Chipmaker Analog Devices to acquire Maxim Integrated for $21B
reuters.com
reuters.com
Joke aside, AD is probably one of the most profitable companies with unprecedented monopoly on analog components and circuits business. Last week somebody mentioned about Visa and Mastercard duopoly in HN, but with this Maxim acquisition, AD is now the Visa and Mastercard combined while TI is the American Express.
To see how profitable AD is please check this article on the AD's 3rd generation transceiver chip and if you are not reading it, basically it's more profitable than drug [1]!
I am using every single versions of this AD's transceiver chip for my work, now it is in the latest 6th gen. Suffice to say if you're building 5G transceiver you will need AD's 5th gen transceiver. It's probably not an exxaggeration to say that ITU radio/wireless cellular standards' bandwidth actually following the AD's transceiver chips generations since 5G system can be built by the 5th gen chip but not the 4th gen chip (it's for 4G). The closest competitors are transceivers from LT (already acquired by AD) and the Lime Microsystems from Cambridge, UK.
[1]https://zeptobars.com/en/read/AD9361-SDR-Analog-Devices-DAC-...
Thier ROIC is 6.7% https://www.gurufocus.com/term/ROIC/NAS:ADI/Return%2Bon%2BIn...
This is below the SP500 ROIC.
Visa / Mastercard ROIC is >20%
Even worse is that basically the only product using the chip is Lime themselves so trying to spin your own product would be a nightmare. ADI has a nice eval board and reference design for their chips and you can control it using the lib-iio framework which is built into mainline linux.
I returned my LimeSDR Mini after a frustrating weekend and bought a USRP instead, which uses the AD chip, and there's really no comparison.
This is one reason why I said in another comment that I thought it'd be difficult for a fabless upstart to complete on a complex / high-performance design like an ADC. The chip Lime designed is an absolute monster -- it has multiple ADCs, DACs, filter banks, switches, oscillators, PLLs. If any one of them isn't perfect, the whole design is shot. I'd still like for them to succeed, but when you're selling a hundred dollar IC, it really does need to be flawless.
[1] https://discourse.myriadrf.org/t/new-limesdr-mini-loopback-t...
For example, when userspace requests a certain TX gain, the chip does some calibration black box magic to determine the best way to distribute that gain across the internal amplifies. Nothing special here, the AD9361 does the same thing. The Lime chip will inexplicably fail the calibration but __will continue to operate__. If the PLL didn't lock, the chip sees something dumb like the loopback being down at -45dBFS so it cranks the gain on all of its internal amplifiers to __correct__ the loss. This results in one or more of the internal amplifiers saturating and you end up with horrible intermod. I was seeing -10dBC ACPR on a simple 1MHz QPSK signal. Sure, you can save off good calibrations or do it by hand, but if the simple loopback cal doesn't work and won't tell the user, the chip is irremediably broken.
You say this like it's a good thing. It's one of the most generic, do everything gigantic overhead drivers I've heard of. On the pluto it limits things to 4 MS/s when without that libiio overhead it can push double that over the USB interface. I suppose things might be different with top of the line AD transceivers.
NXP + Freescale in 2015
Microchip + Atmel in 2016
ON Semi + Fairchild in 2016
Infineon + Cypress in 2020
Analog and low-to-mid complexity digital designs don't usually use the smallest, newest, most expensive silicon processes that you need for processors, GPUs, and FPGAs. You generally need capacitors, precision resistors, and wider voltage ranges more than you do billions of transistors.
Maybe now that these older fabs are being forced to run as actual businesses rather than as bleeding-edge science projects, semiconductor design companies are able to bring them back into the fold to avoid dealing with the headaches of being fabless.
It's too bad though, because this adds a huge capital cost to what would otherwise be a really ripe opportunity for a new competitor. This consolidation has definitely brought higher prices and reduced the diversity of available parts.
The unit economics of analog ICs should be very good -- a product that needs 1/100th the silicon surface area and sells for 1/10th the price, using a much cheaper node than a modern digital IC. There should be plenty of room for a company to compete with Analog Devices on price while still making a killing.
Analog is a bit different in that “node” doesn’t really apply, but they are also not in the high value part of the value chain for the most part.
Even if it's expensive, owning a fab means you have the option to make prototypes of a design, or of parts of a design. You will never, ever do this if you're fabless.
