is it!!!!!!!!!!!
microscopic traces laid out manually.tell me u r joking now!
its such a mess to use manual routing.they must be having some avanced autorouter functions that would allow them to route acc. to desired rules.
see what i said earlier.(specially for you)
''''''
go for it if u have deep pockets.i used its 30day trial,really liked it very much,but cost too much for diy''ing''''''''''
AND
''''''''''eagle is more than sufficient for a diy''er.very good for its price.value for money and a powerful package''''''''
plz read posts b4 posting.
yeah,its like that but seriously speaking,i havent noticed any noise related issue with my pcbs routed using autorouter.
sagarverma said:
Yes, I believe so (manual routing, not the joking part). I have certainly laid out boards approaching the complexity of a small motherboard, and worked on Palm PDA boards etc - with 500-1000 components or more and 5 thou (mil) tracks - done exclusively using manual tracking in Protel. Human designers (or teams) can squeeze more into a smaller area and save on layers, which has clear cost benefits for highly squeezed, high volume (or even medium to low volume) products. Protel's design rule checker has always been good at detecting errors. Now it will guide you as you track which is sometimes useful, and there is lilmited push and shove ability for tracks (ie you want to add a line to a bus, you track it in and all the other ones move out of the way).
I expect mobo designers would have much better tools along these lines - and yes it probably is an autorouter with manual guiding - but my point is that if you look at a cheap (ie high volume) mobo you'll see it is not a "push button and walk away" thing, and the job is certainly not beyond manual routing.
Autorouting is great for low-ish or one-off volume, very complex, digital, expensive products with lots of free space and layers available, with expensive enclosures to keep EMC in/out.
I had a good play with Protel's autorouter (Situs) yesterday to see what it will do, and it's actually kind of ok - it CAN track a board like I would, but there is no "overall vision" and it does dumb things and gets itself cornered. It takes a lot of unrouting and semi automated re-routing to get it past these problems. And of course that's with nicely placed components which are the main battle when laying out a board. Potentially a very useful tool, but no replacement for manual design effort yet.
I agree. Circuit performance is rarely affected by layout, and almost impossible on a 4 layer board. Poor grounding and supply decoupling are to blame for nearly everything (in audio or digital) and these practically disappear on a 4 layer board with power planes. Make the layout as messy as you want, the circuit doesn't care about aesthetics. All that really matters is whether the copper is touching or not. (If you're dealing with high-speed, Protel's autorouter will take care of impedance, delay, crosstalk and even pulse shape distortion.)
Skilled manual routing avoids these risks, produces better EMC performance, and eeeks out that last 1% of performance - which is something we like for audio. But that's unimportant to most people. The cost savings I mentioned above is why manual routing gets used in the commercial world, not performance.
adx
so here we have a person who is 'in bussiness'.
thanks to diyaudio,now i have blacklisted pcb design service out of my would be jobs if it is this much of pain in ***.
u sayin manual routing has no match,it allows extra room for improvements etc etc... 4 me its great untill one day i see for myself that yes,this autorouted board is misbehaving.
listen all those thronging this forum,a professional saying this.
this is the point i wanna drive home.this autorouter is best i have seen till this very second.
yes i totally agree.but the solution lies in 'working' on the autorouted boards manually to thicken the gnd tracks as much as possible.i dunno bout theory behind it but i suppose this is what avoids noise to mingle with useful volts.
its 100% true!!
so here we have a person who is 'in bussiness'.
thanks to diyaudio,now i have blacklisted pcb design service out of my would be jobs if it is this much of pain in ***.
u sayin manual routing has no match,it allows extra room for improvements etc etc... 4 me its great untill one day i see for myself that yes,this autorouted board is misbehaving.
listen all those thronging this forum,a professional saying this.this is the point i wanna drive home.this autorouter is best i have seen till this very second.
yes i totally agree.but the solution lies in 'working' on the autorouted boards manually to thicken the gnd tracks as much as possible.i dunno bout theory behind it but i suppose this is what avoids noise to mingle with useful volts.
its 100% true!!
We were autorouting complex multilayer digital boards in the mid 1980s. There were products that were crap and others that were excellent. There are many ways to guide autorouters and it is naive to think that the entire process should be push button.
It is not true that power and ground planes solve the majority of problems, although they do help. Most advanced high speed logic families are not source and end terminated which often results in ringing, ground bounce, and other noise margin related issues. I've debugged plenty of them.
People should look else where, where there's big money to be made, and you'll find solutions. Do people think that 10-100 million transistor VLSI chips are manually routed? No, and companies did the research to find damn good solutions.
Do people realize that this problem has it's roots in the classic "Travelling Salesman Problem" in mathematics that is to this day unsolved:
http://www.tsp.gatech.edu/problem/index.html
"Indeed, the resolution of the TSP would settle the P versus NP problem and fetch a $1,000,000 prize from the Clay Mathematics Institute."
It is reasonable to manually route a simple audio PC board, after all there are enthusiasts who will look inside and expect to find a "pretty" PCB.
Pete B.
It is not true that power and ground planes solve the majority of problems, although they do help. Most advanced high speed logic families are not source and end terminated which often results in ringing, ground bounce, and other noise margin related issues. I've debugged plenty of them.
People should look else where, where there's big money to be made, and you'll find solutions. Do people think that 10-100 million transistor VLSI chips are manually routed? No, and companies did the research to find damn good solutions.
