Future-Proof
localroger
Posts: 3,451
I was in a discussion last week with a potential customer about a product I created using a QuickStart as its basis. And I was asked, what is the future availability of the product. I know Parallax promised the P8x32A to be available for ~15 years not too long ago, but what about the QuickStart, FLIP, or PropMini? This is not at all meant as a dis, since Parallax has been there for us since forever, but more like what can I tell my customers without being wrong when I say it.
Comments
-Phil
Qty: 5 spare Quickstart modules... $500.00
Siemens have a 6 year EOL on their industrial control components.
$500 should be chicken feed.
?? Quickstart $34.99 X 5 = $174.95
Better yet, FLIP at $27.99.
-Phil
PropMini may be impacted by the EOL QFN P8X32A-M44 chip at some point in the future, although AFAIK Parallax have kept a significant number of chips to support future production. (like they always seem to do for customers).
Interestingly, I just checked the store and notice the EOL tags appear to have gone from the P8X32A-M44 product page. Maybe the manufacturing issues were resolved for that case style and Parallax were able to keep it? I'll ask them.
From what I recall, there are lots of PIC chips that are EOL already.
Sorry, I should have been clearer. I was referring to the selling price to the customer. Nothing wrong with a bit of a markup
Absolutely not, and unless the code and additional hardware was absolutely trivial to produce that's a very low markup.
-Phil
The situation we have is that there are several thousand industrial machines, all costing over $100K new, whose custom OEM embedded controllers are obsolete. The original manufacturer has taken this as an opportunity to, um, perform an unwanted reproductive service on their customers. We are coming in with a cheaper and, more importantly, simpler and easier to maintain alternative. However, all of these customers have already been bitten hard by the obsolescence bug, so they are asking pointed questions. I need to have solid answers in place as we get set to market our solution. The cost to do one of these retrofits will be in the $50K range, most of it travel and labor for setup and installation.
Mickster, that Siemens thinks 6 years is an adequate EOL window for their products is ridiculous. I know people who are using 40 year old PLC's in mission-critical applications. They don't need more power or features, they need it to be reliable without spending money to redevelop a system that's been working for 40 years. They are nursing along old DOS computers to run the obsolete DOS-only software needed to troubleshoot and program those things. And they'll continue to do that until their hand is forced by the smoke escaping from something they really can't replace.
(This is where Parallax goes “hey... you know... we could do a FLIP2...”)
(This is where Ken goes “oh no... please dont let Chip see this...”)
(This is where Chip says “I actually had this in mind when I designed the P2”)
(This is where Peter says “hey! You could do this in Forth!!)
(This is where Phil says “I *told* Roark to stay on his meds”)
Ok. Sorry. My bad. I’ll put the caffeine down now and back slowly away from the computer...
The P2 is not quite ready for prime time and in the industrial world, "Failure IS NOT an option".
When all the bugs have been worked out of the P2 and it has been through stress test hell and back many times over then it could be used.
If you want to make some wiz bang gizmo that customers will tolerate having bugs or the magic smoke getting out then the P2 is your baby.
Killjoy.
To be clear, my post was offered (mostly) as a bit of sidebar humor. That being said, the main takeaway should be that there are many creative options to deal with a manufacturers EOL’ing any given product. Sometimes it helps to think outside the box.
-Phil
It's all open-source; even the P1 itself is open source, so in some kind of worst-case development we could always use FPGA's or even form a group and have more original P1 chips fabbed. This is not a thing we are ever likely to do but it is possible if necessary. Unlike, for example, recreating the ecosystem of the Intel 80386SX, which is too bad so sad like dead, and with it the previous controller for these machines that was built on that ecosystem.
On the PC end all of my software runs just fine under emulation via WINE, although I've never bothered to deploy it that way it's an option if Microsoft really drops a shiv hammer on all of us one day. I might have to move past the ENC28J60 for ethernet one day, but I'm not really seeing that being necessary any time soon. By the time that's necessary P2 will probably be mature enough to reliably do native ethernet on its own.
I think I can walk into my next customer meeting with my head held high. Thanks of course as always and most of all to Parallax. You guys just make amazing stuff.
How about using Quemu to simulate a 80386SX, it is a tad slow but the 386SX was too...
Mike
You mentioned that they are using DOS based machines to support PLC stuff that uses DOS to run things. And that these are 40 year old ones. (The computer that is.) Do you happen to know what brand of PLC this is? If it's Allen Bradley I believe they are some what proprietary in that their OS was home written and they moved from it to embedded Windows around the beginning of this century. In fact the last release of DOS was made available to OEMs for the sole purpose of the embedded device market starting about the time of Windows Seven on the desktop.
GE-FANUC was in the same bucket, I remember an interesting gathering at one of three places, it turned out to be running an industrial strength version of Windows CE, which is what A*B did do as well.
And what's decidedly bizarre about A*B is that during the late 1980s they were a firm supporter of the DEC PDP-11 and even building motion control boards for it. Eventually they even leaned towards the VAX(!) and did much the same thing.
There's more of course, but I felt I'd ask. And contribute.
You should give Ken a call at get it straight from the horses mouth, including something in writing if possible.
At the moment Parallax is focused on education because that's what's paying the bills but they have always been very supportive of their commercial customers.
At a few thousand machines, does it make sense to use a (flip|QS|etc) rather than rolling your own into the project? Then your main worry would be just stocking the prop chip. I liked the GG board and the DIP on a breadboard. And a PPDB (only because the pins are not hardwired to things I may not need).
As an alternative, since the P1 was open sourced for for FPGA, you could always make that a backup counter point to them being left high and dry by explaining the use of FPGA to become whatever the project contains. iNTEL/Altera and Xilinx will not likely go away anytime soon. So while you will want for cost reasons to remain with the Prop, your customer can see that there is a backup plan if needed down the road.
No need to speculate on their future availability. Please design them into your products!
Ken Gracey
Definitely feels like planned obsolescence.
I'm at ~$20K/retrofit for machinery that is $150K - $200 to replace.
However, I have $300K - $500K machines coming up that are currently fitted with 12 axis Siemens 840D controls.
Machines are good as new, mechanically but they can't get support from the OEM nor Siemens.
I am currently working with an enclosure manufacturer to give me the big-name look. Can't go sticking exposed PCBs in place of a Siemens system.