LM293DR
Product Specifications
LM293DR
Texas Instruments — Power Management & Signal Chain
Product Overview
The LM293DR from Texas Instruments is a power management / signal chain IC delivering high efficiency, precision regulation, and robust protection. Texas Instruments (TI) is a global leader in analog ICs, power management solutions, and embedded processors.
Superb Automation partners with leading brands including Qorvo, TI, ADI, XILINX, NVIDIA, Intel, ST, MACOM, Renesas, and Coilcraft to provide complete BOM and supply-chain solutions.
Key Features
- High-efficiency power conversion with low quiescent current
- Wide input voltage range with precision output regulation
- Integrated over-current, over-temperature, and short-circuit protection
- Small footprint packaging for space-constrained designs
- AEC-Q100 qualified options available for automotive applications
- Flexible sequencing and soft-start control
Key Specifications
| Input Voltage | V |
| Output Voltage | V |
| Output Current | A |
| Switching Frequency | kHz |
| Efficiency | % |
| Quiescent Current | µA |
| Topology | — |
| Package | — |
| Operating Temp | °C |
| Protection | — |
Applications
Request Quote — In Stock, 100% New & Original, Full Warranty
LM293DR: the plain dual comparator on a reel, five millivolts at room
LM293DR is the plain grade of the dual differential comparator in the eight pin small outline body, on a large tape and reel, guaranteed from minus 25 to plus 85 degrees. The top side marking is LM293, the reply time is 1.3 microseconds, the supply range is 2 to 36 volts, and the whole part costs less than the labour of placing it.
This is the code that ends up on a bill of materials when a threshold does not need the A sort and the build is large enough to want a reel. The engineering question is which of the two sorts the threshold actually needs, and the answer is usually decided by two rows: the offset window over temperature and the bias current into the source impedance.
The string and the grade
LM is the family, 293 is the dual comparator with the minus 25 to plus 85 window, D is the eight pin small outline body, and R is the reel. There is no A in the string and no G4, so this is the plain sort under the older material declaration, which the addendum still prints as active.
| Orderable code | LM293DR |
| Status on the current addendum | active |
| Comparators per body | 2, independent |
| Body | SOIC, 8 pins |
| Carrier | large tape and reel |
| Part marking | LM293 |
| Offset at 25 degrees | 2 millivolts typical, 5 maximum |
| Offset over the window | 9 millivolts maximum |
| Input bias current at 25 degrees | 25 nanoamps typical, 250 maximum |
| Input bias current over the window | 400 nanoamps maximum |
| Common mode range | ground to 1.5 volts below the positive rail |
| Supply current at 5 volts | 0.45 milliamps typical |
| Response time | 1.3 microseconds typical |
| Temperature window | minus 25 to plus 85 degrees |
Deciding between the plain sort and the A sort
The offset window is 5 millivolts at room and 9 over the temperature range, against 2 and 4 on the A sort. On a 5 volt rail that is 0.18 percent of full scale at room and 0.27 percent at the extremes. Whether that matters depends on what is being compared.
If the threshold is a fault trip with a wide margin, 9 millivolts is noise in the budget. If the threshold is a cell voltage on a battery stack, where the difference between a healthy cell and a weak one can be 50 millivolts, the plain sort still works but the calibration table has to be written against 9 rather than 4. If two comparators form a window, the two offsets add and the worst case is 18 millivolts over temperature, which is where a plain grade circuit starts to need a trim.
Bias current is the second term and usually the larger one. At 250 nanoamps maximum and a 100 kilo ohm source the error is 25 millivolts, which is three times the offset window and twenty five times the difference between the two sorts. A circuit that spends money on the A sort and keeps a high impedance divider has optimized the smaller error.
The output and the pull up
The output is an open collector rated to 36 volts with a 16 milliamp sink at 1.5 volts. Saturation is 130 millivolts typical and 400 maximum at 4 milliamps, rising to 700 over the window. Two outputs can be tied together to make a wired AND, which is the standard way to combine two fault signals without adding a gate.
Pull up sizing sets both the edge and the dissipation. With 10 kilo ohms on 5 volts the sink current is half a milliamp, the low level is well inside the 400 millivolt limit and the dissipation is 2.5 milliwatts. With 1 kilo ohm the current is 5 milliamps, the edge into a capacitive load is sharper and the dissipation is 25 milliwatts, which is still small for the package but not for a 125 degree ambient.
Leakage when the output is off is 50 nanoamps at room and 1 microamp over the window. On a 100 kilo ohm pull up that is 100 millivolts of offset at the hot end, and the next stage sees it as a threshold shift rather than as a leakage figure.
Where the cost saving is real and where it is not
The plain sort is cheaper than the A sort, and the difference is small in absolute terms. The saving is real on a high volume board where the threshold margin is comfortable. It is not real on a board where the change forces an extra trim step, because the labour of a trim is larger than the price gap on a reel.
Our how to reduce component cost page sets out where a bill of materials saving usually comes from, and the pattern holds here: consolidation and carrier choice save more than grade choice on parts at this price. The BOM consolidation page covers the case where a board carries both sorts of the same comparator and one line can go.
The carrier decision is the other lever. A 2500 piece reel on an eight pin body is a common carrier, so the feeder setup is shared and the reel does not cost machine time. A build that only needs a few hundred pieces is better served by cut tape, and a prototype by a tube, and choosing the wrong carrier costs more than the grade difference.
Where the plain sort is the right answer
Three kinds of circuit take the plain sort without a second thought. A fault trip with a wide margin, a supply sequencer with a resistor divider and a coarse level detector on a 12 or 24 volt rail. In each of them the threshold margin is hundreds of millivolts and a 9 millivolt offset window is background noise.
Two kinds need the A sort. A window detector, where the two offsets add, and a sensing node with a source impedance above ten kilo ohms, where the bias current dominates and the tighter offset helps less than a lower impedance divider would. The second case is where a change of the divider values beats an upgrade of the grade, and it is almost always cheaper.
Two other rows worth comparing
The first is the newer B version of the same function, which carries a lower offset, a wider supply range and a faster response in the same eight pin body. It is not a drop in substitute on every drawing, because the input stage changed, but it is the row to price before committing to a long life build of the older part.
The second is the single comparator in the same family, which gives one channel in five pins and a shorter signal path on a dense board. Four comparators in one body save placements, one comparator in a small body saves routing, and the choice is usually made by the board rather than by the price.
A third row is worth a look on a board that carries several thresholds. The quad comparator in the same fourteen pin outline puts four channels in the space of one dual, and on a board with four separate level detectors the package count drops by two. That saving is visible on the placement count and invisible on the price of a single part, which is why it gets missed in a bill of materials review.
Questions that come with this code
Is the A sort worth it here? Only when the threshold budget is tight or the window edges matter. The offset difference is 2 and 4 millivolts against 5 and 9, and the bias current is the same on both.
Can this part replace LM293D? Yes, the die and electrical table are identical and the difference is the carrier, from a tube to a reel.
Does the input work at the positive rail? No, it stops 1.5 volts below at room temperature and 2 volts below over the window, so a threshold near the rail needs a divider or a different part.
How fast does the output fall into a 5 volt pull up? The response time is 1.3 microseconds typical with 5 millivolts of overdrive, and the fall is faster than the rise because the output sinks current rather than sourcing it.