We're good at simulating digital logic, but simulating analog designs is more difficult, and each process tends to have unique quirks. You want your designers to be familiar with these quirks, which is easier to do when everybody designing and using a process is under the same roof.
If you're making a chip that has exotic needs (voltage ranges, threshold voltages, RF performance, noise, thermal properties, bipolar + cmos, etc.) you will have more ability to tweak the process. Foundry type fabs typically offer a smaller "menu" of options that they're comfortable they can support. For example, I think you might have a hard time competing with some of AD's more expensive ADCs as a fabless semiconductor company.
Don't get me wrong... there are plenty of headaches to owning and operating a fab too.
There's certainly the possibility to do analog chips here, but it would take a big team effort.
(have only dabbled a bit with FPGAs with soft-cores, last project I used one for was ~2009: running Linux on an Altera NIOS core, where we sampled at 100 Msps from an input until we filled up the RAM, then more slowly dumped it over Ethernet to a PC)
thanks for the detailed explanation!
* Motorola => Freescale & ON Semiconductor
* Fairchild Semiconductor => ON Semiconductor
* Dallas Semiconductor => Maxim
* Signetics => Philips Semiconductors => NXP
* Freescale => NXP
* National Semiconductor => Texas Instruments
* Linear Technology => Analog Devices
* Atheros => Qualcomm
* Intersil => Renesas
* Atmel => Microchip
How long until we have to sign a license agreement before we can use an OpAmp? Oh, and the license is only valid for consumer applications. Want to use the OpAmp for enterprise applications? That'll cost you more.
That's if we're lucky. In a darker scenario, all OpAmp designs have been bought by Apple, and you can't even use one if you opened your iPhone because the function has been integrated into the CPU.
> In a darker scenario, all OpAmp designs have been bought by Apple, and you can't even use one if you opened your iPhone because the function has been integrated into the CPU.
Scary, because many microcontrollers already have OpAmps built into them...
The margins on an OK product just aren't worth it, and to be competitive you have to build the whole thing and provide a reference design that's within 20% of the best out there to even break even. No surprise the market is all oligopolies. If there wasn't open source and affordable fab services coming up, there would really be no hope.
On the other hand, after Analog purchased Linear, many of the high-performance Linear parts are still sold side-by-side with competing Analog parts today, Analog even created a "Powered by Linear" product line for selling Linear power converter chips. It was a wise decision, apparently the management knew those parts from Linear are of great value. I hope Analog will adopt a similar solution for these Maxim parts.
Discontinued parts also included passive parts originally made by the HP Components subsidiary from the 1980s, such as special Schottky diodes and PIN diodes for RF/microwave applications, up to 10 GHz, still perfectly working today. For example, HP's jelly-bean HSMS‑282x series 6 GHz Schottky diodes was the go-to choice in RF circuits (even at lower frequencies like VHF and UHF) for three decades and still in production as of 2016 - you can find their datasheets with an HP logo, another with an Agilent logo, another with an Avago logo, and the last one with a Broadcom logo - and they eventually came to an end when Broadcom killed them in 2017 after Avago's acquisition.
I was hit by this. On a recent weekend I was tinkering with a DIY software-defined amateur radio receiver design and needed some RF diodes, only to find all of them have been killed, and similar parts from NXP were not stocked by the local distributor, 10-day shipping... A friend told me that they have switched to Skyworks diodes since then. The legendary life and unfortunate death of HP diodes.
Tip: scroll down to pass the sponsored content to see the full news section below the fold.
That being said, I was disappointed with EE Times’s vacuous take on this merger this morning. They have changed hands, and managed to survive, but like most of the vertical literature are a shadow of their former selves.
Here’s the editorial, written by the publisher: https://www.eetimes.com/maxim-is-a-pit-stop-wheres-adi-heade...
And, not in your scope, but AMD acquires ATI, 2006, still has some echos. Also, Atheros/Qualcomm 2011.
I guess Analog Devices makes LTSpice, so it makes sense that they'd be all in for that kind of support.
https://www.analog.com/media/en/technical-documentation/appl...
"High Speed Amplifier Techniques
A Designer’s Companion for Wideband Circuitry
Jim Williams
This publication represents the largest LTC commitmentto an application note to date. No other application noteabsorbed as much effort, took so long or cost so much."
