Parallel port driver - whois?

Hugh Jack: > Who is going to take the lead on this? Thomas, Willy, Petr........ The parport module currently in the CVS is Mario's. Mario - what is the current status of the module? Is it in progress, ready, maintained, orphaned? Jiri -- Jiri Baum <[email protected]> Q: Why did the chicken cross the Moebius Strip? A: To get to the other... um... er... --r.h.f.r _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
A

Andrew Kohlsmith

> > I mean the hardware. > Of course, sorry. If nobody has time I can do this. (lurker mode off, BTW.) Long ago I designed an 8255 board after growing tired of running out of I/O with the parallel port. I currently design embedded industrial controls for Benshaw Canada (http:/www.benshaw.com). An 8255 board with high voltage outs and 240VAC isolation wouldn't be to difficult to whip together. Regards, Andrew _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
W
Well, if you've already done it, then more power to you! Can you use Eagle to do the schema/layout so anyone can look at it/modify it (I'm thinking especially of students here) for free? If you do this one, I'm going to continue to hallucinate on how to make a (really) cheap Ethernet I/O block. Parallel boards are Ok, but I want remote, and Ethernet is the way to go here IMO. Willy Smith _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
A

Andrew Kohlsmith

> Well, if you've already done it, then more power to you! Can you use Eagle to do the schema/layout so anyone can look at it/modify it (I'm thinking especially of students here) for free? I've used Eagle on occasion but have always given up in frustration. While I'll do the initial design in OrCAD I will (once I have it hammered out) put it into an Eagle schematic as well. I fully understand and appreciate the need for a cross-platform and cheap schematic entry / layout program. :) > If you do this one, I'm going to continue to hallucinate on how to make a (really) cheap Ethernet I/O block. Parallel boards are Ok, but I want remote, and Ethernet is the way to go here IMO. I agree with you 100% on this. A side project I'm working on uses the uCsimm for its ethernet capabilites (Modbus/Ethernet, etc.) Regards, Andrew _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
A

Andrew Kohlsmith

One thing I forgot to ask: How does everyone here feel about surface mount components? Personally I love them and find them as easy to install as through-hole but I know that some people do not feel this same way. While I won't be putting fine-pitch devices on the board I would like to be able to use standard PLCCs, SOT-23, SOIC and 1206-size components. Basically 0.1" and 0.05" but nothing finer. Comments? Regards, Andrew _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
C
Hi Willy Me too. I have looked quite seriously at using the MachZ SOC running embedded Linux and using a SIMM socket backplane to allow various I/O modules to be installed. This could be a remote Ethernet rack, part of a distributed control scheme, or even a standalone LPLC all in one. I am hung up on the backplane bus. I2C is cheap and doable as digital and analog chips exist for this bus, but it might be too slow. Higher pin count busses would fix this, but raise the component count and the cost. Maybe there's something there you can use. I couldn't get funding for protos or packaging. Regards cww _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
W
Curt, I used I2C in industrial products since 1988, in the laundry industry. We made little racks with 8574s and 8574As, Opto 22 modules plugged in and they gave you 8 points of I/O. They were connected with d-sub 9s. We used a 4 bit adder with a little LED display so that each time you added a rack, its address incremented and showed you the rack number. They work just fine inside panels, still in use today. Ran em at 100 KHz, not too shabby. But I think Ethernet is tthe way to go now. I'll send you a picture of our latest platform when it's prettier, it's a 1" x 2" board with a 33 MHz ARM7TDMI, 2 megs of flash and 512K of RAM, and a 10baseT interface. We're squeezing uCLinux into it. It ain't easy. I'm going to try to come up with a platform that costs nothing and has Ethernet for I/O modules. I'll post my latest hallucinations when I have them. Willy _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
C
Hi Andrew I have been doing a little digging on this and have found a quad optoisolator that's fairly cheap, ( $1.00 ) in 100's and offers 5000 V isolation at .25 an input. I have the info at work and can mail you the spec sheet pdf if you're interested. I have used the 8311 for an output level shifter and can send a pdf on that too. It gives 16 high current 24V OC outputs in one package. Regards cww _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
C
Hi Andrew. The only place where that would be a problem is on the 24V side. The isolation requirements dictate wider spacing and .5 amps per output continuously requires bigger traces for reliability. The package dissapation will probably require dips and maybe even some planes for heatsinking. The MOS stuff can be surface mount. Discrete hexfets might help by lowering dissapation but are spendy and eat space. Regards cww _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
C
Hi Willy I was going to use the I2C onboard and Ethernet to the controller. But, you're right that this would cost something. The MachZ is overkill but would offer comfortable standalone LPLC operation. Let me know how it goes. I'm intensely interested in doing ethernet IO from Linux both for LPLC and my day job. Regards cww _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
A

