Is this a hit piece by a high-definition display manufacturer? It seems like an unnecessary litany of detractors against one technology without a lot of alternative solutions or trade-offs with others. At the end, there are only 3 sentences that could be called alternatives with the heading "use high resolution displays". Making a "helpful" paper that only lists problems without comparisons or solutions misses all the reasons one would use a 7-segment display (mentioned in the comments here so I won't repeat). There's nothing wrong with this technology. The title "don't use" is too inclusive and is not supported by the paper. I think "Consider not using in certain applications" would have been more appropriate based on the paper's body.
> Is this a hit piece by a high-definition display manufacturer?
This was the exact thought I had. The paper strikes me as being very odd, and the arguments against using 7-segment LEDs are very thin.
I will agree with one thing, though: the right components to use are an engineering decision and as with all engineering decisions, the right choice (i.e., the right set of tradeoffs) needs to be determined on a project-by-project basis. For a given project, that may mean not using a 7-segment LED. Or it might mean that's exactly the right choice.
Considering the paper does not specify that it has no conflicts of interest in the usual place where it would, we should assume the author is involved with such manufacturers
As someone designing electrical circuits that might or might not use a 7-digit display this seems rather to be a guide for which use cases to avoid them.
Electrical engineers sadly sometimes don't think about that at all and slap what they know onto everything. When you have a medical device where a reading error or ambiguity can have serious consequences you might want to be aware of that. There are 7-segment-display usecases where that is not an issue irrelevant, e.g. because you have enough contextual clues to not hold the device wrong and read it upside down or because the potential damage from a reading error is irrelevant.
But this is still a design consideration one should be aware of. Especially in times where 7-segment displays aren't necessarily the cheapest option.
Most other display technologies have their downsides too. For example, fot-matrix LCDs are far more fragile and more easily damaged by sunlight or heat than 7-segment LEDs. If you don't discuss that, but dwell on the downsides of 7-segment displays, it does feel a bit like a hit piece.
In fact, 7-segment displays aren't even a monolithic technology, and not all of the gripes in the article apply to every implementation. You have traditional LED models, LCD 7-segment displays, and then some OLED flavors, VFDs, and even electromechanical designs in niche applications.
Almost any 7 segmented display could be replaced by an analogue gauge. For time, for temperature, for selecting options. When a device has a segmented display, I assume it is to cut costs.
When a device has many analog gauges, I assume it is an unnecessarily expensive device.
Few modern use cases actually benefit from the qualities of an analog gauge. It seems to be more of a status marker, at least for consumer goods. Military and ruggedized applications are a different story.
What would be an application where a segmented display is just as good or better than an analogue gauge? The only one I can think of is a bedside clock or other device that will be used in the dark and can benefit from self-illuminance.
Especially household appliances are worsened by the switch from analogue to digital. A microwave oven used to be so practical to operate. Just dial the clock to your desired time and that's it. Now, you have to press +30s button a bunch of times, or type your time. They're even trying to put digital on stoves, clearly not intended for use by people who actually cook. Induction stoves all have this problem. My air fryer is much worse to operate than it should, because they insisted on digital. It would be perfectly fine with a gauge for time and one for temperature. Now there's all these buttons and a segmented display.
I have a segmented display thermometer. It's not better than an analogue in any way I know of, but they're cheaper. I have a segmented display speedometer on my motorcycle. It is of course getting its input from an analogue measurer, so why transfer it to digital?
Modern speedometers are digital. A magnet or metal disk with cutouts rotating with the transmission produces pulses in a magnetic field sensor, the pulses are counted by a microchip. The data is sent to the dashboard over a packet-based computer network.
I for one find a "71" in large digits much easier to read than trying to figure out where between the lines a little needle is pointing. If the opposite were really true, we'd draw little pictures of gauges to communicate numbers.
In cars I see more analog than segmented displays for the speedometer, but I'm sure you're right that the underlying technology can be digital. For MCs, I kind of agree with you that numbers are nicer.
The current day digital sensors are a lot more reliable. I have multiple cheap digital thermometer etc gadgets and they read the same temperature to within 0.1 C, humidity to ~0.5%, and so on.
The source is analog only in the sense that physics is "analog" until you get to Planck scale or to quantum phenomena.
