Hourly rate for PCB layout consultants...
Marc Reviel
We are budgeting for an employee, or a consultant, to do some basic PCB
layout work, such as SIMMs, DIMMs, etc. What is the average rate now?
I look forward to hearing anyone's experience, whether hiring it, or
charging for it. (...Anyone care to re-locate to Austin, Texas?)
Thank you.
---------------------------------------------------
Marc J. Reviel 928s CAN autocross
"SLOWER TRAFFIC KEEP RIGHT"
mre...@powerlogix.com Hardware Engineering
========= ========= =========
PowerLogix Corporation
"How Fast Do You Want To Go Today?"
http://www.powerlogix.com
---------------------------------------------------
Brian Denheyer
In article <3248EF...@i-link.net>,
Marc Reviel <mre...@i-link.net> writes:
> We are budgeting for an employee, or a consultant, to do some basic PCB
> layout work, such as SIMMs, DIMMs, etc. What is the average rate now?
Depends heavily on the type of work. If you are going to have them do
an 18x12 board with 14 layers expect >= $75/hr.
If it's a 2-layer board - probably not as much. Depends on geography
too.
> I look forward to hearing anyone's experience, whether hiring it, or
> charging for it. (...Anyone care to re-locate to Austin, Texas?)
Why should they have to re-locate ? How about getting the work done
from wherever and sending you gerbers via internet.
--
Brian Denheyer
bri...@northwest.com
Rene Zuidema
On 30 Sep 1996 03:11:20 GMT, bri...@northwest.com (Brian Denheyer)
wrote:
>In article <3248EF...@i-link.net>,
> Marc Reviel <mre...@i-link.net> writes:
>> We are budgeting for an employee, or a consultant, to do some basic PCB
>> layout work, such as SIMMs, DIMMs, etc. What is the average rate now?
>
>Depends heavily on the type of work. If you are going to have them do
>an 18x12 board with 14 layers expect >= $75/hr.
>
>If it's a 2-layer board - probably not as much. Depends on geography
>too.
It takes a *really* good PCB designer to do a single side EMC
compatible PCB! Consumer equipment must be cheap as possible while
still meeting requirements.
Our designer is very happy if he can -for once- use double sided
PCB's.
Adding layers simplifies matters, more room for ground planes etc.
Best Regards, Rene Zuidema <CPS-Europe>
Patrick Rea
Brian Denheyer wrote:
>
> In article <3248EF...@i-link.net>,
> Marc Reviel <mre...@i-link.net> writes:
> > We are budgeting for an employee, or a consultant, to do some basic PCB
> > layout work, such as SIMMs, DIMMs, etc. What is the average rate now?
>
> Depends heavily on the type of work. If you are going to have them do
> an 18x12 board with 14 layers expect >= $75/hr.
>
> If it's a 2-layer board - probably not as much. Depends on geography
> too.
>
> > charging for it. (...Anyone care to re-locate to Austin, Texas?)
>
> Why should they have to re-locate ? How about getting the work done
> from wherever and sending you gerbers via internet.
Although, I wouldn't mind escaping the snow here in Toronto this winter.
--
Patrick Rea 8888
Technical Services O-----88 888
Drastic Technologies 88 888
dra...@netrover.com 88 88----->
88 888
http://www.explorer.net/drastic O-----88 888
8888
Bob Wilson
Brian Denheyer (bri...@northwest.com) wrote:
: In article <3248EF...@i-link.net>,
: Marc Reviel <mre...@i-link.net> writes:
: > We are budgeting for an employee, or a consultant, to do some basic PCB
: > layout work, such as SIMMs, DIMMs, etc. What is the average rate now?
: Depends heavily on the type of work. If you are going to have them do
: an 18x12 board with 14 layers expect >= $75/hr.
: If it's a 2-layer board - probably not as much. Depends on geography
: too.
: > I look forward to hearing anyone's experience, whether hiring it, or
: > charging for it. (...Anyone care to re-locate to Austin, Texas?)
