Former Analog Devices employee here. I used to write programs and that collected all of the data for the data sheets. We would take all of these complex measurements on a part in just seconds using some pretty fancy test equipment. It definitely taught me a lot / inspired me to learn more about programming as a a fresh-out-of-school EE<p>Definitely happy to see some data sheet love ..chip companies are definitely in a whole different world when it comes to documentation, compared to the software world I'm now in!
Data sheet culture at these old-line semiconductor companies - ADI, TI, Linear - descends from the fact that you used to hold it in your hand and read it on paper. When I was in that field we went to the company library to read datasheets because ISO9001 required the company librarians to record which revision of which part an engineer read at what date and time. And it was partly sales literature because you'd be flipping through a bound volume of ADI's current parts looking for something suitable.<p>If you really want to get the whole experience then I urge you to print these out. The graphics are high-resolution.
A couple recent exceptional datasheets I've come across include ADI's ADA4530 [1] and Renesas/Intersil's ISL28110 [2]; both, coincidentally, happen to be for regular old op-amps. The ADA4350 datasheet in particular is absolutely stunning and proof that the old ways are not completely gone. The ISL28110 datasheet is notable when, in an era when I all too often need to read between the lines to determine the type of <i>input transistor</i> used (seriously, guys, it's not hard to show and there are only six possibilities so it's not like you're giving away secrets here), they show a <i>complete device model</i>.<p>That said, I still miss LTC, NSC, and the pre-Renesas Intersil.<p>[1]: <a href="https://www.analog.com/media/en/technical-documentation/data-sheets/ADA4530-1.pdf" rel="nofollow">https://www.analog.com/media/en/technical-documentation/data...</a><p>[2]: <a href="https://www.renesas.com/us/en/www/doc/datasheet/isl28110-210.pdf" rel="nofollow">https://www.renesas.com/us/en/www/doc/datasheet/isl28110-210...</a>
Kind of a rant, so forgive me for venting out :)<p>An untold truth is that datasheets are also the terrain for disputes with tech support. That is, there is a whole team of people called Application Engineers in said companies dedicated to support, that basically knows very little about the practical side of the product and all they do is to read datasheets, just in case there are engineers out there that design products without studying it.<p>So in order to pass the first (and usually not very helpful) line of support you better study the datasheet because it will be a sort of "legal" document you will have opened in your screen for the next week, until the application engineer gives up and connects you with someone who knows the product.<p>Regarding the quality of documentation, yes, nobody beats old monsters like AD. In regards of reference designs... AD is used to give you a 10"x10" board to evaluate a tiny little component; let's say a battery charger. Of course the reference design board has traces of the width of a thumb, very big gold-plated connectors for the power supply, etc. So when you go and integrate the little component in your wrist watch and it doesn't work, the application engineer will go with: "yeah, your design is not much like the reference design, so it's kind you will run into trouble.".... Imagine the rage...
Many years ago I was on a team designing a "smart sensor" box. The heart of it was an integrated precision A/D + 8051-compatible microcontroller chip from Analog Devices:<p><a href="https://www.analog.com/media/en/technical-documentation/data-sheets/ADUC814.pdf" rel="nofollow">https://www.analog.com/media/en/technical-documentation/data...</a><p>None of us has ever worked with microcontrollers before. It was my first ever job out of school. I don't think we would be able to deliver the product if the chip datasheet was not awesome. It had everything. Memory organization, registers, peripherals, board layout and grounding recommendations, ADC driving, etc.<p>Thank you, Analog Devices engineers.
When teaching students about datasheets, I generally pull up an AD datasheet and state, "This is what a good datasheet should look like."<p>They're not the only makers of good datasheets, but they're very consistent. (and many of their chips are awesome)
Linear Technology's AN47 "High Speed Amplifier Techniques" is my favourite document in this category. It is both nicely presented, and a great resource for high performance electronics:<p><a href="http://www.linear.com/docs/4138" rel="nofollow">http://www.linear.com/docs/4138</a>
I have "datasheet-envy".<p>I'm one of the co-founders at satsearch [1], where we're working on fixing procurement in the space industry. We're constantly battling shitty datasheets.<p>As a space engineer, I found it infinitely frustrating to work with flyers that suppliers provide on their website, that also double-up as preliminary datasheets. They're messy, inconsistent, and subject to whatever the supplier thinks their buyer might want to know.<p>We wrote a blog article about the mess that exists in the space industry [2], and a follow-up about what we think can be done to improve this through Electronic Data Sheets (EDS) [3].<p>The EEE industry seems streets ahead in this regard, which is perhaps primarily down to strong commercialization and rapid growth of the market over the past 20 years. I think with companies like Octopart [4] making it patently obvious who does a good job of supporting engineers, it might also have pushed suppliers to improve their documentation.<p>I'm curious though about the consistent use of PDFs. I understand from a usability point of view that they're great for documents, but I would have imagined that the EEE industry would have started distributing "smart documents".<p>To give you an example, the AD datasheet in this article includes a number of high-res graphs: why has the sector not moved to the point where these graphs are interactive elements that you can zoom into, programmatically read off of, export data from, etc.? The way I look at it, a Juypter notebook or something similar would be an awesome way to share datasheet content, allowing much more interactivity.<p>Anyone with any thoughts as to why datasheets haven't progressed in that direction?<p>[1] <a href="https://satsearch.co" rel="nofollow">https://satsearch.co</a><p>[2] <a href="https://blog.satsearch.co/2018-02-21-building-better-datasheets-for-space-mission-design.html" rel="nofollow">https://blog.satsearch.co/2018-02-21-building-better-datashe...</a><p>[3] <a href="https://blog.satsearch.co/2018-01-23-the-challenges-of-an-engineer-in-the-new-space-age.html" rel="nofollow">https://blog.satsearch.co/2018-01-23-the-challenges-of-an-en...</a><p>[4] <a href="https://www.octopart.com" rel="nofollow">https://www.octopart.com</a>
On the other hand, TI's datasheet quality seems to have gone downhill, ST and Infineon's has always been horrible, and Broadcom is infamous for not releasing them at all.
