A load cell transmitter powers a bridge sensor, reads the few thousandths of a volt it returns, and sends the result onward as something a control system can use — a 4–20 mA loop, a voltage, a Modbus register, a relay contact. What it does not do is show you a number.
That omission is the entire design, and it buys more than it costs.
Without a display it is 22.5 millimeters wide and clips onto a DIN rail. No panel cutout, no gasket, no hole in a cabinet door, no real estate on a fascia that somebody else has already allocated. It goes in the cabinet alongside the terminals, the breakers and the I/O modules, in the space a couple of relay bases would take, and it is fastened by pushing it onto the rail. For a machine builder counting door space and drilling time, that is a straightforwardly cheaper way to add a measurement.
More importantly, a display is only useful to somebody standing in front of it. A transmitter takes the opposite view: rather than making a person walk to where the sensor is, it carries the measurement to where the decisions are already being made — the PLC, the SCADA screen, the control room, the alarm annunciator. In most industrial installations that is the more useful direction of travel, and it is why a transmitter and a panel meter are different products rather than better and worse versions of one.
But the specification that actually matters here is isolation, and the page mentions it in a single sentence.
Power, signal input, analog output, relay outputs and communications are each isolated to 250 volts rms, tested at 4.2 kilovolts. That is five separate barriers, and it is worth understanding why a device this small needs them.
Look at where those five connections go. Mains or control power comes from the panel's supply. The signal input goes out to a sensor bolted to a machine, which has its own earth. The analog output goes to a PLC, which is grounded through its own rack and power supply. The relays go to field devices somewhere else again. The serial port goes onto a plant network shared with equipment nobody in your building specified.
Those are five different ground systems, and they are not at the same potential. In a plant they never quite are — motors, drives, welding, long cable runs and ordinary building wiring all see to that. If the transmitter tied them together, it would become the point where those grounds meet, and the difference between them would drive current through the one path they now share: your measurement. That is the classic ground loop, and on a millivolt signal it does not take much of one to matter.
Isolating every port means each side floats independently. The sensor's ground stays the sensor's; the PLC's stays the PLC's; nothing has to agree with anything else. The measurement is unaffected, and a fault on any one of those five circuits has a tested four-kilovolt barrier between it and the other four.
That, rather than the absence of a screen, is what makes this a transmitter rather than a meter with the display left off. It is built to be the point where a sensor meets a control system, which is electrically the most awkward place in the installation to stand.
Excitation goes out to the sensor, millivolts return, the transmitter converts and scales them, and the result leaves by whichever route you configured. The choices worth understanding are the input range, the filter, and how much of the setup the sensor can do for itself.
Five input ranges, and picking the right one is worth a factor of ten. Full scale can be set to 20, 50, 100, 250 or 500 millivolts. Resolution is not uniform across those: the bottom pair resolve ten times more finely than the upper three, at a microvolt against ten microvolts. So the rule is to choose the smallest range your signal genuinely fits inside. Work out what that signal is before selecting — sensitivity multiplied by excitation voltage. A 2 mV/V load cell at the default 10 volt excitation produces exactly 20 millivolts at capacity, which lands precisely on the most sensitive range and gives you twenty thousand counts of working resolution. The commonest sensor in the catalog and the finest setting on this transmitter fit each other exactly.
The input itself is close to invisible to your sensor. Input resistance is a gigohm — roughly three million times the resistance of a 350 ohm bridge — which means the transmitter draws essentially no current from the signal and cannot load the bridge or disturb whatever else is connected to it. It also tolerates up to 35 volts applied to the input without damage, which is more forgiveness than a miswiring in a control panel is likely to need.
The digital filter runs to nine and a half seconds, and that is a feature. Time constants are selectable from eighty milliseconds up to 9.6 seconds, which is by far the slowest deliberate filtering anywhere in our range — and exactly right for what this product does. A transmitter feeding a control loop does not want speed. It wants a number that does not jitter, because a jittery input makes a loop hunt, makes an alarm chatter and makes a trend chart unreadable. Material sloshing in a vessel, an agitator turning, a conveyor running past, a pump cycling: none of that is the measurement, and a long time constant removes all of it. Choose the shortest constant that gives you a steady number, and be prepared for that to be far longer than instinct suggests.