: Why should they have to re-locate ? How about getting the work done
: from wherever and sending you gerbers via internet.
This presumes the design is so layout insensitive that the PCB design
can be done in a vacuum. One of our biggest problems with contract
layout is that even with local layout people, we have problems with
weird component location and routing that may agree with the net list,
but can never work. Examples of this sort of thing are placing high
frequency decoupling caps half an inch or even more from RF devices, and
doing the usual digital thing of running traces willy nilly all over the
place (with the maximum amount of vias).
Having someone doing it by remote control over the internet would be
inviting disaster, unless the layout is a typical digital layout where
little knowledge of electrical criteria is needed. The other option, is
to spend an endless amount of time documenting absolutely every detail.
But then you might as well just do the layout yourself.
Bob.
Tony Hardman
In article <52p19m$g...@viking.mpr.ca>, Bob Wilson <bwilson@newshost>
writes
>inviting disaster, unless the layout is a typical digital layout where
>little knowledge of electrical criteria is needed. The other option, is
>to spend an endless amount of time documenting absolutely every detail.
>But then you might as well just do the layout yourself.
>
>Bob.
Hi
I have a feeling I will regret this, but I do not agree.
As you correctly suggested things go wrong when there is a six foot gap
between designer and the layout man.
Perhaps it is possible to have suitable documentation from the design
notes (which I think are essential to all manufactured and supported
equipment) whithout being the bug bear of all engineers who spend too
much time doing documentation.
What it does require is somebody who understands what the design is
about, and which are senstive stuff and all that.
I do like having reviews of progress at predefined times. I also like
the layout engineers to be involved with the design as early as
possible. I have had many good (if that is possible) meetings using a
conferance phone, and duplicated material. I suspect we should all be
running video conferancing soon which should (but probably not) help
with the communication issues.
So all i'm suggesting is that remote working is possible without being a
documentation nightmare, or a quality disaster. I do agree you have to
be careful, but that is life.
--
Tony Hardman
Bob Wilson
Tony Hardman (to...@f54x19.demon.co.uk) wrote:
: Hi
: I have a feeling I will regret this, but I do not agree.
: As you correctly suggested things go wrong when there is a six foot gap
: between designer and the layout man.
: Perhaps it is possible to have suitable documentation from the design
: notes (which I think are essential to all manufactured and supported
: equipment) whithout being the bug bear of all engineers who spend too
: much time doing documentation.
: What it does require is somebody who understands what the design is
: about, and which are senstive stuff and all that.
: I do like having reviews of progress at predefined times. I also like
: the layout engineers to be involved with the design as early as
: possible. I have had many good (if that is possible) meetings using a
: conferance phone, and duplicated material. I suspect we should all be
: running video conferancing soon which should (but probably not) help
: with the communication issues.
: So all i'm suggesting is that remote working is possible without being a
: documentation nightmare, or a quality disaster. I do agree you have to
: be careful, but that is life.
I cannot disagree with any of this. Especially the part about having
someone who generally understands what the design is all about. I have
seen far too many layout people who got into the business because they
liked to work with computers. Most of them haven't much of an idea about
the actual job they are doing; they are just good at making the tool
work. What is most necessary with a good layout guy is that he
understand the electrical basics of what he is laying out. Being good
with the tool is purely secondary. This would be no different than an
electrical engineer who could operate a spectrum analyzer but had little
idea of the principles behind the circuit he is analyzing.
We do a lot of analog and RF design, as well as some power supply design.
Perhaps in these areas especially, there is no way a typical cad jockey
can lay out a workable PCB without either some basic electrical
knowledge, or the electrical designer sitting beside him.
Bob.
Dale Wlasitz
Bob Wilson (bwilson@newshost) wrote:
: Brian Denheyer (bri...@northwest.com) wrote:
: : In article <3248EF...@i-link.net>,
: : Marc Reviel <mre...@i-link.net> writes:
: : > We are budgeting for an employee, or a consultant, to do some basic PCB
: : > layout work, such as SIMMs, DIMMs, etc. What is the average rate now?
