Cinemecanica End of Life Server

[Gordon McLeod]

We received a Email the the Cinecloud servers were discontinued and deemed End of life and as of May1 no parts or support

[Mark Gulbrandsen]

Mostly, servers reach end of life because of semiconductor supply issues. And right now there is a huge semiconductor shortage because of COVID. And likely why Dolby had to pull the plug on the DSS-200. Probably the longest lived piece of computer based cinema equipment was the DTS-6. I have seen them rebuilt with no less than 4 gestations of the motherboard they originally came with... The last one was actually a miniature industrial 486 MB that mounted to a larger sub board so all the original connectors could still be used. The XD-10 only used two versions of MB.

[Frank Cox]

A couple of applications that I wrote in the 80's in Turbo C on DOS are still working fine.

One is a water vending machine and the other is an oil well production logging program. I rarely hear anything about them -- nobody's squawking so everything is ok, right?

These days they run on cute little industrial DOS computers that that you could put into your pocket. Some change from what it ran on when I wrote the programs in the first place.

[Harold Hallikainen]

Parts going EOL is a real problem. I had to support three different displays in the JSD-60 and JSD-100 due to the original displays becoming unavailable. The firmware would try some different commands to the display controller to determine which one was out there, then use the appropriate functions to write to them. One of the last projects I had was rewriting code to talk to SPI flash devices that went EOL and were used in LOTS of products. Again, the firmware reads an ID from the chip, then decides how to talk to it. One product has a microdisplay that went EOL. There was no suitable replacement, so we had a special run of the part run. That was actually the second microdisplay I had designed into the product. It is a backlit LCD. I originally designed in an OLED display, but the supplier failed in the 2008 financial crisis. My last job was called "sustaining engineering." Just keep the stuff going!

Harold

[Steve Guttag]

The DSS200 was dropped because Dolby acquired Doremi. The CAT745 could have kept the DSS200 going if the CAT862 had semiconductor availability. However, I think the real reason was just that there was no need to compete with themselves. I, personally, think it was a mistake and it definitely cost sales as Doremi was not our #2 server. But to keep both lines going would mean keeping up with not only semiconductors but also keeping both softwares going and such. Now, I wish they had brought out a new server that somehow allowed both platforms to converge into one but that too was not to be.

As to semiconductor EOL...you are going to see a LOT of audio based products go away due to a semiconductor plant burning to the ground (AKM) that made some very popular Digital to Analog converters. At that point, people either have to spend the money to design around those specific chips, presuming on pin and feature compatible versions exist or choose this point to discontinue the products. With 2020 being the year it was, it will be hard to spend the money on R&D to develop new versions of old products.

[Harold Hallikainen]

We used a LOT of AKM chips (DIR, DIT, SRC, ADC, DAC, etc.).

Harold

[Steve Guttag]

All of the JSD series (as well as the DCP series) just recently went "EOL" though there is stock. I have no doubt it was, in large part due to chip availability and the cost of re-designing old product. The DPM was a relatively new product for QSC and the DPM-300 covers a lot of ground where the JSD and DCP were overlapping. The CM-8E seems to have made the cut. Did it use any AKM?

We are fast approaching that the "canned" cinema processors are going to be CP950, the DPM-300 and the Ovation-2.

[Harold Hallikainen]

I did not know about the EOL of the JSD. They are still on the web site. The CM-8E uses an AKM DSP.

[Steve Guttag]

There is still stock on them so they will stay on the web site. They went EOL on the March 2021 price sheet. The JSD100 is OEMed to others so I'm sure it wasn't a casual decision.

The CM-8E either sells enough to warrant redesign or there's so much stock still that they don't want to slow sales any by announcing EOL at this time.

[Marcel Birgelen]

If those manufacturers that decide to pull certain components from the market would actually share their designs with the community, in many cases you could re-implement it using some generic FPGA for example or maybe emulate it in software on a micro-controller. Some long-dead platforms have been revived this way in the past, but many required extensive reverse-engineering of components lost in time...

A couple of applications that I wrote in the 80's in Turbo C on DOS are still working fine.

One is a water vending machine and the other is an oil well production logging program. I rarely hear anything about them -- nobody's squawking so everything is ok, right?

These days they run on cute little industrial DOS computers that that you could put into your pocket. Some change from what it ran on when I wrote the programs in the first place.

There is so much software, both embedded and non-embedded out there that depends on the Intel x86 architecture, that you don't really need to fear going it away in the near future. Someone will always be baking x86-compatible CPUs the next 25 or so years, probably even way past then. Then, there are lots of virtualisation options available right now, that lets you run this software even on another CPU architecture.

You don't have this luxury with all kinds of embedded components, like some of the components mentioned earlier: DSPs, DA converters or even something rather straight-forward like a simple dot-matrix display. If the manufacturer stops producing them and stocks level out to zero, you end up looking for a replacement. While there are often components that do the same task, they're almost never entirely compatible. In the best case, it's simply a question of re-routing some connections, but in the worst case, the behavior of the component is entirely different, hence you need to redesign the circuitry around it, to fit this new component.

Parts availability remains a big problem and is also the reason why more and more stuff is being pushed into software. While microprocessors used to be way more expensive than most custom components, small ARM CPUs can now be had for pennies and the chance of that platform going entirely EOL is far less likely than a specific range of ICs of a certain manufacturer.