Andrew Kohlsmith

> I have been doing a little digging on this and have found a quad > optoisolator that's fairly cheap, ( $1.00 ) in 100's and offers > 5000 V isolation at .25 an input. I have the info at work and can > mail you the spec sheet pdf if you're interested. I have used the > 8311 for an output level shifter and can send a pdf on that too. > It gives 16 high current 24V OC outputs in one package. Sounds great... I'd love to receive PDF's (and the manufacturer's website) from you. Regards, Andrew _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
J

Johan Bengtsson

Well, I like them too, but I realize people might find them challenging. I do however not think it would be a too hard challenge anyway. I like them better, not only for the smaller size, but also for the fact that, when I home-brew something, I have less holes to drill. go for it, especially if it makes the design easier. /Johan Bengtsson _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc ---------------------------------------- P&L, Innovation in training Box 252, S-281 23 H{ssleholm SWEDEN Tel: +46 451 49 460, Fax: +46 451 89 833 E-mail: [email protected] Internet: http://www.pol.se/ ----------------------------------------
 
P

Philip Costigan

Hi All, I think that if I can look at developing a Modbus/Ethernet "slave" protocol for LPLC and others develop the parallel port I/O hardware etc. Then it is then possible to setup remote I/O that can even have a little bit of local inteligence. The master LPLC can then act much like a DCS and the slave LPLC's can be small Advantec like boxes with parallel port I/O. Does this seem reasonable or are we trying for somthing even more compact? -- Regards Philip Costigan P.S. I am just about to get enthusiastic again after a long sleep. _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
C
Hi Phil and all. By all means! If we can get enough bodies working on it, I see absolutely no reason why we can't have our own Ethernet IO and even the first truly Open and Free automation protocol ever on the planet. I am often obsessed with Ethernet as a fieldbus and unlike the automation companies, I can get over caring if someone other than me does it. It has been a pet project of mine, but I will share my research with anyone or help wherever I can to make it happen, because _that's_ what is important. Along this line, it seems that here we have enough interested parties to form a SIG and please collaborate as much as possible. Working together is a powerful secret weapon that no one else in the industry has. The remote rack that can be smart IO, a distributed node, or a standalone micro is a very powerful concept that Linux makes practical even in low volumes. A parallel interface that is well supported can hit an unprecidented bang/buck ratio for education and small projects. It can be a very low investment way to actually do something with LPLC which is crucial for winning people over. A defacto standard 8255 card that can handle "industry standard" automation components in multiples at low cost can turn any old PC into a kickass LPLC for even large applications. With a PC full of these and our Ethernet IO we can offer just about anything anyone else does in terms of digital IO. Same advice on sharing and possibly a SIG here. Working together is important. Sounds like a cottage industry in the making. I had no idea there were so many other hardware hackers available. If these things get organized and are happening, I will look at analog IO, both cheap and dirty and metrology quality. I would like to offer machine vision but my employer owns most of my work on that so I can't promise anything....yet :^) It would be a tremendous boost to the project to have so many others contributing in any capacity. A little effort by a lot of people can make so many more things possible, that we can easily surpass what any one company can do. Can you see that in this burst of triffic on the list we have strengthened the project dramatically?. It's your project and we can use whatever you want to contribute. And your contribution makes a big impact. This is really what the project is about. Regards cww _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
A