: : Depends heavily on the type of work. If you are going to have them do
: : an 18x12 board with 14 layers expect >= $75/hr.
: : If it's a 2-layer board - probably not as much. Depends on geography
: : too.
: : > I look forward to hearing anyone's experience, whether hiring it, or
: : > charging for it. (...Anyone care to re-locate to Austin, Texas?)
: : Why should they have to re-locate ? How about getting the work done
: : from wherever and sending you gerbers via internet.
: This presumes the design is so layout insensitive that the PCB design
: can be done in a vacuum. One of our biggest problems with contract
: layout is that even with local layout people, we have problems with
: weird component location and routing that may agree with the net list,
: but can never work. Examples of this sort of thing are placing high
: frequency decoupling caps half an inch or even more from RF devices, and
: doing the usual digital thing of running traces willy nilly all over the
: place (with the maximum amount of vias).
: Having someone doing it by remote control over the internet would be
: inviting disaster, unless the layout is a typical digital layout where
: little knowledge of electrical criteria is needed. The other option, is
: to spend an endless amount of time documenting absolutely every detail.
: But then you might as well just do the layout yourself.
An experienced designer can easily do your board across town or in
another continent, about the same amount of communication needs to
take place.
|------------------------------------------------------------------------|
| BEHR Micro-Electronics Inc. Dale Wlasitz |
| Calgary, Alberta, Canada |
| e-mail: wlas...@cuug.ab.ca voice: (403) 239-8779 |
| web: http://www.cuug.ab.ca:8001/~wlasitzd fax: (403) 547-1507 |
|------------------------------------------------------------------------|
Larry G. Nelson Sr.
In article <52p19m$g...@viking.mpr.ca>, bwilson@newshost (Bob Wilson) wrote:
>Brian Denheyer (bri...@northwest.com) wrote:
>: In article <3248EF...@i-link.net>,
>: Marc Reviel <mre...@i-link.net> writes:
>: > We are budgeting for an employee, or a consultant, to do some basic PCB
>: > layout work, such as SIMMs, DIMMs, etc. What is the average rate now?
>
>: Depends heavily on the type of work. If you are going to have them do
>: an 18x12 board with 14 layers expect >= $75/hr.
>
>: If it's a 2-layer board - probably not as much. Depends on geography
>: too.
>
>: > I look forward to hearing anyone's experience, whether hiring it, or
>: > charging for it. (...Anyone care to re-locate to Austin, Texas?)
>
>: Why should they have to re-locate ? How about getting the work done
>: from wherever and sending you gerbers via internet.
>
>This presumes the design is so layout insensitive that the PCB design
>can be done in a vacuum. One of our biggest problems with contract
>layout is that even with local layout people, we have problems with
>weird component location and routing that may agree with the net list,
>but can never work. Examples of this sort of thing are placing high
>frequency decoupling caps half an inch or even more from RF devices, and
>doing the usual digital thing of running traces willy nilly all over the
>place (with the maximum amount of vias).
>
>Having someone doing it by remote control over the internet would be
>inviting disaster, unless the layout is a typical digital layout where
>little knowledge of electrical criteria is needed. The other option, is
>to spend an endless amount of time documenting absolutely every detail.
>But then you might as well just do the layout yourself.
>
familiar with the electronics and not just a high end tape jockey. I have
several friends that do excellent layout work but do not know electronics.
Those that know the design end get more money but can see the problem areas of
the circuit and worse case need a page or two of notes from the principle
design engineer as to the critical areas.
I have done this type of work at times and bid on the jobs submitted to me
knowing where most of the problem areas are.
Larry G. Nelson Sr.
L.Ne...@ieee.org
http://www.ultranet.com/~nr
Marc Reviel
Brian Denheyer wrote:
>
> In article <3248EF...@i-link.net>,
> Marc Reviel <mre...@i-link.net> writes:
> > We are budgeting for an employee, or a consultant, to do some basic PCB
> > layout work, such as SIMMs, DIMMs, etc. What is the average rate now?