I also have some digital thermometers, and personally don't have any preference digital vs analog in them. For home use, +/- one degree or percentage of humidity don't really make a difference. As for reliability I don't think you can beat analog, which will last for decades without changing a battery.
I repair old timing equipment (think Race America, JACircuits, etc) as a side hobby, and a lot of it is large 7 segment displays with individual LEDs in series for each segment.
I am so glad people didn't try using an LCD display or similar, I'd never be able to fix them when the pixels/display panels go bad and the original manufacturer is long out of business. LEDs, shift registers, mosfets... all serviceable.
Users fixing things make the stock market sad. Better to force them to spend hundreds on a new thing then chuck the old one in the garbage even though it's 99% functioning.
Unfortunately some motorsport event organizers barely break even, especially post-covid in CA, although it is picking up slightly this year. It is hard for them to justify $2-4k for a new display or timing system, whereas it's a few hours of work for me to do most repairs.
LCD is doable, it would just take you more time and research to come up with stable cheap solution. LCD are driven by AC at rather pedestrian frequencies, nothing an RP2040 (rp pico) couldnt handle.
Yeah you could do it. I'm sure I will at some point be replacing the guts in some devices with STM32s or RP2040 as you said. Or beef it up with ESP32 :)
but if the display itself is damaged, that's tough, pretty much just reuse the case at that point I bet.
I meant use rp2040 to tap into raw signals meant for glass LCD in devices using microcontrollers with integrated LCD driver. Interpret and redraw into new framebuffer, spew out to modern LCD/OLED/VGA/hdmi.
Replace old custom unobtanium cracked LCD with modern LCD/OLED.
oh yeah, that's also a good idea! If they're using the software I sell to organizers it supports regular old computer display out, too, so you could just replace the display board and use HDMI directly these days.
The publication date is unclear, but the latest citation is 2011. It doesn't make economic sense anymore to use seven segment displays in new products, small TFT/LCD and even OLEDs are cheaper.
It probably does make sense to use them with legacy equipment that require a large certification effort (most of the examples are medical).
The advice the author gives for situations where seven segment displays are used is good, though.
What really doesn't make sense is using a seven segment font on a high resolution device to make the product seem more technological, I guess, but I don't think that was mentioned.
> It doesn't make economic sense anymore to use seven segment displays in new products, small TFT/LCD and even OLEDs are cheaper.
Maybe cheaper, but what about power consumption? I would have expected them to need significantly more, which might matter for battery-powered devices.
While this display is really slick, it seems like it is still not available under multiple suppliers? If thats the case. that means its going to be more pricey than other options. I cant seem to find many cheap options on alibaba/aliexpress.
Unlit liquid crystal 7-segement displays are the lowest-power & lowest-cost solution. You're absolutely right. The person you're replying to clearly doesn't have experience in the low-power/low-cost embedded world. In that space even 1KB of SRAM is a liability.
> Maybe cheaper, but what about power consumption? I would have expected them to need significantly more, which might matter for battery-powered devices.
Seven segment displays are quite power hungry relative to what information they can display. LCDs sip power, the issue is their backlight.
> I can read a 7-seg display from clear across a factory floor from almost any angle in the dark.
Ironically, this is true at lower power/brightness than higher. A few gas stations near me got newer super-bright displays and you can no longer make them out at a distance.
Which is why standalone digital clocks generally still use them instead of LCDs, and why I'd like to build a 7-segment clock at some point. That condition-agnostic readability is hard to beat.
The reason why most standalone digital clocks use LED seven segment displays is that to this day almost all of them are built around the same weird PMOS Sanyo ASIC from late 1970's.
Some equipment is placed directly under bright lights.
> And what is it that you need to read across the factory floor?
I don't understand the question. I mean, machines in a factory display something for the operator. If a machine doesn't need to display anything, why have a conversation about 7-seg LED vs LCD?
Legacy 7-segment displays use LEDs. Specifically, one LED per segment. The LCD version arose when the power consumption of the LED-based ones was an issue.
I'm not convinced that matters much in 2024, especially when this article singles out safety critical systems. Such systems are likely to be fairly expensive to being with, so the marginal cost of adding a display would be minuscule compared to testing and regulatory costs (not to mention the increase in fault tolerance versus a failed segment causing an incorrect reading).
