The recent discussion about the “new” 6502 from WDC got me thinking about other old and/or unavailable chips that could use a re-introduction, or perhaps just a process facelift. This post is mainly to solicit reactions and suggestions from viewers like you, but I’ll provide my own example too.
Having recently gotten back in to analog synth design after a long hiatus, I was rather shocked to discover that OTAs (operational transconductance amplifiers) suited to such work are few and far between. With the exception of the NE5517 and the LM13600/13700, there are no other audio OTAs currently in production. TI (via BurrBrown) manufactures other OTA devices, such as the OPA860/861, but these are optimized for high-speed analog (RF and transmission line) and they are rather expensive. For many synthesizers, the OTA forms the heart of nearly every voltage-controlled module and using them greatly simplifies the process vs. building with discrete components (such as a Moog ladder). The best OTA for audio was probably the CA3280, which is sadly no longer in production, due to the process becoming obsolete, among other factors. This 2005 blog post by Don Tillman, which explains the whole problem perfectly, was part of the inspiration for this blog post, and I’m going to throw in with Don and humbly request that someone please update the CA3280 to the latest fab, give it 21st-century specs, and bring it back into production.
If the 3280 is a no-go, for whatever reason, then my second choice would be an update of the LM13700 (originally from National). The LM13700, which was designed by Don Sauer, is currently available from both National and NJR, but it’s got its problems. It’s somewhat noisy, for one. It also has an output offset current which scales with the transconductance (a problem when building integrators), and the buffer is just a simple darlington pair. It would be great (read: totally awesome!) if someone were to update the chip to a BiCMOS process, while changing the design to drop the noise floor and increase the dynamic range, reduce the output offset current and replace the darlingtons with FET-input opamps. This might sound like a very involved request, but the expertise is already there to do it — analog devices have come a long way in 30+ years, and the basic design of the 13700 is not very complex, compared to modern opamps.
In case you’re wondering what incentive there is for somebody to do this, I submit that, compared to when I was doing it 10+ years ago, there’s A LOT more information out there now about DIY synths, and the increase in homebuilt synth projects has paralleled the increase in DIY ‘tronix in general. There’s definitely a willing (and growing) market for these things, and with a chip redesign they could certainly charge more per chip. There’s also a resurgence in analog electronic music in general right now, which means these new chips may find their way into commercial products as well.
So anyway, that’s my request. I suspect there are lots of different opinions on the subject, however. If you have a chip you’d like to see resurrected and brought back into production, or just updated to modern spec, post it in the comments! Try to include some info with your suggestion, and reasons why you think such an update would be successful.
The answer is easy: I want whatever chip that JUST went obsolete to not be obsolete anymore. I also want to give a stern talking to whomever decided not to keep making it.
A bigger problem is that as chipmakers hope to move up in the value-chain, they keep pulling in more functionality and system level design (think SoCs). They also want to differentiate from their competitors so you’re often hard-pressed to find two chips that are pin and function compatible. I’m predicting we’ll all have to dole many more stern talking-to’s to chip execs in the near future.
Commodore SID chip
LM308
The SID chip from Commodore has a huge following but won’t be produced for many reasons.
T.I. has a lot of old chips that could be produced but aren’t.
It is time to create something new using microcontrollers or FPGA and forget people who won’t.
Instead of dealing out stern talking-to’s to unheeding chip company CEOs, wouldn’t it be better if the chips could never go obsolete at all. Like, what if we had some magical machine in our basements that could crank out a chip for us whenever we needed it. I’m not sure what you would call it, but it seems like everyone would want THAT!
I miss the LM3909 LED blinker.
While we are asking manufacturers to produce old chips, I would like to make a plea for manufacturers and trade publications to allow their future users and those who help mentor future users access to material (digital) that costs them virtually nothing to distribute. I am speaking of young hobbyists that are interested in technology and older unemployed or retired engineers that could be encouraging younger people. The IEEE could even have a hobbyist section with free membership. I think that our existing programs (ones I’ve read about)to encourage STEM are targeting students that are too old.
I cannot remember when I became interested in electronics, but I got my first multimeter when I was 11. I subscribed to every "Popular" magazine I could afford and went down to the local pharmacy to read those beyond my meager budget. Delivering these digitally to youngsters would not cost much. While mailorder electronics parts stores are great, shipping costs add substantially to the cost of experimenting. Arduinos sold in Radio Shacks are a good first step. I would encourage electronics parts manufacturers to use Radio Shack or similar outlets to provide parts, inexpensive development boards (like TI MSP430 MCU Value Line LaunchPad development kit) for the latest technology. Companies just might want to think in terms of "raising" future customers.
I think analog synths doesn’t make sense nowadays, except for nostalgia, because there are so powerful and cheap microcontrollers available, if you don’t use a PC anyway for it, or something like the Raspberry Pi. A SID, without the rest of the C64, could be emulated in a small microcontroller. If you are really good, you can implement chiptunes in a small AVR with 8 MHz and 8 K flash: http://www.linusakesson.net/hardware/chiptune.php
Same with many other old chips: Often they are not produced anymore, because there are betters chips. There are some exceptions, like good chips from Maxim, where you need to buy 1000s of it to force Maxim to start production again.
How about applying the “If you’re going to kill it, open source it!” rule?
Many of these chips are decades old, outliving any patents that would restrict their distribution. Publish the masks and let the community get on with it! Overall it seems a lot less fraught than opening previously closed software. (Then again I’ve published a lot more software than made semiconductors, so there’s a good chance I’m missing something important.)
@Scott Harris – I loved that chip! It’s my Google+ profile pict.
Har har Dave. Don’t mess with my dreams! 😉
@Steve: that idea was in the back of my mind when I was writing this, and from a design standpoint that could definitely work, but from a production standpoint, the cost to do small runs of chips is prohibitively expensive for most individuals (or even hobbyist groups). This will change in the future (of that I am certain), but for now it’s untenable.
But that doesn’t mean we couldn’t learn from the designs. If it’s going to be obsolete anyway, with no hope of being reintroduced, that it makes sense to open-source it.
On the same lines as SID, how about Commodore’s other masterpieces, Agnus, Denise and Paula? For those who are too young or too old to remember, these were the Amiga’s core chipset, designed by the brilliant Jay Miner. He also designed the 2600’s TIA and the Atari 800’s ANTIC.
http://en.wikipedia.org/wiki/Jay_Miner
On a more practical (and recent) note, the S1D13806 was a great chip. Basically an LCD and TV display driver with 1.25MB of embedded DRAM, with integrated RAMDAC. QFP package and (at the time) really cheap and easy to use.
STP16C596 was a popular constant current led driver chip used in the early Arduino days. People kept making new designs for it even though it was deprecated, then it dried up. It’s a very powerful chip and there is no easily sourced pin compatible replacement (there is one by Allegro, but it isn’t easy to source in the USA)
I’d also love to see the Commodore SID and the Atari GTIA (video) chips brought back.
I would personally love to see the Curtis/SSM synthesizer chips resurrected and available again. Would make repair of a ton of classic synthesizers alot easier, and to have the opportunity to build new hardware around those classic chips (without killing a perfectly good synthesizer) would be very welcome.
In particular:
CEM3360 Dual VCA,
CEM3340 VCO
CEM3310 EG
CEM3320 VCF
and they synth-on-a-chip ones like:
CEM3394 1 Synth Voice: VCF,VCA,Env (ADSR),VCMix