EmbeddedRelated.com
Forums

Offpage connector symbols

Started by D Yuniskis January 19, 2010
WangoTango wrote:
> In article <hj88ic$vam$2@speranza.aioe.org>, not.going.to.be@seen.com > says... >> Hi AL, >> >> LittleAlex wrote: >>> On Jan 19, 1:17 pm, D Yuniskis <not.going.to...@seen.com> wrote: >>>> As a side-comment to the schematic preferences thread >>>> (hopefully not another lengthy thread :> ), I'm curious >>>> as to what folks use as an offpage connector symbol. >>> I use a CAD program. It has "input", "output", "bidirectional", and >>> "passive" (none of the above, AKA don't care) for off-page and off- >>> sheet. >>> >>> I've never seen a reason to change them from the default. >> Yes, all of the tools I use do this. I am just not happy >> with their symbol choices. And, since I can change them, >> I have. >> >> E.g., I don't like an output on the right side of the page >> drawn as < >> > I guess your CAD package doesn't have a rotate or flip?
Sure! Then you get a symbol that looks like > -- but now the pin connection is on the *right* side of the symbol instead of on the *left* (since this example was describing an offpage connector for an output to be located on the right edge of the page!)
> Funny what these guys will forget. ;)
Funny how these posters fail to think things through! :)
Hi Colin,

colin_toogood@yahoo.com wrote:
> > Something to bear in mind is that many users of your design won't use > the schematic, eg your layout guy isn't going to look at the schematic > for every net he picks up, nor is your firmware guy going to > constantly check when he defines and uses an FPGA pin.
Yes. Though the schematic is the driving document in both cases. I.e., it wins all arguments ("Oh, I thought RESET was on pin 23..." "No, it's not.") The biggest consumer of the document is the end user. He needs to be able to quickly understand what the design is trying to do and how it is trying to do it. To that end, you have to balance "information" with "clutter".
> We name almost every net on the board :- > > {source}_{destination}_{major function name}_{minor function name}
Ouch! Your schematics must be very "dark" :> I only name things that *need* names. E.g., if I have an RC snubber across a switching diode, I don't name the signal *between* the R and the C. Chances are, I will never have to refer to it in my written commentary. And, if I actually *do* need to refer to it (e.g., to tell a technician to probe the signal there), I would simply say "the junction of Rx and Cy".
> With not much thought you can define three letter acronyms for every > source and destination and probably major function name. Suddenly you > have a schematic where you don't have to drill up and down through > hierarchy and fewer mistakes are made.
Yes, but everything you put on a document is one more thing that has to be maintained. It's like putting comments on each line of code in a program. Or, using "FirstArrayIndex" and "SecondArrayIndex" (as in array[FirstArrayIndex][SecondArrayIndex]) instead of array[i][j]). I.e., it's just more than you need. <shrug> YMMV. The whole point of this was to elicit *preferences* as none of these things are cast in stone...
In article <hj9vf5$idp$1@speranza.aioe.org>, not.going.to.be@seen.com 
says...