Andrew Kohlsmith

> If we can get enough bodies working on it, I see absolutely no > reason why we can't have our own Ethernet IO and even the first > truly Open and Free automation protocol ever on the planet. What about ModBUS/Ethernet? Sure it's not fancy (the term "fieldbus" has many meanings) but I was under the impression that Modicon had released ModBUS into the wild years and years ago. If fieldbus can be used to describe a low-level industrial communications protocol then ModBUS can be the lowest common denominator. Any kind of *process* control though would require something far more ballsy than straight register reads and writes. > Can you see that in this burst of triffic on the list we have > strengthened the project dramatically?. It's your project and > we can use whatever you want to contribute. And your contribution > makes a big impact. This is really what the project is about. I agree. Personally I flip-flop a lot from being a bottom-up designer to a top-down designer as different things require different design methodologies. In my opinion some kind of I/O design and lowlevel communications protocol should very much be a bottom-up design. Get the bugger working and make sure it's industrial-grade that won't break down and *then* add the glitz and high level design components as necessary. It's like designing a PWM controller. Get that PWM output absolutely perfect and it doesn't matter what kind of high-level control you throw at it from above... the output will always be clean and you won't have trouble with component failure due to dead-time issues or low-level code not understanding what the command was. Putting that back into I/O-land: if the hardware meets UL/EU (VDE?) standards (CSA can be met if you do careful UL testing, probably similar with the few other standards groups) and design has margin, and if the lowlevel protocol can be guaranteed to get a message from one place to the next without buggering it up then you can trust the base systems which makes further product enhancement so much easier. :) Anyway this little rant went a little far... I'm still recovering from a buck and doe last night. :) I do have a question regarding what kind of time people are able to put into this... Like many lurkers here I do other things besides hang around the list... Three kids (one brand new), a full-time job and a design contracting business on the side makes for some interesting time constraints... What are other people's schedules like and when and if it comes down to crunch time do many (any?) of us on the list have the ability to give additional time? Regards, Andrew _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
C
Andrew Kohlsmith wrote: > > > If we can get enough bodies working on it, I see absolutely no > > reason why we can't have our own Ethernet IO and even the first > > truly Open and Free automation protocol ever on the planet. > > What about ModBUS/Ethernet? Sure it's not fancy (the term "fieldbus" has > many meanings) but I was under the impression that Modicon had released > ModBUS into the wild years and years ago. If fieldbus can be used to > describe a low-level industrial communications protocol then ModBUS can be > the lowest common denominator. Any kind of *process* control though would > require something far more ballsy than straight register reads and writes. I am a fan of Modbus/TCP also. It's the closest thing to a standard Ethernet IA protocol. We have stuff for it but, it's not Open and it's not Free as in freedom (not free beer) The hardware will work with either, but if someone wants to work on a protocol that is Free and Open that's great in my book. > > > Can you see that in this burst of triffic on the list we have > > strengthened the project dramatically?. It's your project and > > we can use whatever you want to contribute. And your contribution > > makes a big impact. This is really what the project is about. > > I agree. Personally I flip-flop a lot from being a bottom-up designer to a > top-down designer as different things require different design > methodologies. In my opinion some kind of I/O design and lowlevel > communications protocol should very much be a bottom-up design. Get the > bugger working and make sure it's industrial-grade that won't break down and > *then* add the glitz and high level design components as necessary. > > It's like designing a PWM controller. Get that PWM output absolutely > perfect and it doesn't matter what kind of high-level control you throw at > it from above... the output will always be clean and you won't have trouble > with component failure due to dead-time issues or low-level code not > understanding what the command was. Putting that back into I/O-land: if the > hardware meets UL/EU (VDE?) standards (CSA can be met if you do careful UL > testing, probably similar with the few other standards groups) and design > has margin, and if the lowlevel protocol can be guaranteed to get a message > from one place to the next without buggering it up then you can trust the > base systems which makes further product enhancement so much easier. :) > > Anyway this little rant went a little far... I'm still recovering from a > buck and doe last night. :) I do have a question regarding what kind of > time people are able to put into this... Like many lurkers here I do other > things besides hang around the list... Three kids (one brand new), a > full-time job and a design contracting business on the side makes for some > interesting time constraints... What are other people's schedules like and > when and if it comes down to crunch time do many (any?) of us on the list > have the ability to give additional time? > That's why I am encouraging people to collaborate as much as possible. A little time from several people can maintain good progress and prevent burnout. And as much as we hate to admit it, other people have good ideas too. It's hard to share, but it's worth it. In the beginning I was gonna do this whole thing myself. It's hard to give up any control but I'm glad I did. regards cww _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
Top