Do people realize that this problem has it's roots in the classic "Travelling Salesman Problem" in mathematics that is to this day unsolved:
http://www.tsp.gatech.edu/problem/index.html
"Indeed, the resolution of the TSP would settle the P versus NP problem and fetch a $1,000,000 prize from the Clay Mathematics Institute."
It is reasonable to manually route a simple audio PC board, after all there are enthusiasts who will look inside and expect to find a "pretty" PCB.
Pete B.
I was thinking about that, even old products had really usable place+route+compact which is neat to watch. C compilers are another example where automation has "won the race" pretty much conclusively. There are people (like me) who like to argue that assembly language is better, especially for small micros, but in reality good C compilers usually code just as well or better than humans except in very limited cases. That's not to say "C is better", there can still be very good reasons to avoid an extra layer of automation (or to favour it) which go beyond traditional measures of performance (speed, size, development time, etc).
I don't know why PCB routers haven't "made it" to that extent (yet), possibly because humans are naturally good at visual problem solving. Some people will disagree of course, as they have only ever used autorouters. I have never used an autorouter in the way it was intended. Probably because I've never found / couldn't afford one that won't "put tracks everywhere".
I stand by my comments about 4 layer boards and grounding/decoupling. I have seen too many problems blamed on layout, when it was something dumb circuit-wise. Of course I'm playing the devil's advocate here. The fact is, the nastiest layout on a 4 layer board will probably work (just) if the circuit is right - up to say 100MHz. Signals will look pretty nasty at the point things start failing, but if nothing fails then nobody might think to look (it happens!). Beyond 50-100MHz, most "push the button" people would start thinking seriously about setting some high speed design rules. Protel takes care of all the overshoots and undershoots and I/O buffer types, and will even suggest termination solutions. Pretty freaky really.
Anyway, I found another autorouter yesterday that looks really promising...
TopoR
http://www.freestyleteam.com/en/products/topor/
Pity it "puts tracks everywhere" (as in it does what it wants to do, not what you want to do).
For 15 years I used Amiga computers. Two electronic CAD programs have appeared at the end of the life of this computer, the best ever commercially sold : AmiCAD and AmiPCB.
Both were the works of one guy only, called Roland Florac. Both were freeware. Both were bug free. Both combined occupied less than 1 Mb of disk space (all programs are very frugal with memory)
I only had a look at the documentation for printing purposes. No need any tutorial time. Every command was evident.
Last october, the european magazine Elektor has released a DVD-rom full of Windows electronic CAD demo packages. I had a quick look at each of them. I had to abandon almost all of them because of their complexity and unease of use. I think the first quality of a program is to be able to be used in a few minutes without any lengthy (and forgettable !) tutorial.
Ergonomics of most programs are very complex and of a poor conception. I am still looking for Windows or Linux programs as friendly, simple, powerful and efficient as AmiCAD and AMIPCB... and some others like the Amiga file browsers.
~~~~~~~~ Forr
§§§
Both were the works of one guy only, called Roland Florac. Both were freeware. Both were bug free. Both combined occupied less than 1 Mb of disk space (all programs are very frugal with memory)
I only had a look at the documentation for printing purposes. No need any tutorial time. Every command was evident.
Last october, the european magazine Elektor has released a DVD-rom full of Windows electronic CAD demo packages. I had a quick look at each of them. I had to abandon almost all of them because of their complexity and unease of use. I think the first quality of a program is to be able to be used in a few minutes without any lengthy (and forgettable !) tutorial.
Ergonomics of most programs are very complex and of a poor conception. I am still looking for Windows or Linux programs as friendly, simple, powerful and efficient as AmiCAD and AMIPCB... and some others like the Amiga file browsers.
~~~~~~~~ Forr
§§§
emulated hardware, bug free software
bug free I don't believe, not for ANY software, ...
But that is not the point of this message ...
I suspect that even the last Amiga was a relatively slow computer by today's standards, so
you could probably run the Amiga sw on an emulated Amiga on top of windows.
bug free I don't believe, not for ANY software, ...
But that is not the point of this message ...
I suspect that even the last Amiga was a relatively slow computer by today's standards, so
you could probably run the Amiga sw on an emulated Amiga on top of windows.
Hi MIRLO,
"Bug free". I mean "almost", of course.
Despite its 50 Mhz, and due to its conception, drawing using either of the two Amiga programs I mentionned, speed was as good as any software tested with Windows, at least sufficient for the basic works that most diyers need. I prefer ergonomy, which means human speed, to 3 Ghz computer speed with human error prone software.
I tried Amiga emulation on Windows, but did not solve the problems of high screen resolution needed for Amiga CAD programs.
However it is great fun to have the possibility to use an extraodirnary Amiga file commander called Dirwork under emulation to manipulate Windows files.
~~~~~~~ Forr
§§§
"Bug free". I mean "almost", of course.
Despite its 50 Mhz, and due to its conception, drawing using either of the two Amiga programs I mentionned, speed was as good as any software tested with Windows, at least sufficient for the basic works that most diyers need. I prefer ergonomy, which means human speed, to 3 Ghz computer speed with human error prone software.
I tried Amiga emulation on Windows, but did not solve the problems of high screen resolution needed for Amiga CAD programs.
However it is great fun to have the possibility to use an extraodirnary Amiga file commander called Dirwork under emulation to manipulate Windows files.
~~~~~~~ Forr
§§§
pinkmouse said:
Just ran the Windoze version in VPC and it looks like a well developed project.
This is the stuff we should be supporting; it's GPL. Help the guy build us a good suite of software and at the same time expand his reputation.
I hope someone can persuade him to compile it for OSx.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.