> WangoTango wrote: > > In article <hj88ic$vam$2@speranza.aioe.org>, not.going.to.be@seen.com > > says... > >> Hi AL, > >> > >> LittleAlex wrote: > >>> On Jan 19, 1:17 pm, D Yuniskis <not.going.to...@seen.com> wrote: > >>>> As a side-comment to the schematic preferences thread > >>>> (hopefully not another lengthy thread :> ), I'm curious > >>>> as to what folks use as an offpage connector symbol. > >>> I use a CAD program. It has "input", "output", "bidirectional", and > >>> "passive" (none of the above, AKA don't care) for off-page and off- > >>> sheet. > >>> > >>> I've never seen a reason to change them from the default. > >> Yes, all of the tools I use do this. I am just not happy > >> with their symbol choices. And, since I can change them, > >> I have. > >> > >> E.g., I don't like an output on the right side of the page > >> drawn as < > >> > > I guess your CAD package doesn't have a rotate or flip? > > Sure! Then you get a symbol that looks like > -- but now > the pin connection is on the *right* side of the symbol > instead of on the *left* (since this example was describing > an offpage connector for an output to be located on the > right edge of the page!) > > > Funny what these guys will forget. ;) > > Funny how these posters fail to think things through! :) >
OK, well then who makes a symbol that isn't grouped into a cohesive unit, or what CAD package can't handle such things? All in/out/bi symbols I have EVER seen are not treated as individual parts/lines/primitives. They are a equivalent to a symbol or part that is manipulated as a unit. So ---> becomes <--- when flipped or rotated. ---< becomes >--- rotating around what would be the electrical connection point. I guess I assumed you spent more than $1.50 on the software..... :)
WangoTango wrote:
> In article <hj9vf5$idp$1@speranza.aioe.org>, not.going.to.be@seen.com > says... >> WangoTango wrote: >>> In article <hj88ic$vam$2@speranza.aioe.org>, not.going.to.be@seen.com >>> says... >>>> Hi AL, >>>> >>>> LittleAlex wrote: >>>>> On Jan 19, 1:17 pm, D Yuniskis <not.going.to...@seen.com> wrote: >>>>>> As a side-comment to the schematic preferences thread >>>>>> (hopefully not another lengthy thread :> ), I'm curious >>>>>> as to what folks use as an offpage connector symbol. >>>>> I use a CAD program. It has "input", "output", "bidirectional", and >>>>> "passive" (none of the above, AKA don't care) for off-page and off- >>>>> sheet. >>>>> >>>>> I've never seen a reason to change them from the default. >>>> Yes, all of the tools I use do this. I am just not happy >>>> with their symbol choices. And, since I can change them, >>>> I have. >>>> >>>> E.g., I don't like an output on the right side of the page >>>> drawn as < >>>> >>> I guess your CAD package doesn't have a rotate or flip? >> Sure! Then you get a symbol that looks like > -- but now >> the pin connection is on the *right* side of the symbol >> instead of on the *left* (since this example was describing >> an offpage connector for an output to be located on the >> right edge of the page!) >> >>> Funny what these guys will forget. ;) >> Funny how these posters fail to think things through! :) >> > OK, well then who makes a symbol that isn't grouped into a cohesive > unit, or what CAD package can't handle such things?
Sure. "Create a symbol". I think if you read upthread, that's where this discussion started.
> All in/out/bi symbols I have EVER seen are not treated as individual > parts/lines/primitives. They are a equivalent to a symbol or part that > is manipulated as a unit. > So ---> becomes <--- when flipped or rotated.
Sure! But the PoE moves, also! Or, graphically: --->X becomes X<--- where X is the PoE (i.e., where the signal connects). If, as I had stipulated in the discussion, you are creating an output for the right side of the page, then you really want: --->X If you are starting with; ---<X I think you will find "you can't get there from here".
> ---< becomes >--- rotating around what would be the electrical > connection point.
Only if that connection point is located in the *center* of the symbol. (Many eCAD packages put the PoE's on the *edge* of the symbol boundary).
> I guess I assumed you spent more than $1.50 on the software..... :)
This was OrCAD 9. I'll check Altium/Protel this afternoon if I get a chance. I know STRIDES would do it correctly (because I could always move the PoE manually if need be). I *really* don't want to fire up the Mentor Graphics workstation to see how *that* does it...
D Yuniskis wrote:
> Or, graphically: > > --->X becomes X<--- > > where X is the PoE (i.e., where the signal connects). > If, as I had stipulated in the discussion, you are > creating an output for the right side of the page, > then you really want: > > --->X > > If you are starting with; > > ---<X > > I think you will find "you can't get there from here".
Argh! I mispoke (confusing placement of signal name with signal PoE). You have: X--< SIGNALNAME (standard OrCAD symbol "OFFPAGELEFT-L") You want: X--> SIGNALNAME Tell me some combination of flips and rotates (remember, you're alleging that "silly me" doesn't need to bother editing the symbol itself -- creating a new one) will transform the first into the second? Unhappy with "OFFPAGELEFT"? You can always try OFFPAGERIGHT: SIGNALNAME >--X But, I think you will find you can't rotate *that* either to get to X--> SIGNALNAME (even if you are willing to manually *move* "SIGNALNAME" each time you place a connector). Perhaps your eCAD program works in N-dimensional space??
In article <hja8b5$4cq$1@speranza.aioe.org>, not.going.to.be@seen.com 
says...