The noise rejection figures are quoted before the filter helps. Common mode rejection is 130 decibels, which as a ratio is better than three million to one, with 250 volts rms of common mode voltage withstood. Normal mode rejection at the power line frequency is 90 decibels with no digital filtering at all — that is the starting point, and any filtering you select is on top of it. The conversion rate matches the power line for the same reason: sixty times a second on 60 Hz supplies, fifty on 50 Hz, so each conversion spans a whole line cycle and the mains largely cancels itself.
Excitation is isolated, and the sensor can have a say in it. The default is 10 volts at up to 120 milliamps, with 5 volts at 100 milliamps and 24 volts at 50 milliamps also available. Nearly everything we make is characterized at ten volts, and it is also the setting that gets you the most counts. Check the current against your bridge rather than assuming: 24 volts across a 350 ohm bridge asks for more current than that setting provides, so the highest voltage suits higher impedance sensors rather than a standard load cell. The excitation output is isolated by 50 volts from the instrument's own ground, which keeps the sensor's end of the circuit separate from everything else.
With a TEDS sensor, much of that setup happens without you. Connect a transducer carrying an IEEE 1451.4 memory and the transmitter reads what the sensor says about itself — what it is, what it needs and what its calibration is — and configures accordingly. For a sensor without that memory, setup runs from a PC over the serial connection using the instrument setup software, which is the easier route anyway when several transmitters need the same configuration.
And the outputs are unusually complete for something this size. A sixteen-bit isolated analog output in your choice of 4–20 mA, 0–20 mA, 0–10 V or plus and minus 10 volts, accurate to two hundredths of a percent of full span. Serial in RS232 or RS485, half or full duplex, speaking Modbus RTU, Modbus ASCII or plain ASCII, all software-selectable rather than ordered as options. And two solid state relays for limits. A single module handles the analog path, the digital path and the alarm path at once.
Specifying one of these is mostly about deciding what has to receive the measurement, and being honest about how quickly the thing you are measuring moves. Show our applications engineers the loop end to end rather than the transmitter on its own; that is where the answers usually are.
Start with who is consuming the number. A PLC or a chart recorder expecting a process signal wants the analog output, and 4–20 mA is the right default over any distance because a live zero at the bottom of the range distinguishes a dead loop from a genuine reading of nothing. A supervisory system wants the serial port and Modbus. A local device that simply has to act on a threshold wants a relay. You can have all three at once from one module, which is worth knowing before anyone specifies three devices.
Then ask whether anybody actually needs to see a number at the machine. If a maintenance technician standing at the vessel needs to read the weight, a transmitter alone will not give it to them — you need a display fed from the analog output, or a panel meter instead of a transmitter. If the reading is only ever consumed by a control system, a display is a hole in a door and a thing to keep clean. Decide deliberately rather than by habit.
Choose the input range from arithmetic, not from the middle of the list. Multiply your sensor's sensitivity by the excitation you intend to use, and pick the smallest range that number fits inside. Selecting a range far wider than you need throws away resolution you already paid for and cannot recover downstream.
Set the filter last, and set it by experiment. Start short, watch the output while the process runs, and lengthen the time constant until the reading is as steady as your control strategy requires. What you are trading is response time, so the right answer depends on the loop rather than on the sensor — a tank that takes ten minutes to fill can afford a long constant, a machine that has to react to a load change cannot.
Match the power option to the panel. The standard version accepts a wide range of line voltage and will also run from high voltage DC. There is a low voltage version for panels built around a 24 volt control bus, which saves bringing line voltage to the module and simplifies both the wiring and the paperwork. Decide before ordering.