>
> Depends heavily on the type of work. If you are going to have them do
> an 18x12 board with 14 layers expect >= $75/hr.
>
> too.
>
> > charging for it. (...Anyone care to re-locate to Austin, Texas?)
>
> Why should they have to re-locate ? How about getting the work done
> from wherever and sending you gerbers via internet.
For our needs, as an *employee*, relocation would be required.
Kent Farnsworth
Marc:
This thread is interesting, in that I am Chief Engineer at a company
in Oak Hill. You should be familiar with that area of Austin. I have
hired a layout person, and the results are the same as when I used
outside/remote people. It's like "I told you to put this list of
traces on the ground plane", or, "Why did you put this chip clear
across the board, when I told you that the speed is critical!". I have
always prepared a list of critical areas, though. It doesn't take long
to generate, and is mandatory, in my opinion. I generally create
critical lists in less than an hour (with 20 pages of schematics).
I know where you are coming from. If you have a local or remote
designer that you can talk to, and fax/email/etc. the current HP/GL or
postscript files on a regular basis, then it really doesn't matter
where the guy is. It's faster if you are 'over his shoulder', but you
can beat him up up close, or remotely. It's the same. Obviously the
goal is your design interpretation and it's implementation.
Kent
Roy McCammon
Bob Wilson wrote:
> We do a lot of analog and RF design, as well as some power supply design.
> Perhaps in these areas especially, there is no way a typical cad jockey
> can lay out a workable PCB without either some basic electrical
> knowledge, or the electrical designer sitting beside him.
The way we do it is that we expect our pcb design technicians to know the
tools (schematic capture, pcb layout, data base systems for partrs lists),
the factory's requirements (such and such spacing, no board bigger than
X by Y, traces no closer than W), and the process (you have to have an artwork
for each layer of copper, one for the silkscreen, upper solder mask, lower
solder mask, certain tooling holes, certain cad formats, pcb fabricator
capabilities). The engineer that designs the circuit is responsible for it
working correctly. We generally schedule 1 or 2 weeks for him to spend time
with the design tech.
For a purely digital board, it may be sufficient to specify the stack up
(two power planes, two ground planes, 4 signal layers), placement of the
major components (cpu there, memory here, connectors there) and help route
the clock lines.
For mixed analog and digital, I generally do all the above, specify the
design of the power and ground planes, design every summing junction, and place
the guard rings.
Here is a useful artifice:
I once had several design techs that had trouble keeping the high speed digital
signals away from the sensitive analog. I finally had them route a ground trace
across every layer, one above the other, so that no trace could cross that
barrier. After all the other circuits were in place, I let them remove the barrier
in selected places to complete the circuit. It worked well, recommended.
Once and only once, did I let the design tech do the design without my supervision.
I got an incredible mess of ground loops and digital busses woven in with summing
junctions. We had to scrap the whole thing. I took responsibility for the mess
(I deserved it). We all learned that nothing beats good communication and mutual
respect.
Opinions expressed herein are my own and may not represent those of my employer.
AbdulraHman Lomax
bwilson@newshost (Bob Wilson) wrote:
>This presumes the design is so layout insensitive that the PCB design
>can be done in a vacuum. One of our biggest problems with contract
>layout is that even with local layout people, we have problems with
>weird component location and routing that may agree with the net list,
>but can never work.
It helps if the designer does know at least a bit about things like
trace impedance and the principles of noise control.
But this is not really a question of whether or not the designer is
physically close to the engineer, not since it became possible to send
files by modem and now internet. Rather, it is a question of designer
skill as well as the ability of the engineer to anticipate possible
problems and to give special instructions.
>Examples of this sort of thing are placing high
>frequency decoupling caps half an inch or even more from RF devices, and
>doing the usual digital thing of running traces willy nilly all over the
>place (with the maximum amount of vias).
One only sees that kind of design with either (1) a radically
inexperienced designer, or (2) an autorouted design, with bad
placement, or (3) both. One question I always ask customers, if it is
not obvious from the schematic, is whether or not the design is layout
sensitive. If it is complex and insensitive, then I might autoroute
the board. After all, it will save the customer a good chunk of cash.
