[sci.electronics] PCB info request update. .

tac@sei.cmu.edu (Timothy Coddington) (03/15/88)

	I hope what follows is informative to those who are interested in
	PCB design.  I've included my original post to this newsgroup, and
	some of the better/more informative responses I got back.  I've
	done some editing and have removed some names.  If you have
	anything to add, I'd appreciate it if you'd e-mail me.  

	As for AMS (see my original post, 2nd para), I Have an IMPORTANT
	WARNING - IT STINKS!!!!!!!!.  It is full of bugs, and I've only 
	gotten the run-around since I bought it from them.  Since November
	I've been told - next week, next week.  I've just about given it up
	for a loss.  If anyone wants more information let me know.

	Otherwise, I'm still looking for a schematic capture/PCB
	layout/auto routing  package/programs to use.

    >>>>>>>>>>>>>>>>>>>My original post<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

	CAD/CAE tools are in wide spread use today. I lack the experience,
	time, and/or money to locate and evaluate those products/tools
	which can help me generate printed circuit boards.  I would like to
	find any public domain, or commercial software tools which support
	the printed circuit board development process. For example,
	Schematic Capture, Net List Generator, PCB Layout Editor,
	Auto-Router, etc.  I have a Compaq Portable, PC AT, or MicroVax II
	to run the tools on (prefer PC compatible).  Many of the commercial
	vendors offer demonstration copies which can be freely distributed.

	**I've purchased and attempted to use Advanced Microcomputer System's
	**Cell, Pcpro, and Pcroute tools.  The so called users manual is
	**inaccurate and incomplete.  So far I've provided AMS with a list of
	**at least 30 bugs to correct.  Several of them have cause complete
	**lockup of the system, and I've had to reboot.  The advice and 
	**instructions I've obtained over the phone, from AMS, have half 
	**the time been incorrect.  And, after all the trouble and hours of
	**frustration I went through I found out it doesn't route on single
	**layer boards, only 2 or more.

	If they could get rid of the bugs and make some of my suggested
	changes it would be a good product.  However, I'm now stuck without
	a way to generate several prototype PCBs, which I don't have time
	to do by hand.

	I would like to hear from anybody who can offer some advice, knows 
	where I can obtain this kind of software, or has experience
	with a particular system/package.  

	I would also like some advice on etching boards, wave soldering, 
	and pricing and availability of commercial services to
	do this kind of stuff (since these are prototype, maybe one-time
	boards I want to do as much as possible myself).

	Any input would be greatly appreciated.  I prefer any response 
	to come through mail. Thanks for taking the time.   

	Tim Coddington
	412-268-7712(w)
	412-244-8557(h)

	arpa==>  tac@sei.cmu.edu
	bitnet==>  tac%sei.cmu.edu@uma.cc.cmu.edu
	csnet==>  tac%sei.cmu.edu@relay.cs.net

	Disclaimer:  These are my opinions.

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
----------------------Mail I received------------------------------------

>Subject: PC Layout
>
>You might wish to try the Visionics product.  It does schematic
>capture, simple simulation, and layout.  Runs on a PC.  They
>have a new "greatly improved" version that will be available
>at the end of February.  I have yet to actually build a board
>that I have layed out with their software but I have been happy
>so far.
>
>
##############Name removed##############
>----------------------Mail I received-------------------------------
>Subject: Re: PCB design software
>
>
>Well, I'm only a hobbyist, what do I know for expensive systems?
>
>But, the above aside, I bought the EE-designer package and I am very pleased
>with it. It is, admittingly, in the low-end market ( i.e. don't build a service
>bureau on it), but for me it is just great. Background: I make projects;
>that means I design the device, I draw the schematics, I design the PCB and
>I debug the hardware. If needed, I write the software, too. So I'm not using
>the software all the time, maybe 20-30% of the time I use for electronics
>(wich gives about 5-10 hours a week).
>
>The software has schematic capture, simulation and PCB design (with autorouter)
>all integrated. Not very easy to learn, but easy enough. Not extremely powerful
>but powerful enough for professional work (if not for full-time PCB design).
>
>I recommend it, provided your need is not too heavy and cost has some
>meaning. (For the rich and heavy users, there exist much better systems; I
>know none with better cost effectiveness.)
>--
############Name removed########