Check the rail and the space, not just the width. It clips to standard 35 millimeter rail, and at 22.5 millimeters it takes about the width of a couple of terminal blocks — but allow for the detachable connectors and the cable bend radius in front of it, which is where DIN rail layouts usually run out of room rather than along the rail itself.
Specify TEDS on the sensors if you want the setup to look after itself. The Cal-Teds plug and play option puts an IEEE 1451.4 memory on the sensor so the transmitter can read what it is connected to. On one permanently installed sensor that matters less. Where sensors get swapped, recalibrated or moved between machines, it removes the step where the right sensor gets the wrong configuration — and on a transmitter with no display, that error is even harder to notice than usual, because there is no number in front of anyone to look wrong.
Send us the sensor, its sensitivity and bridge resistance, what has to receive the signal and in what form, how fast the measurement changes and what the panel supply is, and we will confirm the configuration before you order.
SST-HV Load Cell Transmitter Applications.
The Transducer Techniques SST-HV is a plug and play smart TEDS IEEE 1451.4 compliant transmitter that self-calibrates to a TEDS load cell or torque sensor, mounts on standard 35 mm DIN rail, and provides isolated analog, serial and relay outputs from a single module.
- PLC and Control System Integration: An isolated 4–20 mA or voltage output delivers force, weight or torque to a controller as an ordinary process signal.
- SCADA and Plant Networks: Modbus RTU or ASCII over RS232 or RS485 lets a supervisory system read the measurement directly without an intermediate gateway.
- Tank and Vessel Monitoring: Long filter time constants produce a steady contents reading despite agitation, sloshing and machinery running nearby.
- Panel and Cabinet Installation: A 22.5 mm DIN rail module adds a measurement channel without a panel cutout or door space.
- Electrically Hostile Plant: Isolation on power, input, analog output, relays and communications keeps five separate ground systems from meeting in the measurement.
- Process Control Loops: Selectable filtering delivers a signal steady enough that a control loop settles instead of hunting.
- Overload and Limit Alarms: Dual solid state relays act on thresholds locally, independently of whatever is reading the analog or serial output.
- Machine Builder OEM Use: A compact, CE approved, US-made module that installs on rail as a standard component of a control system.
- Multi-Sensor Installations: One transmitter per sensor on a shared RS485 line gives independent channels that a single supervisory system can poll.
- Sensor Interchange: TEDS self-configuration allows a transducer to be replaced or moved between machines without reprogramming the transmitter.
The SST-HV works with our full range of load cells, torque sensors and pressure transducers, and is supported by TEDS reader-editor software, data logging software and serial cable accessories.
Frequently Asked Questions
What is the SST-HV, and how does it differ from a panel meter?
It is a DIN rail mounted transmitter: it excites a bridge sensor, converts the millivolt signal and sends the result onward as an analog output, a serial message or a relay action — but it has no display. That single difference changes what it is for. A panel meter needs a cutout in a cabinet door and shows a number to whoever is standing there. A transmitter is 22.5 millimeters wide, clips onto rail alongside the other control components, and carries the measurement to the PLC, the SCADA system or the control room instead. Both read the same sensor equally well; they deliver the answer to different places.
What does it mean that every port is isolated?
Power, signal input, analog output, relay outputs and communications are each isolated to 250 volts rms and tested at 4.2 kilovolts — five separate barriers. It matters because those five connections go to five different places with five different grounds: the panel supply, a sensor on a machine, a PLC in a rack, field devices, and a plant network. In a working plant those grounds are never quite at the same potential, and a device that tied them together would become the junction where the difference between them drives current through the one path they share — your measurement. Isolating each port lets every side float independently, so nothing has to agree with anything else.
How does TEDS work with this transmitter?
Connect a sensor carrying an IEEE 1451.4 memory and the transmitter reads what the sensor says about itself — identity, requirements and calibration — then configures itself to suit, so there is no scale factor to enter and none to get wrong. That is worth more on a transmitter than on a meter, because a transmitter has no display: a wrong configuration produces a confident, plausible, entirely incorrect signal with no number in front of anyone to look wrong. For a sensor without TEDS, setup runs from a PC over the serial connection using the free instrument setup software.