The idea mentioned in another post that somehow multilayer design
would command a higher design hour rate is not new; but it has always
puzzled me. If a board is higher density, it is more work, so it is
more expensive to do, but, other things being equal, multilayer design
is easier than 2-sided or 1-sided design. For me the most common board
is 4-layer and so, with those boards, there is almost zero worry about
power routing. What can make design more difficult is surface mount
and the complexities it introduces. But, still, one pays according to
the experience (and chutzpah!) of the designer, not according to how
many layers the design will use or how difficult it is. A good
designer may be able to reduce the number of layers....
>Having someone doing it by remote control over the internet would be
>inviting disaster, unless the layout is a typical digital layout where
>little knowledge of electrical criteria is needed. The other option, is
>to spend an endless amount of time documenting absolutely every detail.
>But then you might as well just do the layout yourself.
I do complex analog/digital boards over the internet. I live in the
Blue Ridge mountains of Western North Carolina, but I moved here from
California. My customers are still all in California. My business
*increased* when I moved away. There has been no problem communicating
with customers.
When I was in California, a few customers would be tempted to ask me
to come in for a conference with the engineer. My general experience
was that nothing was accomplished in the conference that would not
have been better accomplished with written instructions plus a chat on
the phone. I make sure all my customers have the Tango demo package,
and I will sometimes send them a preliminary placement. I very rarely
have an engineer come back with problems about placement.
One of my customers is a very, very picky RF engineer. I used to go
and design boards at his house. It took longer and the results were no
better than what can be done remotely.
Yes, if a board is sensitive and the designer does not understand
that, and the designer does a week's worth of work without showing it
to the engineer, this is a formula for disaster. This is true whether
or not the designer is in the next cubicle or the next state or even
in the next country.
Daniel Lomax
Trace Engineering
(704) 274-5900, fax 274-0059
email mar...@ioa.com
Dave Baldwin
Bob Wilson (bwilson@newshost) wrote:
: Having someone doing it by remote control over the internet would be
: inviting disaster, unless the layout is a typical digital layout where
: little knowledge of electrical criteria is needed. The other option, is
: to spend an endless amount of time documenting absolutely every detail.
: But then you might as well just do the layout yourself.
I suppose for larger projects, this might be true. I do layout work for
small companies from California to New York and I find myself having to
explain to them what's important in the layouts and the circuits. Showing
up onsite would mostly be a waste of time.
--
-=-=-=-=-=-=-=-=-=-=-=- Check out DIBs and TCJ -=-=-=-=-=-=-=-=-=-=-=-=-
Dave Baldwin: dib...@netcom.com | The Computer Journal 1(800)424-8825
DIBs Electronic Design | Home page "http://www.psyber.com/~tcj/"
Voice : (916) 722-3877 | Hands-on hardware and software
TCJ/DIBs BBS: (916) 722-5799 | TCJ/DIBs FAX: (916) 722-7480
-=-=-=-=-=-=- @#$%^&* I can't even quote myself! Oh,well. -=-=-=-=-=-=-
Alan Baker
Tony Hardman <to...@f54x19.demon.co.uk> said through the grapevine:
>In article <52p19m$g...@viking.mpr.ca>, Bob Wilson <bwilson@newshost>
>writes
>>Having someone doing it by remote control over the internet would be
>>inviting disaster, unless the layout is a typical digital layout where
>>little knowledge of electrical criteria is needed. The other option, is
>>to spend an endless amount of time documenting absolutely every detail.
>>But then you might as well just do the layout yourself.
>>
>>Bob.
>
>Hi
>I have a feeling I will regret this, but I do not agree.
>
>As you correctly suggested things go wrong when there is a six foot gap
>between designer and the layout man.
>
>Perhaps it is possible to have suitable documentation from the design
>notes (which I think are essential to all manufactured and supported
>equipment) whithout being the bug bear of all engineers who spend too
>much time doing documentation.
