A multi-channel load cell amplifier is a chassis that conditions several sensors at once, each on its own card, with a meter and a channel selector on the front so any one of them can be looked at individually. The TIO-3000 has five slots, and what goes in them is up to the job.
It exists because of a problem that sounds trivial and is not: adding up several load cells.
Put a vessel on four mounts and each mount carries part of the weight. Total the four and you have the contents. That is the theory, and it fails immediately in practice for a reason worth understanding.
The four load cells are not identical. Manufacturing variation means each has a slightly different sensitivity, so a pound resting over one corner does not produce the same signal as a pound resting over another. Move the load around the platform and the indicated weight changes even though nothing was added or removed. Rice Lake's guidance on the subject states the objective plainly: trimming exists to ensure the scale weighs correctly regardless of where the load is applied. It is made worse by the fact that weight distribution between the cells is rarely equal anyway — an agitator, a discharge chute, a pipe connection or a powder that mounds to one side all load some corners harder than others.
The industry's usual answer is a summing junction box, and it works by throwing signal away. Cells are wired in parallel in a box, their outputs tied together, and one summed signal goes on to the indicator. The mismatch is then trimmed out with potentiometers, of which there are two kinds and both subtract. Signal trim puts resistors across each cell's output so some of its signal leaks away before it reaches the summing point. Excitation trim puts resistance in series with the supply so the stronger cells receive less excitation and therefore produce less output. Either way you are correcting by making the good channels worse until they match the weakest one.
The TIO-3000 does the opposite, and that is the whole idea. Every cell gets its own amplifier card with its own adjustable gain. Matching is done by turning a channel up rather than turning the others down, so nothing is discarded. The summing happens afterwards, on signals that are now volts instead of millivolts, in a precision module built for the job rather than in a junction box in a damp corner of the plant.
And there is a second advantage that matters more over the life of the installation. A summing junction box produces one number. When that number starts misbehaving — drifting, jumping, refusing to return to zero — it tells you nothing about which cell is responsible, and finding out means going out to the vessel with a meter and disconnecting things. With a channel per cell and a selector on the front, you turn a knob. A cell that is drifting, a mount that has begun to bind, a corner that is being propped by a pipe someone added, a cable that is failing — each of those looks like a fault in one channel and normal behaviour in the other three. The diagnosis is a five second job at the front panel instead of an afternoon in the plant.
The chassis was originally built for precision summing on two, three or four load cell signals for weighing, and has grown into a modular instrument that also handles peak capture, alarm and setpoint work, and current loop outputs, on weight, force, pressure, strain and deflection signals.
Five slots, and you decide what goes in them. Understanding the modules is understanding the instrument, because the chassis is essentially a power supply, a meter and a way of connecting cards to each other.
The amplifier and conditioner card is where a sensor lands. It supplies a constant 8 volts of excitation to the bridge, takes the millivolt signal back and amplifies it to a usable zero to plus or minus ten volts, with gain adjustable anywhere from seventy-five to a thousand. Work that range out and it corresponds to full scale inputs from about ten millivolts up to a hundred and thirty — which comfortably covers the sensors you are likely to own. A 2 mV/V load cell at eight volts of excitation produces sixteen millivolts at capacity, landing near the middle of the adjustment range with room either side. Nonlinearity through the card is a twentieth of one percent of full scale. There is a version that produces a 4–20 milliamp current loop instead of a voltage, for feeding a controller directly.
Each channel carries its own excitation supply, which is a bigger deal than it sounds. Every amplifier card can deliver up to a hundred milliamps at eight volts, and it delivers it to one sensor. A 350 ohm bridge draws about twenty-three milliamps of that; even a 120 ohm bridge draws under seventy. So the specified range of 120 to 1000 ohms is covered with margin on every channel, and — more importantly — the cells are not competing for one supply. In a parallel arrangement every cell shares the same excitation, so anything that loads it affects all of them together. Here they are independent, which is one less way for a fault in one corner to contaminate the other three.
The summing card adds up to four conditioned signals to within a tenth of a percent of the inputs. It holds that to five hundredths of a percent over a day and drifts a hundredth of a percent per degree Celsius. Those are the numbers that matter for a vessel that has to read the same on Friday afternoon as it did on Monday morning, and they are achievable precisely because the summing is being done on volt-level signals rather than on the raw output of a bridge.
The bandwidth is deliberately narrow, and for weighing that is correct. The amplifier card responds from DC to sixteen hertz. On a vessel, everything above that is a nuisance rather than information — an agitator, a pump, a conveyor running nearby, product sloshing, a forklift going past. Rolling off there removes all of it before it reaches the display, which is why a properly conditioned weighing signal sits still while a wideband one dances.