> WangoTango wrote: > > In article <hj9vf5$idp$1@speranza.aioe.org>, not.going.to.be@seen.com > > says... > >> WangoTango wrote: > >>> In article <hj88ic$vam$2@speranza.aioe.org>, not.going.to.be@seen.com > >>> says... > >>>> Hi AL, > >>>> > >>>> LittleAlex wrote: > >>>>> On Jan 19, 1:17 pm, D Yuniskis <not.going.to...@seen.com> wrote: > >>>>>> As a side-comment to the schematic preferences thread > >>>>>> (hopefully not another lengthy thread :> ), I'm curious > >>>>>> as to what folks use as an offpage connector symbol. > >>>>> I use a CAD program. It has "input", "output", "bidirectional", and > >>>>> "passive" (none of the above, AKA don't care) for off-page and off- > >>>>> sheet. > >>>>> > >>>>> I've never seen a reason to change them from the default. > >>>> Yes, all of the tools I use do this. I am just not happy > >>>> with their symbol choices. And, since I can change them, > >>>> I have. > >>>> > >>>> E.g., I don't like an output on the right side of the page > >>>> drawn as < > >>>> > >>> I guess your CAD package doesn't have a rotate or flip? > >> Sure! Then you get a symbol that looks like > -- but now > >> the pin connection is on the *right* side of the symbol > >> instead of on the *left* (since this example was describing > >> an offpage connector for an output to be located on the > >> right edge of the page!) > >> > >>> Funny what these guys will forget. ;) > >> Funny how these posters fail to think things through! :) > >> > > OK, well then who makes a symbol that isn't grouped into a cohesive > > unit, or what CAD package can't handle such things? > > Sure. "Create a symbol". I think if you read upthread, > that's where this discussion started. > > > All in/out/bi symbols I have EVER seen are not treated as individual > > parts/lines/primitives. They are a equivalent to a symbol or part that > > is manipulated as a unit. > > So ---> becomes <--- when flipped or rotated. > > Sure! But the PoE moves, also!
Not in any package I have used. Maybe in some I evaluated and tossed out.
> > Or, graphically: > > --->X becomes X<---
Not in any package I have seen. The rotation point should be around the electrical connection.
> > where X is the PoE (i.e., where the signal connects). > If, as I had stipulated in the discussion, you are > creating an output for the right side of the page, > then you really want: > > --->X > > If you are starting with; > > ---<X > > I think you will find "you can't get there from here".
I have never had to get there to begin with.
> > > ---< becomes >--- rotating around what would be the electrical > > connection point. > > Only if that connection point is located in the *center* > of the symbol. (Many eCAD packages put the PoE's on > the *edge* of the symbol boundary).
Never seen a package where you couldn't add or expressly set the reference point location. I guess maybe I have just been lucky?
> > > I guess I assumed you spent more than $1.50 on the software..... :) > > This was OrCAD 9. I'll check Altium/Protel this afternoon > if I get a chance. I know STRIDES would do it correctly > (because I could always move the PoE manually if need be). > I *really* don't want to fire up the Mentor Graphics > workstation to see how *that* does it...
I used P-CAD which became Altium and it doesn't have any such issues, my ANCIENT DC-CAD no such problems, and I know CAD-Star handles things just fine too. Like I said, maybe I have just been lucky. Strange.
In article <hja8qe$54e$1@speranza.aioe.org>, not.going.to.be@seen.com 
says...
> D Yuniskis wrote: > > Or, graphically: > > > > --->X becomes X<--- > > > > where X is the PoE (i.e., where the signal connects). > > If, as I had stipulated in the discussion, you are > > creating an output for the right side of the page, > > then you really want: > > > > --->X > > > > If you are starting with; > > > > ---<X > > > > I think you will find "you can't get there from here". > > Argh! I mispoke (confusing placement of signal name with > signal PoE).
I see. Now that makes sense. I never saw where you were referring to a signal name, just the symbol itself.
> > You have: > > X--< SIGNALNAME > > (standard OrCAD symbol "OFFPAGELEFT-L")
Well, now there's your problem. That doesn't even make sense for an off page left signal to me at all. Shouldn't that be : <--X off pageleft X--> off pageright <->X Bidir left X<-> Bidir right Where X is your electrical Connection AND rotation point?