>
>What it does require is somebody who understands what the design is
>about, and which are senstive stuff and all that.
>I do like having reviews of progress at predefined times. I also like
>the layout engineers to be involved with the design as early as
>possible. I have had many good (if that is possible) meetings using a
>conferance phone, and duplicated material. I suspect we should all be
>running video conferancing soon which should (but probably not) help
>with the communication issues.
>
>So all i'm suggesting is that remote working is possible without being a
>documentation nightmare, or a quality disaster. I do agree you have to
>be careful, but that is life.
>
>--
>Tony Hardman
If the guy at the other end of the link is also a electronics engineer
I see no problem whatsoever, its just a question of experience I guess
and well worth a try as it would be a darn sight cheaper than the 75
dollar am hour someone quoted, me I do it for fifty cents a pad and
provide full engineering drawings and Gerber files with Excellon drill
files and if one has access to the shareware GCPREVUE you can view the
resulting PCB. So its horses for courses and just a question of
picking the right designer and the internet is ideal for transfering
the completed files.
--
73 Alan
Alan Baker, G3KFN
Plymouth, England
email: al...@g3kfn.avel.co.uk
Simon Peacock
In message <52p19m$g...@viking.mpr.ca> - bwilson@newshost (Bob Wilson)30 Sep
1996 17:50:46 GMT writes:
:>This presumes the design is so layout insensitive that the PCB design
:>can be done in a vacuum. One of our biggest problems with contract
:>layout is that even with local layout people, we have problems with
:>weird component location and routing that may agree with the net list,
:>but can never work. Examples of this sort of thing are placing high
:>frequency decoupling caps half an inch or even more from RF devices, and
:>doing the usual digital thing of running traces willy nilly all over the
:>place (with the maximum amount of vias).
:>Having someone doing it by remote control over the internet would be
:>inviting disaster, unless the layout is a typical digital layout where
:>little knowledge of electrical criteria is needed. The other option, is
:>to spend an endless amount of time documenting absolutely every detail.
:>But then you might as well just do the layout yourself.
Well actually, no. A good designer takes into account the type of work being
done. In the same regards a RF designer won't nessasarly do a good digital
design and a digital designer generally won't do a good RF design. Thats why
we haveboth here.
You must know your limits.
Good layout comes from experience.
"The difference between an educated man and an uneducated man is that an
educated man knows that he knows what he dosen't know, but an uneducated man
doesn't know that he dosen't known what he doesn't know."
--
Simon Peacock Telephone: +64-4-388-8964 (home)
si...@digitech.co.nz +64-4-389-8909 (work)
si...@actrix.gen.nz Fax: +64-4-389-9901 (work)
Home address: 5 Countess close, Mapuia, Wellington, New Zealand
Work address: 102-112 Daniell Street, Newtown
(PO Box 20-002, Wellington South)
Wellington, New Zealand
Wen-King Su
In a previous article wlas...@cuug.ab.ca (Dale Wlasitz) writes:
<: Having someone doing it by remote control over the internet would be
>: inviting disaster, unless the layout is a typical digital layout where
<: little knowledge of electrical criteria is needed. The other option, is
>: to spend an endless amount of time documenting absolutely every detail.
<: But then you might as well just do the layout yourself.
>
<An experienced designer can easily do your board across town or in
>another continent, about the same amount of communication needs to
<take place.
If the engineer has to know enough about board design so that he can list
all the constraints in a clear and concise format so it can be related
to the designer with a minimal amount of interaction, why do you even
need a designer? The engineer may just as well feed it to one of the
commercial cad tools himself. We have used outside designers before, and
now that we are doing it in house, we are much happier with the result.
Brian Land
To all following this thread:
What are some good PCB layout software packages??? We have PADS at work and I refuse to
use it because it is such an awful program. (I will spare you my rant because I really
hate that program)
I would like to find some software that has excellent layout capabilities, with the
ability to easily generate and manage libraries of custom artwork.