>----------------------Mail I received---------------------------------
>Subject: Re: PCB design software
>
>
>
>The following is a mail-message in response to a previous more or less simular
>question (although the focus was different). Apart from the small remark I
>added at the end, I think this is all I can do for you. Succes !
>
>***********************************************
>>
>>Two years ago, at my previous job with Philips in the Netherlands, I became
>>heavily involved with selecting one of the then available CAD-software
>>packages for schematic capturing and pcb-design. As we were involved with the
>>design of switch mode power supplies, simulations were out of the
>>question (then at least). So no advice or experiences on the simulation field.
>>
>>Perhaps the first thing to do is to make a list of your prerequisitions;
>>things that such a CAD-tool absolutely should provide.
>>Scratch all alternatives that don't meet your prerequisitions. (tough if there
>>aren't any left :-) )
>>The rest is just a question of preferences and/or budget.
>>
>>Some of the most important items to us were:
>>- What to do with the output
  >>A perfect pcb design on screen or floppy doesn't help you a lot if there is
  >>not an appropriate way to either make the film(s) yourself or have them
  >>post-processed for you. The same applies for the ability to make drilling
  >>tapes.
  >>If you decide to have your designs postprocessed (which you are most likely
  >>to do) the first thing you have to do is standardise on pad sizes. Most CAD
  >>packages will support 32 or 64 pad codes. This might seem a lot, but be
  >>aware that every combination of pad size and drilling diameter requires a
  >>unique pad code. Especially when you plan to design mixed prints (i.e. both
  >>leaded and surface mount devices in the same library) you run out of pad
  >>codes quite easily. Changing the pad code reference file at the
  >>postprocessing facility for every new design is not impossible but very
  >>hard to manage.
>>- Our designs merely consisted of discretes. Therefore we had to be able to
  >>call a component during PCB design by its name, assigned during schematic
  >>capturing (e.g. R1 or C22). Amazingly, a number of packages only supported
  >>calling by type number (e.g. bc337) and/or library number. If you used more
  >>than one of such a component in your design, you never knew which one would
  >>be called.
  >>If you design circuits  using a lot of IC's having more than one gate,
  >>latch, flip-flop, opamp etc. on one chip, it is most likely you want to call
  >>the component by its type number. Than, a very important feature might be the
  >>possibility of gate swapping. Some packages don't support this at all, some
  >>only allow you to interchange both in- and outputs of the gates (or
  >>whatever) on one chip, some even support multiple chip gate swapping.
  >>If this is not clear: suppose you use two four-gate nand chips. During
  >>schematic capturing you have to connect all gates. During PCB design, you
  >>discover it would be much more convenient to use gate #3 of chip #2 for a
  >>specific function than gate #1 of chip #1. Interchanging all connections to
  >>those gates is called gate swapping. The system should know you are allowed
  >>to swap two nand gates (if using the same supply) but not a nand and o nor
  >>gate.
>>- Back annotation
  >>If using gate swapping (see above) or renaming components after having
  >>completed PCB design (useful for service purposes) you should change your
  >>schematics accordingly, in order to have a matching circuit diagram.
  >>As this is the reverse of the design direction, this is called back
  >>annotation. Depending on the nature of your designs this might be an
  >>important issue.
>>- The use of off-grid components
  >>Most packages enable you to design in a grid of 1/80 or 1/40 inch. If you
  >>use special components with metric dimensions this might be unacceptable.
  >>If this is important to you, make sure off-grid components are also
  >>supported.
>>- Variable track width
  >>Larger currents require wider tracks. Make sure you can already enter a
  >>track code at schematic capturing. There it is not to hard to indicate where
  >>higher currents might occur.
>>- True outline screen representation
  >>If you design a PCB and only see lines for the tracks and x's for the pads
  >>it is very hard to judge if another track can be routed trough a certain
  >>narrow path. This is even  harder if your designs contain a lot of different
  >>track widths.
>>- Design rule checking
  >>After completion of your PCB design, a design rule checker can check if 
  >>all design rules are met. These rules can concern *track to track distance
  >>*pad to track distance, *pad to pad distance etc. and can be very useful
  >>for a PCB design that is to be used in an production facility.
>>
>>Regarding the names you mention (pcad, orcad) I suppose you (as we did) are
>>looking for a CAD package that can be ran on an IBM-pc compatible machine.
>>If this is true make sure to run it on an AT (80286) instead of using a XT.
>>Always use a hard disk. Get yourself a decent hi-resolution screen, preferably
>>in colour. Use at least a 13 inch crt, but if you can afford it, bigger.
>>
>>In case you might be interested, we ended up buying Redcad from Racal Redac.
>>It was the only one that met all our requirements. We had a special interest
>>in the postprocessing part.
>>
>>The aspects involved with simulation were not evaluated by us, because only
>>logical simulators were available.
>>
>>Hope this has been of help !
>>

########Name removed########
>************************************
>End of original mail-message
>
>The only thing I can add concerning your specific questions is, that Redcad
>(our choise) also didn't support single layer autorouting. Neither am I aware
>of any software package that does. Two layer autorouting does not perform that
>well either. We didn't care; we hardly ever designed other than single layer
>boards.
>
>
#########Name removed###############
>>--------------------------------------------------------------------------


	Tim Coddington
	412-268-7712(w)
	412-244-8557(h)

	arpa==>  tac@sei.cmu.edu
	bitnet==>  tac%sei.cmu.edu@uma.cc.cmu.edu
	csnet==>  tac%sei.cmu.edu@relay.cs.net

	Disclaimer:  These are my opinions.