Which input range should I select?
Whichever is the tightest fit around your signal, because the choice is worth ten to one. Only the 20 and 50 millivolt settings give you microvolt resolution — step up to 100 millivolts or beyond and each count is worth ten times as much. Calculate your full scale signal first — sensitivity multiplied by excitation voltage. A 2 mV/V sensor at the default 10 volt excitation gives 20 millivolts, which lands exactly on the most sensitive range and yields twenty thousand counts across the sensor's working range. Choosing 500 millivolts for that same sensor would work perfectly well and would discard most of the resolution.
What excitation does the transmitter supply?
Ten volts at up to 120 milliamps as supplied, with five volts at 100 milliamps and 24 volts at 50 milliamps also available, and the output isolated from the instrument's own ground. Our load cells and torque sensors are calibrated at ten volts as a matter of course, so leaving it there keeps your numbers consistent with the certificate. Check the current against your bridge rather than assuming, because the highest voltage setting supplies the least current: 24 volts across a 350 ohm bridge asks for considerably more than that setting provides, so it suits higher impedance sensors or powering a transmitter rather than a standard load cell.
What analog output does it provide, and what will it drive?
Sixteen-bit isolated output, user selectable between 4–20 mA, 0–20 mA, 0–10 V and plus or minus 10 volts, accurate to 0.02 percent of full span. In current mode it will drive up to 500 ohms of loop; in voltage mode it wants 5 kilohms or more. The bipolar range is worth noting if you are measuring in two directions, since tension and compression or clockwise and counterclockwise arrive as positive and negative without anything extra. For most plant work 4–20 mA remains the right default, because a broken wire reads as zero current, which is outside the valid range and unmistakably a fault.
What serial options and protocols are supported?
RS232 or RS485 in half or full duplex, user selectable, speaking Modbus RTU, Modbus ASCII or a plain ASCII format, all chosen in software rather than ordered as separate options. Baud rates run from 300 up to 19,200, and the connections are detachable screw terminal plugs. RS485 is the one to use for a plant network, since it runs a long way and lets several transmitters share one line; RS232 suits a single device close by, and is also how you configure a non-TEDS installation from a PC.
What is the digital filter time constant for?
Steadying the reading, and it goes much further than most people expect — from eighty milliseconds up to 9.6 seconds. That is the slowest deliberate filtering in our range and it is appropriate here, because a transmitter feeding a control loop wants a number that does not jitter far more than it wants speed. A jittery input makes a loop hunt, an alarm chatter and a trend chart unreadable. Agitation, sloshing, a conveyor running past, a pump cycling: none of that is the measurement, and a long constant removes all of it. Set the shortest constant that gives you a steady number, and expect that to be longer than instinct suggests.
What can the relays switch?
Two solid state relays, single pole normally open, rated at 120 milliamps at 140 volts AC or 180 volts DC, with 28 ohms of series resistance. That combination — high voltage, modest current — suits an indicator lamp, a beacon, a small solenoid or a logic input on another controller. It is a signaling output rather than a power output, so anything drawing real current needs an interposing relay or contactor driven from it. Being solid state, there are no contacts to wear, which matters where a threshold is crossed on every cycle of a process.
How is the transmitter mounted and powered?
It clips onto standard 35 millimeter DIN rail and measures about 120 by 101 by 22.5 millimeters, so along the rail it occupies roughly the width of a couple of terminal blocks. All connections are detachable screw plugs, which means the wiring can be unplugged as a block if the module ever has to come out. The standard version runs from a wide range of AC line voltage and will also accept high voltage DC; a low voltage version is available for panels built around a 24 volt control bus. Operating temperature runs from freezing to seventy degrees Celsius, which is generous for a cabinet-mounted device.
Questions From The Field
The transmitter and my portable indicator disagree on the same sensor.