The peak card is faster than the amplifier in front of it, and it is worth knowing which one sets the limit. Peak hold captures in under half a millisecond and holds what it caught, drooping less than two millivolts a minute, with polarity selection and a remote reset. But a peak detector can only capture what reaches it, so if it is fed from a sixteen hertz amplifier the system's real response is sixteen hertz, not half a millisecond. For slowly building loads — a proof test, a vessel filling, a gradual pull — that is fine. If you need to catch something genuinely fast, that is a conversation about the amplifier channel rather than about the peak card, and worth having with us before you specify.
The alarm card is a window rather than a pair of switches. It has two settings and produces three states: below the window, inside it, and above it. That is a go / no-go answer in a single module, which is a different thing from two independent setpoints and a much better fit for checking that a part, a fill or a force is within tolerance rather than merely above or below a line. Its relays are rated modestly, at ten watts and half an amp, but they are specified for a hundred million operations — a card intended to switch on every cycle of a production process for years.
Configuring one of these is a matter of listing the signals and listing the jobs, then filling five slots. Our applications engineers do this regularly and would rather draw it out with you than have you guess.
Count your sensors first, then count what has to happen to them. Four load cells summed into one weight uses four amplifier cards and a summing card, and that fills the chassis exactly — which is not a coincidence, since that is the arrangement it was designed around. Three cells and a sum leaves a slot for an alarm or a peak card. Two cells leaves two. Work out the arithmetic before ordering, because slots are the constraint that bites.
Decide what form the answer has to take. A voltage output suits a data acquisition system or a recorder on a bench. A 4–20 milliamp loop suits a plant controller and survives distance better, because the live zero at the bottom of its range means a dead loop is distinguishable from a genuine reading of nothing — a distinction a voltage output cannot make. Both the amplifier and the summing cards come in either form, so decide per channel rather than for the whole chassis.
Choose the meter resolution to match what you are doing. A three and a half digit display is ample for monitoring and for setting up. The four and a half digit version, reading to nearly twenty thousand counts, earns its place where the readout is the measurement rather than a check on it. It is a chassis-level choice made at order rather than a card, so it is worth deciding deliberately.
Be realistic about speed. The narrow bandwidth that makes this excellent for weighing makes it unsuitable for capturing fast transients, and that is a design choice rather than a shortcoming. Tell us how quickly the thing you are measuring actually changes and we will tell you honestly whether this is the right architecture. Getting that wrong is the one specification error on this product that cannot be fixed by swapping a card.
Check your bridge resistance and your wiring. Sensors from 120 to 1000 ohms are supported, which covers essentially everything we make and most of what we do not. The amplifier cards expect six-wire shielded connection to the sensor, so if you are reusing an existing installation, find out what is actually in the conduit before you plan the commissioning.
Confirm the supply. The chassis runs from 115 volts at 60 hertz. If your facility or your destination country runs on something else, raise it with us at the enquiry stage rather than discovering it on a loading dock.
Then think about the plant, not the instrument. The best summing system in the world cannot rescue a vessel that is being propped by a rigid pipe, restrained by a conduit, or standing on mounts that bind when it expands. Those faults are real load paths and every instrument will faithfully report them. What this one adds is the ability to see which corner is doing it.
Send us the number of sensors, their resistance and sensitivity, what the output has to feed and how fast the load changes, and we will put together the card list. Tell us if this is replacing a summing junction box, because that usually changes the answer.
TIO-3000 Multi-Channel Load Cell Amplifier Applications.
The Transducer Techniques TIO-3000 is a versatile five channel industrial amplifier and conditioner instrument, originally created for precision summing on two, three or four load cell signals for weighing applications, and now a modular system with five plug-in slots, a built-in meter and a channel selector.
- Tank and Vessel Weighing: Four load cells conditioned individually and summed precisely, with each mount still visible on its own channel for checking and diagnosis.
- Platform and Floor Scales: Corner matching performed by adjusting channel gain rather than by trimming signal away, so no output is discarded to achieve it.
- Hopper and Silo Contents: Narrow bandwidth conditioning rejects agitator, conveyor and material movement so a level reading settles rather than dancing.
- Multi-Point Force Measurement: Several sensors on one structure read separately and together, for load distribution studies as well as totals.
- Test Stands and Laboratories: Up to five conditioned channels of weight, force, pressure, strain or deflection from a single bench instrument.
- Batch and Fill Control: Window alarm modules provide a three state low, in-tolerance and high output to drive valves, feeders and indicators.
- Go / No-Go Inspection: A single window alarm module answers whether a force or weight is inside tolerance, rather than merely above or below a limit.
- Proof and Overload Testing: Peak hold with polarity selection and remote reset records the maximum reached during a load application and holds it for recording.