Most electronic engineering component-related documentation is good. It has a long history tracing back through National Instruments.<p>Intel's 80386 documentation was superb.
One other interesting repository of design knowledge: Texas Instruments' reference design catalog.<p>TI will <i>give</i> you the plans to build all sorts of machines and gizmos using their chips. ATMs, industrial robots, blood oxygen meters, even oscilloscopes - you can find detailed design information (schematics, BOMs, layouts, etc) on TI's website, free of charge.
I owe a great deal to National Semiconductor and their application notes. Prior to college I tought myself a lot of what I know about analog design from this handbook [0]. I remember in middle school calling up the support line and having an extremely patient applications engineer walk me through my design and point me at resources to learn. He always answered my emails, even though I'm sure he had more important questions to answer. I tried several years ago to track down the emails to contact him again and thank him, but those emails seem long gone.<p><a href="http://mirror.thelifeofkenneth.com/lib/electronics_archive/NationalSemiconductorLinearApplicationsHandbook1994_text.pdf" rel="nofollow">http://mirror.thelifeofkenneth.com/lib/electronics_archive/N...</a>
IMO - This article is an over simplification. Data sheets need to be book worthy documentation due to the behavior of complex semiconductors. I believe Datasheets are law abiding documentation? If something is out of spec or out of characteristics is is considered faulty which means loss of money or competitors can get in.<p>What I'm really interested is in what peoples opinions are regarding the navigation of content and material of these large scale Semiconductor Sites. They have 50k plus products which are all very complicated and specific to a use-case. I find there is a constant struggle to route people to the right solution especially when engineers can use different parts in new creative ways. ADI, Texas Instruments, and Maxim all fall in this boat.
ADI has gone down hill. For example, the HMC6300 datasheet is crap, having wrong register values, and lacking the interfacing details. Hittite, prior to ADI, would have never had a poor data sheet like that. The ADI web site also stinks, compared to what was Linear Technology.
I believe the Signetics 25000 series 9C46XN document is arguably the pinnacle of datasheets:<p><a href="http://www.repeater-builder.com/molotora/gontor/25120-bw.pdf" rel="nofollow">http://www.repeater-builder.com/molotora/gontor/25120-bw.pdf</a>
The data sheets of AD, National, TI et al have also followed a pretty standard format in terms of layout, order of descriptions etc (something that I have always assumed predated semiconductors). This helps make them particularly useful.
Reminds me of the cartoons in Tektronix manuals:<p><a href="http://w140.com/tekwiki/wiki/Tektronix_Cartoons" rel="nofollow">http://w140.com/tekwiki/wiki/Tektronix_Cartoons</a><p>I think it was recently discussed on HN.
To really understand what the author is getting at it helps to have some experience with lower quality data sheets from other manufacturers. Take, for example, the WS2801, a popular LED driver chip from World Semi[1]. All of the necessary information is there (I used it to successfully wire them up and bit bang the LED colors) but it's just not pretty. The blue header is jarring and the graphs are screenshots from a scope.<p>[1] <a href="https://cdn-shop.adafruit.com/datasheets/WS2801.pdf" rel="nofollow">https://cdn-shop.adafruit.com/datasheets/WS2801.pdf</a>
They do make beautiful documentation, so clear concise and well ordered. Would be nice to see who the authors are but I suspect they might be introverts. Certainly show a love for their work
"hehehe, I remember my first krabbie patty".<p>Looks like someone stumbled across a data sheet, and never read any before. Almost all of them look like that. Ti, NatSemi, ON, DIODES-inc, NEC, Toshiba, etc, etc. Ti and National Semiconductor used to be great reading for new engineers.
I used to worked at Intel as a design engineer and I had to look through data-sheets to find the ICs with features and specs that I needed for the design.