Check that both are configured for the same sensor before suspecting either. The commonest cause is two instruments carrying two different scale factors for one transducer, which produces two confident and incompatible answers with nothing visibly wrong. Confirm the excitation voltage each is supplying, since output is proportional to it and a sensor read at five volts by one device and ten by another will genuinely differ. Then confirm the input range and scaling on each. If both are TEDS-configured from the same sensor and they still disagree, that is worth reporting to us with both model numbers.
My analog output will not reach 20 mA.
Add up the loop. A current output can only push its full current through so much resistance, and this one is specified for up to 500 ohms: that is the cable there and back, plus the receiving device's input resistance, plus anything else in series such as an isolator, a display or a barrier. A bench test with a short lead and a meter is a much easier load than a long run into a panel with two devices in the loop. Also confirm that only one device is powering the loop, since a transmitter driving into a supply that is also driving it is a common and confusing fault. If the loop is simple and it still falls short, come back to us with the arrangement.
Do I still need a display somewhere?
Only if a person has to read the number where the sensor is. Ask who looks at it and when. If the measurement is consumed by a PLC and watched on a SCADA screen, a display at the machine is a hole in a cabinet door and something to keep clean. If a technician commissioning or troubleshooting the machine needs to see the weight while standing at it, then either feed a display from the analog output or use a panel meter instead of a transmitter. The awkward middle case is a machine that is normally unattended but occasionally worked on, and there the usual answer is a portable indicator carried to the sensor rather than a display fitted permanently.
My PLC and my sensor are grounded in different places. Is that a problem?
It is precisely the problem this transmitter is built for, and with the isolation it provides it should not affect your measurement. Because the signal input and the analog output are separately isolated, the sensor's ground and the PLC's ground are not connected through the transmitter, so a potential difference between them has no shared path to drive current along. What isolation cannot fix is a ground loop you create elsewhere — a sensor cable shield grounded at both ends, for instance, gives the difference somewhere else to flow. Ground the shield at one end only, keep the sensor cable away from motor and drive wiring, and let the isolation do the rest.
My control loop hunts and the reading is jumpy.
Lengthen the filter time constant before touching the loop tuning. A jittery process variable makes a controller chase noise, and the usual instinct — detuning the loop — treats the symptom while leaving the cause. Step the constant up and watch what happens to both the reading and the loop; on a slow process such as a tank filling you can often go to several seconds with no practical loss of response. If the jitter survives even a long constant, it is probably not noise but real movement in the process or the structure, and that is a mechanical investigation rather than an instrument one.
Can I mount this on the machine rather than in a cabinet?
Only inside a suitable enclosure. The module is designed for DIN rail in a protected environment and carries no ingress rating of its own, so whatever surrounds it provides the protection against dust, moisture and mechanical damage. Mounting it near the sensor in a small rail enclosure is a perfectly good arrangement and has a real advantage — it shortens the low level sensor cable and sends a robust 4–20 mA or serial signal the long distance instead, which is exactly the right way round. Watch the temperature inside a small sealed box in a warm location, and tell us the environment if you are unsure.
My sensor has no TEDS memory. How much extra work is that?
A one-time setup rather than an ongoing burden. Connect a PC to the serial port, run the free instrument setup software, and enter the sensor's details from its calibration certificate. Everything the transmitter does afterwards is identical. What changes is where responsibility sits: with TEDS the sensor carries its own configuration wherever it goes, and without it the configuration lives in the transmitter and has to be kept correct by whoever maintains the installation. If you have several transmitters to set up the same way, doing it from a PC is quicker than any other route, so the software is worth using even where TEDS is available.
Can one transmitter handle two sensors?
No — it is a single channel device, and that is deliberate. The usual arrangement for several sensors is one transmitter each, which has real advantages: the channels are independent, each can be scaled for its own sensor, a fault in one does not affect the others, and on an RS485 line they can all be polled by the same supervisory system. If what you actually need is several load cells summed into one total, that is a different requirement with a different answer, and it is worth telling us which of the two you have before ordering anything.