A cracked LCD screen also makes the display unreadable. There are all sorts of ways for anything to fail, so calling out one failure method of one thing while ignoring similar failures for the device being compared comes across as very short sighted
The thing I hate about low cost LCD displays is that terrible motion blur when you have changing or scrolling text. It just screams cheap in my mind. 7 segment never had that problem.
Yes but a cracked LCD is much more obvious than a single LED segment burning out. Also usually physical damage is required to crack an LCD, where as LEDs can burn out (albeit rarely)
That’s only part of the equation. Grid displays require a lot more memory and often power to drive. By contrast, a seven segment display requires 7 (or 8) LEDs and 8 bits per digit on the output stage, 4 bit in BCD, which can be fed via a shift register.
Almost any other display that you will reasonably use in a application that makes sense for segmented multiplexed LEDs will integrate internal controller, so instead of having to refresh the thing somehow you have a thing that is simply an peripheral hanging of some or other bus that just does its thing. And well, such things tend to be cheaper than seven segment LEDs (there is a reason why random almost disposable crap has graphical monochrome OLED panels)
Granted, but as a sibling states, the other factor is firmware complexity. I can theoretically drive a high-resolution display on an 8-bit micro (although it may require with some clever bit banging to get I2C to work at all) but its trivial with a 7-segment display.
And as a bonus, you don't have to involve anymore UI/UX design than to answer the question "what number should go here?"
It'll be microcontroller equivalent of running full GNU/Linux to control an Internet of e-waste tea kettle, but CH32V003 + SSD1306 OLED + passives takes less than 10 parts total, no crystals needed, has plenty sample codes online, and costs ~$2 combined. Depending on the price of the final product, it could make financial or product design sense.
In a medical or aerospace application, that added firmware complexity could be a non-starter. Not from a price, or even e-waste standpoint. Its that proving the thing is safe is a lot harder when its complex. That's the real point of keeping it dumb.
Most LCD displays are custom: there are Chinese manufacturers that will design your LCD display and embed a SPI/I2C decoder for very low cost, so the simplest 8-bit PIC can drive one at 1 MHz (or less!)
If you look at distributor single-unit quantities, they are close in cost. If you look at hobbyist stores, they are the same price (or more expensive for the seven-segment version).
If you ask a factory in China to make you 1 million displays, the seven-segment version will be a lot cheaper. That's why new products use them. Even at 1k quantity from distributors, seven-segment displays are substantially cheaper.
Some Chinese display manufacturers are dumping lots of small, easy to use, postage stamp sized monochrome blue-white OLED at ~$1.5 apiece from 1 unit with free shipping, which had made it a viable replacement for 7-segments, couple power LEDs, even decorative logo displays, for small scale productions. It's not the case at all for real mass produced products, but not everyone is delivering thousands of units per month to Walmart.
I personally don't necessarily support "just go graphical" sentiment, but such local minima do exist.
I had considered that maybe new 7-segment displays aren't being made, so maybe they cost more for that reason, but even that doesn't seem like it'd be true.
Seems like many suppliers are coalescing on certain sizes/designs of OLEDs at the low end. All I see are these tiny blue/white displays or 1 inch full color. They all look the same and are boring at this point. I was hoping that we would get an explosion of different OLED display sizes and designs.
"... For Critical Applications" is the missing second part of the headline. Where critical applications are basically anything involving healthcare, piloting, or safety. I don't fully agree with all of the arguments presented in the article, but I feel like a lot of commentors here are going to miss that.
I don't fully agree either, but some of them make sense. But, some of them do make enough sense for some kind of critical applications, that it is worth to consider the alternatives (although power usage should also be considered).
As items (iii) (although it incorrectly says "(ii)") and (iv) of the caption of figure 1, are reasonable considerations.
There is also the consideration of faulty displays, in case a segment stops working (although you can mitigate this problem somewhat, by adding a test switch to light up all of the segments).
About decimals, one possibility (depending on the application) is to require a fixed position for the decimal point and add a wider gap between the digits at that position, which improves the clarity.
They also mention, "the user cannot see the difference between entering zero, or entering zero then decimal point". This is a problem on many hand-held calculators but it is easy to avoid, by not displaying the decimal point unless it has been entered explicitly (or if the result is being displayed).
Upsidedown (also mentioned in the article) is also a reasonable consideration, although there are ways to mitigate that too.
If anything, 7-segment displays are better when it comes to partial failure, as it's reasonably easy to add self-test functionality which measures the current passing through a supposed-to-be-lit / supposed-to-be-off segment.