- Process Signal Conversion: Current loop amplifier and summing modules deliver 4–20 mA directly to a plant controller without an intermediate converter.
- Summing Box Replacement: An upgrade path for installations where a junction box gives one number and no way of telling which load cell is responsible for a problem.
The TIO-3000 accepts weight, force, pressure, strain and deflection signals from bridge sensors of 120 to 1000 ohms, and works with our full range of load cells, torque sensors and pressure transducers.
Frequently Asked Questions
What is the TIO-3000, and what is actually in the chassis?
A five channel industrial amplifier and conditioner instrument. The chassis provides power, five plug-in slots, a built-in meter and a channel selector that lets you display any one channel on demand; the cards you fit decide what it does. It was created for precision summing on two, three or four load cell signals for weighing, and has grown into a modular system covering amplification, summing, peak capture and alarm functions, on weight, force, pressure, strain and deflection signals. Sensors connect through a 9-pin D connector and data and control travel through a 25-pin D.
What modules are available and what does each one do?
Four functions, some in two output forms. The amplifier and conditioner card excites a bridge sensor and amplifies its output to a voltage, with a current loop version available. The precision summing card adds two to four conditioned signals into one total, also available as a current loop version. The peak hold card captures and holds a maximum, with polarity selection and remote reset. The window alarm card compares the signal against two thresholds and reports which of three zones it is in, with relay outputs. You mix them across the five slots to suit the job.
Why can I not simply wire my load cells in parallel and read the total?
You can, and a great many systems do, but the result is only correct if the cells are identical and they are not. Manufacturing variation means each has a slightly different sensitivity, so the same weight produces a slightly different signal depending which corner it sits over — and the indicated total therefore changes as the load moves around the platform, with nothing added or removed. That is the problem trimming exists to solve. It is compounded by the fact that the weight rarely distributes evenly in the first place: an agitator, a discharge point, a pipe connection or product mounding to one side all load some mounts harder than others.
How is this different from a summing junction box?
Two differences and both matter. First, a junction box corrects mismatch by subtraction — signal trim leaks output away through resistors across each cell, excitation trim lowers the supply to the stronger cells — so the good channels are degraded until they match the weakest. Here each cell has its own amplifier with its own gain, so matching is done by turning a channel up, and nothing is discarded. Second, a junction box produces one signal, which means when something goes wrong you have one number and no way of telling which cell caused it. With a channel per cell and a selector on the front, you look at each one in turn.
What does the amplifier card do, and what gain range does it cover?
It supplies excitation to the bridge and raises the millivolt signal that comes back to a usable zero to plus or minus ten volts, with gain adjustable from seventy-five to a thousand and nonlinearity of 0.05 percent of full scale. That gain range corresponds to full scale inputs of roughly ten to a hundred and thirty millivolts, so it accommodates most sensors comfortably — a 2 mV/V load cell at the card's eight volt excitation produces sixteen millivolts at capacity and sits near the middle of the adjustment. Frequency response runs from DC to sixteen hertz, which is a deliberate choice discussed below.
How much excitation does each channel supply?
A constant eight volts, up to a hundred milliamps, per card. Because each sensor has its own card it also has its own supply rather than sharing one, which is worth more than the numbers suggest. A 350 ohm bridge draws around twenty-three milliamps and even a 120 ohm bridge stays under seventy, so the full 120 to 1000 ohm range is supported with margin. More usefully, the channels are independent: in a parallel arrangement every cell shares one excitation supply, so anything that loads or disturbs it moves all the readings together. Here a problem in one corner stays in one corner, which is exactly what you want when you are trying to find it.
How accurate is the summing?
Within a tenth of one percent of the inputs, holding to five hundredths of a percent over twenty-four hours and drifting a hundredth of a percent per degree Celsius. Those stability figures are the ones to look at for a vessel that must read consistently across a shift or a week, and they are achievable because the addition is being done on conditioned volt-level signals rather than on raw bridge outputs. Note what the specification is measuring: the accuracy of the addition itself. The accuracy of the total also depends on the load cells, the mounts and how well the channels were matched at commissioning.
What is a window alarm and how is it different from two setpoints?
Two thresholds that define a band, with three output states rather than two: below the band, within it, and above it. Two independent setpoints tell you whether a value has passed each of two lines. A window tells you directly whether the measurement is in tolerance, which is a different and usually more useful question. For checking that a fill, a part or an applied force is within specification, that is a complete go / no-go answer from one module. The relays are rated at ten watts and half an amp, which suits driving indicators, logic inputs or an interposing relay, and they are specified for a hundred million operations.
What does the peak module capture, and how fast?