More of his app notes for the curious: http://readingjimwilliams.blogspot.com/
Even reading their datasheets, I often felt like I was being taught a full lesson on the topic, which is great when you're an amateur and lack some of the foundational knowledge.
https://gist.github.com/dantler/f26d75f1051db13649aaedf73f7d...
https://www.xilinx.com/about/xcell-publications/xcell-journa...
Avoid them at all cost.
ADI's documentation--relative to many chip makers--is outstanding. They have EVBs for almost all of their parts, open source hardware designs as well as firmware--and support that does not require you to tell them you'll order 100,000 pieces next year.
I wish as much as anyone else that they sold their products for less, but the market clearly supports their existing prices.
$5 for a $2 op-amp or a regulator is one thing, but when you need something slightly special, it goes to $10 or $20, just because TI doesn't have a comparable part. If you need more than a handful in your design, you're screwed.
Yeah, their parts generally are very good, but the industry isn't healthy if they're the only company with good designs.
All this consolidation has left a lot of room for a good competitor. We're not seeing one here in the US, because students today think analog stuff is terrifying black magic and avoid it like the plague, but I'm a little surprised we haven't seen any serious competitors out of China, since analog designs don't usually need the crazy expensive silicon processes that digital designs do.
Also, for the same $20, would you rather sell a 0.5 cm^2 microprocessor at 45 nm, or a 0.1 cm^2 analog to digital converter at 130 nm? Analog ICs should be a really good business.
In the United States, today, an RF design engineer's job prospects are mostly limited to military contractors.
>> All this consolidation has left a lot of room for a good competitor.
Which will come from Asia. US companies seem to avoid competition as much as possible. The rather obvious outcome of that is they end up uncompetitive. The thing is, new stuff has to be developed while you are still profitable, not after someone is winning away your business.
Since my home electronics projects are one-off things... the 250% increase for better documentation and better LTSpice models is well worth the $3 cost increase.
They make very exotic and state-of-the-art products. Where else would you get stuff like LTZ1000 or AD797?
The idea that prices should depend in any way of costs is ridiculous.
If you're a hobbyist like me, you don't really care about spending $5 on a chip instead of $2 on a chip... especially when you're paying multiples of that for the PCB and other parts of the equation.
I built a digital audio interface with an ADI ultra-low-noise clock generator. It was a $20 part. That's a ridiculous price but it's also only $20.
From the perspective of a hobbyist with limited space: you want to minimize the SKUs that you stock in your personal shelves. Buying a higher-end op-amp and spamming it everywhere (even when its specs aren't needed) is far simpler than buying 10x different op-amps at the $0.50, $1, $1.50, $2, $3, $4, and $5 price points.
Just keep a supply of higher-quality $5 rail-to-rail low-bias op-amps at the voltage-level (3.3V for most Arduino projects).
Yeah, there's probably a $0.50 op-amp that does the job. But do you really want to keep another SKU on your shelf and keep track of it? There's simplicity in just buying over-specced parts for personal hobby projects.
--------
Well... maybe keep a stock of the $0.50 stuff too (EDIT: Ah right, LMV358. That's the cheap part I keep around). But I think you get my gist. Personal-supply closet management is certainly a problem for the personal hobbyist.
Yes, it will work in general. Unless the expensive opamp has a significantly wider bandwidth, then one must use with caution. Switching from a low bandwidth opamp to a high bandwidth opamp is especially problematic, the loop characteristics are different and it may need additional frequency compensation or bandwidth limiting, the original layout may become inadequate, etc. More often than not, a blind replacement make it oscillate like crazy.
Sometimes having a bunch of cheap opamps makes life easier.
And financially - your heart and wallet won't start bleeding when your $0.5 opamp is unexpectedly zapped by ESD, cooked in soldering, or fried in an overload. Killing a $10 ASIC is a big deal (just did that last week while prototyping an ASIC controller, but there isn't anything cheaper on the market).
https://www.analog.com/en/support/customer-service-resources...
The march of progress continues forward. Its not computer tech per se... but power-tech and analog-tech is incredibly important still.