This would allow for failure detection without the user having to explicitly press a test button and look for any misbehaving segments. You can't do this as easily (if at all) with LCD or OLED.
Usually when these displays initialize they turn on every segment before displaying data for that reason.
Also, fairly trivial to detect if a display has a segment out and throw a fault if needed. It's also not likely you burn out a whole segment at once if you have it wired up with that in mind.
> , obviously examination of cost and design trade-offs may lead to other decisions, and such decisions should be backed by competent empirical evaluation.
It’s almost like that is what electronic design means, creating a curcuit that works is perhaps the easiest part. Choosing all the parameters for the parts to be used and keeping it under the target cost is the hardest!
One very nice thing, dependent on use case of course, is that 7 segment displays are very simple to wire up and use. They can be easily used for analog circuits, without any digital micros at all if needed.
You can just grab all the pins to say 8 and wire them to vcc and it works. They are just led lights, arranged a particular way
You can even drive them with something like a 555 timer or other 74 series ICs if needed to make basic and very robust monitoring systems. They are very tough, can’t easily break or be destroyed by heat or cold
This is an example with a 555 and a cmos IC, but 74 series TTL IC could be used too. Personally more comfortable breadboarding TTL circuits myself but that’s just because I’m old
I have a love of electro-mechanical seven-segment displays. For scoreboards and timers, I feel like these are less intrusive than modern electronic display scoreboards.
There is something magical about seeing the score "flip" when a point is scored.
There's definitely a certain style with VFDs that I'm a little surprised hasn't been adopted by people looking for a retro/cyberpunk aesthetic. I was looking into them when it seemed my decades old alarm clock might need a replacement, and they seem quite rare which I assume the economies aren't there in making something more complex and delicate. Interestingly there's a few clocks repurposing cooker displays, I'd guess a manufacturer had a large stockpile of displays for models they discontinued early and liquidated them.
That's sadly all scientific papers. It makes a lot of sense on a printed page, and possibly on a vertical (portrait) monitor, but rarely on a 16:9 laptop screen or phone.
"Just use a high resolution display" doesn't seem like the corect remedy.
Next cheapest would be a low resolution LCD display, (30x80) that has its own set of issues. 7 segments at minimum provide for high visibility even if they fail at high readability.
note: price doesn't have to be money, complexity is expensive
Why isn't there a market for 14-segment displays or other simple alphanumeric displays?
edit: Childen comments use the phrase "for a few more $" a lot and i think they miss the point that if that was acceptable and easy we wouldn't be talking about it.
Agreed. "Just use a pixel matrix display" sounds an awful lot like, "Just use an Arduino."
There are plenty of places where a simple and small microcontroller paired with a 4-digit segment display is the cheapest, simplest, and most defensible design decision.
It appears we live in a universe where embedded systems aren't assumed to be a mission critical device by default.
I'm not going to bother using a microcontroller to solve a problem if isn't a use case where the solution needs to be as simple and robust as possible.
The market is flooded with 128x64 1.3" or 0.96" OLED displays. You can pick them up for anywhere from $1-$5, probably less if you bulk order.
Sure, they bring some complexity. Instead of just driving some pins you now need a display library and at least some bitmap font. But they provide a lot more fidelity, are vastly more readable, and fail in more obvious ways without needing periodic test patterns (the flashing 8's or snakes mentioned in the paper so you can detect failed segments)
For larger displays the choice becomes more interesting between OLED, LED matrices or 16 segment displays, but for small devices like those shown in the paper cheap OLEDs make the choice easy
The most common models of small OLED displays (SSD1306, etc) have a (write-only) framebuffer in the display controller, and can be written to incrementally. If you structure your visual interface with certain simple constraints in mind, you can update it without using any additional dynamic memory.
That would probably be 100x more complicated than the actual business logic on a small microcontroller and there are limits to program memory for the "streaming this string to the screen by constantly computing if a given pixel is on one at a time" algorithm.
edit addendum;
If it was using a 7-segment display, it is either really cheap or really simple "a graphical GUI on a screen" IS NOT A REASONABLE NEXT STEP.
I think we have radically different expectations of a simple embedded system.