> > You want: > > X--> SIGNALNAME > > Tell me some combination of flips and rotates (remember, > you're alleging that "silly me" doesn't need to bother > editing the symbol itself -- creating a new one) will > transform the first into the second?
I wasn't alleging anything. I guess I misunderstood your original point. I didn't understand that you were starting with a bizzaro symbol layout to begin with.
> > Unhappy with "OFFPAGELEFT"? You can always try > OFFPAGERIGHT: > > SIGNALNAME >--X
That looks more like onpage left to me.
> > But, I think you will find you can't rotate *that* > either to get to > > X--> SIGNALNAME > > (even if you are willing to manually *move* "SIGNALNAME" > each time you place a connector). > > Perhaps your eCAD program works in N-dimensional space??
No, it just didn't straddle me with some RPN-esque symbols from the get go. I'm sorry to have wasted your time. I think that CAD-Star's default is to place the signal name over the 'wire' and then you move it to where you want. I will have to play again to find out. I only use P-CAD to support our older products, so I don't remember EXACTLY how it handles the names.
On Jan 21, 6:08=A0am, "Meindert Sprang" <m...@NOJUNKcustomORSPAMware.nl>
wrote:

> I know (from your earlier posts :-) ). Lucky for me I'm on the other end =
of
> the scale. I'm the only engineer here, accounting for 50% of the eployees > and 100% for management :-)
I don't suppose you're hiring? :) This company has crossed the event horizon where it is no longer possible to do anything except procedure.
On Thu, 21 Jan 2010 11:24:32 -0700, D Yuniskis
<not.going.to.be@seen.com> wrote:

>Hi Colin, > >colin_toogood@yahoo.com wrote: >> >> Something to bear in mind is that many users of your design won't use >> the schematic, eg your layout guy isn't going to look at the schematic >> for every net he picks up, nor is your firmware guy going to >> constantly check when he defines and uses an FPGA pin. > >Yes. Though the schematic is the driving document in both cases. >I.e., it wins all arguments ("Oh, I thought RESET was on pin 23..." >"No, it's not.")
Not necessarily. In fact, it's dangerous to have the schematic be the definition of correct. The specification is *supposed* to be the definitive document. Also, it's not unusual to 'fix' the I/Os on an FPGA first. The schematic then picks up the pinouts from the FPGA.
>The biggest consumer of the document is the end user. He needs >to be able to quickly understand what the design is trying to do >and how it is trying to do it. To that end, you have to balance >"information" with "clutter".
The "end user" never sees our schematics. They see the product.
>> We name almost every net on the board :- >> >> {source}_{destination}_{major function name}_{minor function name} > >Ouch! Your schematics must be very "dark" :>
Confusing, too.
>I only name things that *need* names. E.g., if I have an RC >snubber across a switching diode, I don't name the signal >*between* the R and the C. Chances are, I will never have >to refer to it in my written commentary. And, if I actually >*do* need to refer to it (e.g., to tell a technician to probe >the signal there), I would simply say "the junction of Rx >and Cy".
Test point: TP1234 ;-)
>> With not much thought you can define three letter acronyms for every >> source and destination and probably major function name. Suddenly you >> have a schematic where you don't have to drill up and down through >> hierarchy and fewer mistakes are made. > >Yes, but everything you put on a document is one more thing >that has to be maintained. It's like putting comments on each >line of code in a program. Or, using "FirstArrayIndex" and >"SecondArrayIndex" (as in array[FirstArrayIndex][SecondArrayIndex]) >instead of array[i][j]). I.e., it's just more than you need. > ><shrug> YMMV. The whole point of this was to elicit >*preferences* as none of these things are cast in stone...
Sure, and it's been a good discussion. I think a lot of this stuff is driven by the tools used, though.
"larwe" <zwsdotcom@gmail.com> wrote in message
news:9461facb-5dd6-4b27-a068-36c8247ece12@p24g2000yqm.googlegroups.com...
> I don't suppose you're hiring? :) This company has crossed the event > horizon where it is no longer possible to do anything except > procedure.
No, unfortunately not. Unless you bring in a couple of fine long term jobs, of course... :-) Meindert