I'd argue that without modern IGBTs, it'd be impossible to make an electric car today. Its probably one of the most major sources of efficiency in the modern electric vehicle. The power-switch that controls "on" vs "off" needs to be incredibly efficient when you're shoving 200+ kW of energy through it!!
Case in point: lets say 200kW of power is being used by the motor at 375V. That's 500+ Amps of current. A resistance of 0.1 Ohms would waste 25,000 Watts of energy (power == current^2 * resistance)... in the switch / transistor itself before the motors even got the energy.
Yeah, when I was working in EVs everything was about reducing losses. Not just for efficiency but because a few percent loss is a LOT of heat.
I ran an inverter at 200kW that was not rated for that. The weak point was actually the DC connector. I didn't find that out the hard way, just looked it up as I was closing in on 600 Amps DC.
BTW I just needed to find 2 values empirically and was unable to because of that DC current. Strange enough I figured running at a lower voltage would reduce max power and I could get my values. Unfortunately half the voltage is half the power but still the same current. Higher speed had similar issues. Project got cancelled before I found that motors limits.
Er, everybody switched to SiC MOSFETs, so... definitely not.
> The power-switch that controls "on" vs "off" needs to be incredibly efficient when you're shoving 200+ kW of energy through it!!
Even before SiC became economical, FETs were edging out IGBTs in most applications. The voltage drop across IGBTs is just too high.
In fact SiCs have worse on/off efficiency, but are still preferred because they can switch more quickly and in the end that makes them more efficient, overall.
> Its probably one of the most major sources of efficiency in the modern electric vehicle.
I would give that to the battery or the motors before the controller. Controller have been >97% efficient for ages and ages, but motors are another thing entirely and better computers, salient rotors and permanent magnets, and better characterization have led to bigger gains than the switching elements have seen.
Neither hold a candle to batteries though. The resistance of switches is <10 milliohms, and motor resistance isn't much higher. The resistance of a 400 volt, 250 Ah NiMH battery is around 180 milliohms. The same battery built with venerable NCR18650Bs (original model S) is 68 milliohms. That's a full 7x improvement, or an 86% reduction in loss, and by far the biggest inefficiency besides aerodynamics and rolling friction.
Power dissipated by resistance is only linearly (not quadratically) related to resistance when current is relatively fixed: P=(I^2)R. So reducing the parasitic resistance from 180mΩ to 68mΩ is only a 2.6x improvement, albeit still quite significant.
Definitely not everybody are on SiC parts, let alone MOSFET only designs. Though, the trends is for at least some hybrid SiC parts as they are getting cheaper.
When switching speeds are low enough (as is in EVs,) you don't need that much switching performance. Current Si IGBTs are quite viable in EVs, if you switch high enough voltage.
There is a reason why rail is still preferring high voltage AC for traction. That's because Si IGBTs for that are cheap, and efficient enough.
I always thought of ADI as a company with ASICs like PLLs and MEMS and whatever. Are IGBT drivers an important line of business for them (I don't follow the company closely).
If you look at another company that makes a shitload of power semiconductors, like Infineon, they are struggling.
CREE is a large but shrinking LED business tied to a small but growing SiC business. At some point the SiC part will overshadow the LED part and CREE will be seen as a growing successful tech company again.
IGBTs were more of a TI-comment. I see TI as more of the power-company.
I like talking about IGBTs because its superficially a simple subject. It just a switch: power turns on, power turns off. But when you really consider the shear level of efficiency being developed today, its mind-boggling.
-----------
You're right that I don't see ADI as an IGBT (or similar) company. But I just sorta went on a tangent, hooking into your TI-comment more so than ADI.
- the idea of injecting money into a lame economy causing inflation is preposterous
- comparing historic stock market prices w/o inflation data is perfectly sensible
or
- that inflation data is readily available to everyone and doesn't need to be dragged into the discussion here?
Well, I'm afraid, we'll never know.
These chips cost hundreds or thousands of dollars and are meant for base stations. Analog Devices wants to get a contract with Nokia or anyone else and then also be able to sell the LDOs, PLLs, and clock devices that go along with it (https://www.analog.com/en/products/AD9213.html#product-tools). Linear Tech was the leader in power and the name is still used for new power products under the "Power by Linear" name with ADI. Hittite offers a number of RF chips. If you're paying $1k for a data converter, what's an extra $50 for all the analog/RF support circuitry? They may even throw it in for free. If you have a complex system with analog/RF/digital circuitry, offering a full reference design with other parts from your own catalog is a great way to ensure that those parts are used.