RP2040 should cost $0.70 in some quantity (https://www.digikey.com/en/products/detail/raspberry-pi/SC09...) and in low-quantity it's $1 from places like CanaKit. That's actually shockingly cheap for an ARM Cortex M0; I was assuming your ATmegas and PIC18s with a similar pin count (40, 48, 64) would be a bit cheaper, but they seem to be around $2-4 in similar low-production quantity. Power efficiency won't be anywhere near as good with the ARM part, but if you're making a battery-powered device, you might be able to use the extra dollar or so to make it rechargable, then people care a bit less if it eats batteries.
Power consumption on the RP2040 is quite poor, actually; there's much better parts to use if you care about that.
If you just want a really cheap ARM microcontroller, Puya has some parts like the PY32F002A in the <$0.10 range. Single core, lower Fmax, and fewer peripherals than the RP2040, but the price is hard to argue with.
> That would probably be 100x more complicated than the actual business logic on a small microcontroller
It really isn't. I've implemented simple UIs on SSD1306 displays, and it's very easy to work with. The hardware is specifically set up to optimize for rows of 8px tall text (or multiples of that with a bit more effort). Once you have row/column addresses set up, each byte you write to the display writes one 8px column of pixels to a row and moves the column pointer over by a pixel. Printing text is a matter of streaming a sequence of bytes to the display.
Yes, I'd also like to note, the market is flooded with crap. Multiple devices I've owned with these "cheap" displays have been on crappy, battery-intensive devices. The devices with simpler segment (or dot matrix) displays last longer and work better.
I get some of the points where extreme legibility trumps (nearly) all other concerns, however I'd propose the issue could have been solved more simply with more specialized segment displays.
Dot-matrix LED display panels seem bulletproof, but I just checked the catalog for the 5x7 display driver I know and use and it costs $30 per. I guess the people who have reliability and legibility concerns like those expressed in the article aren't worried about spending $15 per digit.
Some subways have 111 segmented displays, I really like them, the refresh rate is stunning. Here is a page selling them [0] and here is how they look [1], [2].
I would say that if you're going to use a 7-segment display, it would be helpful to add a "test/diagnostic" button that does a pattern in which all segments are lit up in a predictable order, does a TOP pattern in which only the top segments are lit (to allow for device orientation), and then goes back to displaying exactly what was on the display prior to pressing the button.
Most of that is just due to a lack of contrast. Add a slightly-tinted window in front of it to reduce the contrast between unlit segment & background, and it becomes a lot easier to read.
I actually prefer 7-segment displays as a user over high-resolution alternatives. I love the 7-segment displays in my Casio watch, calculator, multimeter, lab PSU and until recently Canon 5D Mark IV (which had a 7-segment-ish display on the top). I switched to Nikon Z8, which unfortunately has a higher-resolution display, which I dislike the look of. It’s all overly smooth sans-serif[1]. I like the jaggedness of 7-segment displays. I’ve also never experienced any of the listed issues with them. Unfortunately, all mirrorless cameras have those high-resolution smooth ones now (if they have the top display at all). The one in Z8 is actually the least bad one I’ve seen.
[1]: On higher resolution I generally prefer serif fonts. In particular, for coding I use Go Mono, which is one of very few serif monospace fonts.
One thing 7-segment displays does better than LCDs and OLEDs is refresh rate. They can have a very high refresh rate, and this can help detect noise or variation in the signal.
I have two lab power supplies, one with a 7-segment display and one with an OLED. The 7-segment one allows me to easily see issues with the circuit due to the refresh rate, such as unwanted oscillations, which can be completely masked by the comparatively slow refresh rate of the OLED.
I won't necessarily see each value, but I'll notice a slight blurring of the least significant digit for example.
I suppose, but an analogue clock can be misread if held upside-down also, so what's the permitted degree of operator error? Depends on the consequence & domain I suppose.
Ultimately, this is trying to solve a human problem with tech, which is rarely a solution. If a display has garbage data on it because it's broken, the untrained person trying to read it is going to misread it regardless of the failure mode.
That and you can make 7-segment displays fault tolerant, ie, detecting when segments are out/partially out, because you can measure current draw. Not exactly difficult, also not exactly common because they don't really fail that often.
Over the past 10 years, I've noticed that most gas stations near me have replaced their pumps. They look like they have 7 segment displays, but if you look close (as you do when you're standing next to it pumping gas), they are actually high resolution LCDs showing numbers in a 7 segment font. I'm sad for what improvements could be had, but aren't.
