If you can afford Altium Designer for home use then you can afford to get a double sided plate through board made! Heck, you can afford to hire someone to lay the board out for you.
Do you board double sided plated through and get it manufactured by someone if single sided is turning out to be such a big problem for you. The novelty of making your own boards wears off pretty quick now that professional boards are pretty cheap to manufacture. Use Futurlec, PCBcart, or one of the other cheap prototype places and get a nice double sided plated through solder masked board made, it won't cost you all that much, and you get a nice silkscreen and solder mask.
Actually I'm playing around with a friend of mines that uses it for work. (He does power stuff but just got out of college and doesn't really anything)
Well, I will do that eventually but I kinda like the feeling of doing things myself at this point. Its a good learning experience and once I get the hang of it I can get a simple board done in a day(I've etched about 5 boards with about 30 components on each). I used the pcb express software thats free and doesn't have autorouting or a large footprint library.
I mean, its not so much that I'm looking just to get the board done but to prototype. I have several designs and once I prototype them I'll probably send them off all at once and get several made. At this point the designs are fairly simple that I can do them myself and I don't need it professionally done. Nonetheless I still want to strive for for the best I can do.
I guess the problem is I just thought you meant that doing it in 10 mins means doing it very well and getting a pretty good optimized layout. If you just mean routing the thing so it works then I can do that in probably 15-20 mins(cause I don't know the hotkeys to altium and there are still a lot of features that I end up playing around with(alignments and stuff)).
So what I'm going to do now is route it and probably etch it today. not going to worry about perfection. (going to use larger traces just encase I have problems with etching it). There are some fears I have of the smt because they are so small which is why I want to reduce the number of jumpers and pads so its just easier to debug and work with in general. (to many jumpers might get in the way or cause problems if they are close to smt's and I misjudged the size). Hell, my soldering irons tip is probably 40 mills so its not going to be easy to work with them. (I was thinking about using a hot air gun but not sure how well that would work)
Anyways I'll try and get that stuff done today if I have done. I don't think routing will take more than an 20-30 mins if I don't run into trouble or try and perfect things. (in fact I could just place 0-ohm resistors on all the ports and net aliases and probably get away with autorouting doing it if I wanted to be lazy about it).
I've never used one that would work worth a damn, although I once saw a Japanese (?) one that was supposed to work well. The only way I've found with SS boards is to use quadrille paper (preferably light blue
5 squares to the inch) and a pencil. Placement and orientatation are way more important than with 2 or more layer boards, and I used to go through a lot of paper before I was happy with the layout.
Best regards, Spehro Pefhany
--
"it\'s the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
No, it's an extremely easy problem. Your trouble is that you're fighting an uphill battle against geometry. Your circuit does require a certain number of crossover points, period. You need the jumpers. There's no way to minimize the number of jumpers by redistributing the parts on the board.
It's either possible or imposssible, and it's very easy to figure out which one.
It shows.
With your simple N-fold circuit, the solution is immediately obvious depending of whether you can route between 0603 or SOT-23 pads.
Of course it is. Buy by now you must have realized that each intersection you avoid by repositioning a part will create another.
With SMD and realistic design rules, yes. Actually I've never seen a single-sided SMD board, with or without jumpers. Double-sided PCB is a reality with SMD.
What I said was easy is to figure out how many, if any, jumpers the design needs per repeated block. After you've figured that out it's easy to route. To you the problem seems difficult because you're stubbornly insisting on doing the impossible.
Here's a simple experiment: You seem to have an autorouter. Now lay out your components on a BIG board. Plenty of space everywhere. Make vias extremely expensive for the autorouter. Assuming that the autorouter will try to brute-force its way through the board (which is what it's designed to do), it'll find the solution with a minimal number of jumpers. Using that number of jumpers you can start improving (shrinking) the board.
Don't try to fight mathematics.
We were, and probably still are, talking on completely different levels here. You're facing a problem that no amount of intelligence will solve. Once you've realized that you'll see that it is a very simple problem indeed.
I have found a use for an autorouter on small designs. I let it route everything that is a straight shot then hand-route the rest, ripping out the occasional straight- shot connection as needed. This saves time; I don't need to hand-route a connection betweel two adjacent pins on a DIP or between two resistors right next to each other. Other than that, I manually route everything.
LOTS of single-sided boards these days have SMT components on them. Particularly when there's one big ASIC, but also there are a lot of SMT passives as well. Of course the SMT parts are on the opposite side of the board from any through-hole parts.
Best regards, Spehro Pefhany
--
"it\'s the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
I had not done a single sided board for 10 years, but I did one a few weeks ago for a simple power supply. I had forgotten how crappy they can be - plated through holes provide a lot of strength that you don't get with single sided. You have to be very careful not to apply a side load when trimming component leads after soldering, otherwise the pads move and crack the trace. Same for external strain on connectors and (large parts in general).
Some of my projects have ended up using single-sided SMD boards, but those were boards with quantities in the millions where every fraction of a cent matters.
While composing the post, I accidently erased my sig, so I cut and pasted from the header. Looks like while doing so I pasted whatever snippet of text I had last selected.
Actually, not retired in a sense. Today's programs don't use dongles - thankfully Altium doesn't believe in them - and the users have an immortal license to use the program, which works very well right now, so no problem. Furthermore they've made some nice improvements to it recently, and promise at least one more before they're done. Also, not only can we keep on using PCAD, but we all got free licenses to Designer, and of course are encouraged to "switch," but even after starting to use Designer we can continue to use PCAD, of course.