If so, that's a pretty big change to the industry.
If you want practical examples of what they do, go to digikey.com, search for the company names and click randomly on one of the thousands of hits.
Here's a microcontroller: https://www.digikey.com/product-detail/en/maxim-integrated/M...
Here's an I2C temperature sensor: https://www.digikey.com/product-detail/en/maxim-integrated/M...
Here's an RF amplifier: https://www.digikey.com/product-detail/en/maxim-integrated/M...
Here's a 3v3 DC-DC: https://www.digikey.com/product-detail/en/linear-technology-...
As an aside, AD has a large number of parts which are non-public. You will not find the part numbers (or any information about them) without an NDA and personal assistance from an AD rep.
I agree that the categorization can be a bit tricky at times, but once you're used to it it's not terrible.
Also often times you'll find some great component on e.g. Maxim's website, only to find no one distributes it, so it's effectively useless in the immediate sense.
As a hobbyist, this sucks so much. With PCB manufacturing reliable enough for high speed signalling being affordable and hackerspaces with soldering ovens in every major cities (and there was some startup - in NYC I believe - that you could give Eagle files and a product order list and they would buy you the components at Digikey and manufacture the board for you... anyone remember their name?!), it really really sucks that all the interesting stuff is hidden beyond such obscure walls.
You won't even find Intel Thunderbolt datasheets on obscure Chinese sites. Wtf? And then Intel complains about why people aren't adopting it?
Are you thinking of MacroFab (https://macrofab.com/) located in Houston? I believe they have been featured here before.
I don't know about that specific startup, but there are tons of firms in China which will manufacture and assemble PCBs for you. Most of the big hobbyist firms, like JLCPCB, Elecrow, and PCBWay, offer some form of PCBA service. JLCPCB in particular is closely integrated with LCSC, and will source parts directly from them.
Not a microcontroller.
There is overlap, but the portfolios are complementary. It seems to be a good acquisition save the sticker price seems too high.
Maxim tends to make a bunch of almost generic odds and ends, are middle-to-low prices, with fairly large volumes. They compete directly against Texas Instruments for a lot of interface products.
This acquisition is especially interesting in light of AD's (relatively) recent acquisition of Linear Tech, which was basically a consolidation of the high-performance low-volume integrated circuit market.
Does your hardware need power? It needs power controller and converter chips, such as a MOSFET driver, a DC-DC converter, a charge pump, a linear regulator, a voltage monitor, or a power multiplexer (if it has multiple power sources). Does it have a battery? It needs a battery charging controller, or a battery gauge. Does it need to communicate via a data interface or a cable? It needs interfacing chips, such as I2C chip, SPI chip, IIO chip, RS-232/RS-485/RS-422 chips, CANbus chipss, LVDS chips, voltage level translation chips, and I/O extending chips. Does it have USB? It may need a USB multiplexer chips to switch between different signal sources (for example, USB Type-C can be an audio port or a data port), or a USB charging controller chip. Does it need wireless communication? It needs RF front-end chips, RF amplifier chips, RF frequency synthesizers, and SoCs for Bluetooth, Wi-Fi, GPS, etc. Does it need data acquisition capabilities? It needs ADC chips, DAC chips, amplifier chips, and signal conditioning chips. Does it need NFC/RFID? It needs NFC and RFID chips. Does it have a screen or a LED light, like a screen backlight or a flashlight? It needs LED controller chips and power chips. Does it need any kind of software control? It needs a microcontroller. The list goes on.
And Maxim's products cover almost all of these applications.
So does Analog Devices' products, but with an emphasize on high-precision, high-performance components.
(The caveat to the above is that LTC pricing was previously absurd.)
I guess inevitable. The high-performance business is getting smaller in volume and consolidation allows prices to go up.
Who does that leave for high performance ADC/DACs? Texas Instruments. Microchip has a limited selection. That's about it.
Lovely.
https://www.marketwatch.com/story/chip-maker-analog-poised-t...
Crazy values nonetheless
I don't understand why they never replaced them; the glorious MC1648 itself was upgraded as MC12148 and MC100EL1648.