports single-point span, multi-point span for installations where mid-range linearity matters, and digital corner balancing, which trims the load cells so the same load reads the same regardless of where it sits. On an analog junction box corner trimming is done with potentiometers, by hand, in the box. Here it is a number, set remotely, and it does not drift.

Configuration, calibration and diagnostics are all done over the RS-485 port, so a commissioned unit can be re-scaled without opening the enclosure.

This is the argument for a transmitter over a junction box, and it is worth being precise about. A summing junction box adds four analog signals into one before any instrument sees them. When one load cell drifts, corrodes or loses a leg, what reaches the indicator is a single number that is quietly wrong. Nothing in the system knows which load cell caused it, and often nothing knows anything is wrong at all until a product weight is out of tolerance.

The LCT-2 converts each load cell separately and runs real-time diagnostics on the system, so a fault is detected as it happens and can be raised as an alarm in the controller rather than found later. The host can read the A/D counts and the weight value for each individual load cell as well as the compound weight, which turns a fault into a specific instruction: replace the load cell on channel three.

It also offers dead load cell emulation, which substitutes a value for the failed channel so the system keeps producing a usable reading until the load cell can be changed. That is a deliberate trade: the reading is an approximation while emulation is active and should not be treated as a measurement of record. What it buys is the choice of when to stop, rather than stopping now.

The same architecture helps before anything fails. Because conversion happens at the transmitter, the long analog homerun cable that usually sits between a junction box and an instrument is removed from the signal path, and with it most of the noise and voltage-drop error that run picks up.