I've noticed a few nice features of PCAD are missing from Designer, but I've also noticed some of these showed up in the latest update.
That's too bad -- I agree with Win that, while it had its warts like all EDA tools do, PCAD had evolved into a pretty good package.
Their emphasis on the "problem" with P-CAD being a "point tool" is greatly exaggerated, I think... from a business perspective I'm sure they love the idea of their selling you everything from schematic capture to logic synthesis to C compilation to analog or RF simulation to PCB layout to electromagnetic simulation, but from a practical perspective no vendor has ever managed to pull this off in a respectable fashion -- it's just too broad of an area for any one company to try to cover. The best approach, IMO, is to make the @#$%#$% tools OPEN in the sense that the netlist (or whatever) formats are well-documented, so all the "point tools" can cleanly communicate with one another. (AWR took this approach with Microwave office, and at least from their returns, it seems to be working well for them.)
PCAD had a very well documented netlist format (PDIF) before the mid
90s. I still have a copy of the PDIF reference. I think Altium is trying to emulate Cadence in killing a product meant for the mid to small size company. Now folks are forced into their expensive big-boy suites so they can get more money. It's a shame that nice tools get swallowed up by the bigger players so they can kill them.
We've never had any problems with "point tools" involving designs with programmable devices. It seems that 70% of our boards have some sort of FPGA on them these days, including analog boards. Some crappy integrated suite sure isn't going to make it any easier when you're stuck in some convoluted methodology. Weenie managers are easily lured into enticing terminology like "integrated" and "IP". They should be forced to use it!
Orcad Layout can do single sided autorouting / manual routing with pre-defined standard size jumpers. Jumpers can be converted to components afterwards so you can use 0 Ohm links.
--
Reply to nico@nctdevpuntnl (punt=.)
Bedrijven en winkels vindt U op www.adresboekje.nl
I recently got tossed from a project because I wasn't speedy enough using Virtuoso/Composer (that program is SO bad, and so klutzy to use (*) it must have been written by a leftist weenie ;-) Nevermind that PSpice simulates 11X faster and matches the results exactly.
(*) Can you imagine having to write a script (OCEAN) just to obtain corner simulations?
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |
America: Land of the Free, Because of the Brave
PCB design is 90% placement. If you get that right at the start, the actual routing becomes trivial. That is why I said routing 100 nets takes no time at all, and the use of an autorouter for such a task is just silly. I was not implying a really good job all up would take 10 minutes, but it almost could if you were practical about it and used copy and paste for a multichannel design like this (as another poster demonstrated). The figures were designed to demonstrate how easy this problem is if you approach it correctly.
ARGHHH! You said above that you wanted to "strive for for the best I can do", and now you say you still might use an autorouter??? If you *really* want a single sided board with no holes it is possible by using 1206 0 ohm resistors as jumper links. 1206 would allow you to route two tracks underneath each "jumper" if you are careful.
Also, why are you going 0603? If you want the absolute smallest board possible then fair enough, but in that case you shouldn't even be considering a single sided board. It seems like all the stuff you are saying is contradictory. You can't have your cake and eat it too.
life is full of contradictions. First off, I can only use what I got. I have 603's because thats what I ripped of a motherboard and I brought some SOT23's. So sure it looks ike a contradiction to you but its cause your not working within the parameters of the problem I have created.
Heres the difference between what you say,
"If you want the absolute smallest board possible then fair enough,"...
Its not that I want the absolute smallest board possible.
Its that I want the smallest board possible that I can do that works for my application(obviously if its to small then its useless). I also want to minimize jumpers because its more work to have more jumpers. Actually I don't care to much about having the smallest board because its already probably going to small enough.
So actually I'm just trying to minimize the work. You say, then why not just get someone else to do it? Thats not the point(and infact doesn't minimize the work because then I have to work to pay them). I just remember when I had to drill about 100 holes will a dremel and it wasn't any fun. So I want to try to say as far away as drilling as I can. I also want to get a little work in with SMT because I have not done any.
So if you want my "rules" its,
minimize jumpers
minimize holes
small size(not necesssarily smallest because it doesn't matter to much after a certain point. I don't want something to be 10'x10' when 99% of the area is not used).
minimize jumper length(actually probably use a standard length)
--
BTW, each channel has 5 "ports" so its much more difficult then channels
with 3 or 4 ports. Think of a square. If you have a component in a square
then you cannot route out that component and it takes 2 jumpers or 2 vias.
Now no matter how I route, because I have 5 connections that must connect to
each channel, and there are 8 channels, there will be "squares" formed
around some components. Now they are not complete squares and maybe one
could find a way not to box in the components but I doubt it will happen in
10 mins... although maybe I'm wrong and maybe with enough experience one
easily sees how one can connect things as to minimize those issues. (But
just from the shear number of variables it seems like its much more
difficult)
Anyways, I forgot I have to add some more components and do some other
circuits before I etch.
Jon
SMD parts do not take kindly to being removed and reused. Do yourself a big favour and buy new ones. 0805 or 1206 will be much easier to solder, easier to route tracks under (if needed), and cost practically nothing.
You are taking my post out of context, I only said that in relation to the price of Altium Designer.
Then I would not start with 0603's, use the bigger and easier to handle 0805's at least.
"Rules" are the reason this is turning from a molehill into a mountain. If you want a standard jumper length and no holes on your board then that's possible using 1206 or 0805 zero ohm resistors as jumpers.
That's why you need to practice. Routing is an art, and single sided can require a different methodology to double sided, which in turn can use a different methodology to 4 layers, and so on.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.