It captures a maximum in under half a millisecond and holds it, drooping less than two millivolts a minute, with polarity selection so it can follow either direction and a remote reset input so it can be cleared without anyone standing at the instrument. The important caveat is that a peak detector can only catch what arrives at it, so if the channel feeding it rolls off at sixteen hertz, that is the system's real limit rather than the module's half millisecond. For loads building over seconds — a proof test, a gradual pull, a vessel filling — it works exactly as intended. For genuinely fast events, talk to us about the channel in front of it.
What are the chassis' size and power requirements?
Twelve inches wide, seven and a half high and eight deep, running from 115 volts AC at 60 hertz through a 250 milliamp fuse. That is a bench or shelf instrument rather than a panel-mounted one, which suits where it is normally used — a control room, a laboratory, a test cell or an enclosure near the vessel, somewhere a person can reach the channel selector. If you need it to live in a panel door or run from a different supply, raise it with us early because those are not configuration options.
Questions From The Field
My vessel reads differently depending on where the load sits in it.
Classic corner mismatch, and it is the reason this instrument exists. The load cells have slightly different sensitivities, so the total shifts as weight moves between them. The fix is to match the channels: apply a known weight over each mount in turn, note what each channel produces, and adjust the gains until they agree. Because you are raising the weaker channels rather than trimming the stronger ones down, you keep the full output of every cell. Do it after the vessel is fully installed and connected, not before, since piping and restraints change how load reaches the mounts.
Something is wrong with one corner but I cannot tell which.
Switch through the channels on the front panel and compare them. This is the diagnostic that a summed installation cannot give you, and most faults declare themselves immediately: a cell that drifts while the others are steady, a channel that will not return to zero after a load is removed, one corner that barely changes when the vessel is filled, a reading that jumps when someone walks past a particular mount. Note what each channel reads empty and again with a known load, and compare against what you recorded at commissioning — if you did not record it then, record it now, because the comparison is worth far more than any single reading.
The reading wanders and I cannot tell whether it is the vessel or the electronics.
Look at the channels individually, because the pattern tells you which. If every channel wanders together, suspect something common — the supply, grounding, or genuine movement of the whole vessel. If one channel wanders and the rest are quiet, the problem is that corner: its cell, its cable, its mount, or something touching the vessel near it. If the wander correlates with a machine starting somewhere, remember the conditioning already rolls off above sixteen hertz, so anything surviving that is either slow or is arriving as a real force. A pipe that transmits pump vibration into a vessel is a real load, and the instrument is right to report it.
Can I use this to catch the peak of a fast impact?
Probably not, and the reason is worth understanding. The peak module itself is quick, but the amplifier channel feeding it is deliberately band-limited for weighing, so a genuinely fast event is filtered before the peak detector ever sees it and what gets captured will understate the true maximum. That band-limiting is the same feature that makes the instrument steady on a vibrating plant, so it is not something to regret — it is the trade. Tell us how long your event actually lasts, in milliseconds, and we will tell you plainly whether this architecture suits it or point you at one that does.
I only have two load cells. Is this too much instrument?
Not necessarily, and the question to ask is what else you want it to do. Two cells and a summing card use three slots, leaving two for a peak module, an alarm module or a spare channel for a different sensor entirely. If all you need is two signals added together and displayed, there are simpler arrangements and we will say so. If you want the total, the two corners visible separately, a window alarm on the result and room to add something later, then the chassis is doing five jobs rather than one and the arithmetic looks quite different.
Our facility runs on 230 volts at 50 hertz.
Raise it with us before ordering rather than after. The chassis as catalogued runs from 115 volts at 60 hertz, so a different supply is a conversation to have at the enquiry stage. There are usually ways to address it, and all of them are easier to arrange before the instrument ships than after it has arrived on the wrong side of an ocean. The same applies if the instrument is being specified for an export project by someone working in a 115 volt country.
Can I add a module later, or take the summed output somewhere else?
The chassis takes plug-in cards in five slots, so adding a function is a matter of a free slot and the right card — check with us with your configuration to hand and we will confirm what fits alongside what you already have. Getting the output elsewhere is straightforward: the amplifier and summing cards produce voltage or 4–20 milliamp signals through the 25-pin data and control connector, so a data acquisition system, a recorder, a controller or a larger display can all take the result. If you know now that something upstream will eventually want the data, it is worth saying so while the configuration is being decided.
My load cells are 120 ohm, not 350. Does that matter?
Not to this instrument. The amplifier cards accept bridge resistances from 120 to 1000 ohms, and the excitation supply has the current to drive the whole of that range on every channel — a 120 ohm bridge at eight volts draws under seventy milliamps against a hundred available. What lower resistance does affect is your cabling, since more current in the excitation leads means more voltage lost along them, so a low resistance bridge on a long run wants heavier conductors. Tell us the resistance, the sensitivity and the cable length together and we will confirm the arrangement works before you commit to it.