The finished PCB is eventually mounted somewhere, so we also need to generate an
accurate mechanical drawing for the completed board/layout. I would need the ability to
dimension and annotate the board outline and the copper pad sizes and locations. (If the
software can reliably create .DXF or .DWG files for Autocad, then that would be ideal)
The software would need to run on a PC platform.
Any suggestions?
Who makes Protel PCB?
Brian
n...@junk.thanks.com wrote:
>
> But the biggest problem is that instructions to do such-and-such are
> often ignored. I used to sub-contract all my boards, until the man
> made a big error in an area where I specifically told him to be
> careful.
>
> This prompted me to buy Protel PCB, and I have never looked back.
>
>
Robert K. Johnson
On Thu, 03 Oct 1996 21:58:59 -0700, Brian Land <bl...@ix.netcom.com>
wrote:
PRotel is the GREATEST...
_____________________________________________________
Robert K. Johnson
r...@earthlink.com
Oliver Bartels
In <325499...@ix.netcom.com>, Brian Land <bl...@ix.netcom.com> writes:
>To all following this thread:
>
>What are some good PCB layout software packages??? We have PADS at work and I refuse to
>use it because it is such an awful program. (I will spare you my rant because I really
>hate that program)
(Marketing On)
Of course, this opinion is biased, but: You should have a look at our home page,
the BAE software includes several drawing utilities plus DXF IN/OUT (ULC source!)
plus an object oriented database plus an own programming language (ULC) for
special problem solutions as well as the Bartels Neural Router with things like
P/G-swap during rip-up routing. Different versions for Win, Win NT, HP-UX and other
Unixes as well as good old 386/DOS/DPMI are available. An earlier DOS demo is
available for download, the full set is available on a CD (on EMail, may take a few
days as the new NT demo is currently with 1000 CD's in production). The software
is used by major European accounts for things like GSM base station PCB designs.
(Marketing Off)
Biggest disadvantage: Made in Europe, "not invented here", no hillybilly marketing ;-)
Greetings Oliver
------------------------------------------------------
Oliver Bartels + Erding, Germany + obar...@bartels.de
http://www.bartels.de + Phone: +49-8122-9729-0 Fax:-10
------------------------------------------------------
Brian Denheyer
In article <52p19m$g...@viking.mpr.ca>,
bwilson@newshost (Bob Wilson) writes:
> This presumes the design is so layout insensitive that the PCB design
> can be done in a vacuum. One of our biggest problems with contract
> layout is that even with local layout people, we have problems with
> weird component location and routing that may agree with the net list,
> but can never work. Examples of this sort of thing are placing high
> frequency decoupling caps half an inch or even more from RF devices, and
> doing the usual digital thing of running traces willy nilly all over the
> place (with the maximum amount of vias).
Of course, the idea is to hire someone with RF layour experience if
you are going to do RF layouts. Also, _good_ schematic capture/layout
programs allow you to put in important details like :
this net can't be more than xxx long
or trace width must be =x
You can also give the layout person a placement, and then let them do
the routing.
>
> Having someone doing it by remote control over the internet would be
> inviting disaster, unless the layout is a typical digital layout where
> little knowledge of electrical criteria is needed. The other option, is
> to spend an endless amount of time documenting absolutely every detail.
> But then you might as well just do the layout yourself.
>
> Bob.
Yeah, usually you don't need much experience to do a digital layout
unless you have 32 data lines and 30 addr lines running all over the
board with 5ns rise times and you want to pass FCC - the _first_ time. :-)
I completely disagree that it is "inviting" disaster. It is entirely
possible to accomplish a design remotely with a layout person. I have
done it and seen it done. I do believe that it will take longer.
--
Brian Denheyer
bri...@northwest.com
Marty Goldgrub
> One of my customers is a very, very picky RF engineer. I used to go
> and design boards at his house. It took longer and the results were no
> better than what can be done remotely.
>
> Yes, if a board is sensitive and the designer does not understand
> that, and the designer does a week's worth of work without showing it
> to the engineer, this is a formula for disaster. This is true whether
> or not the designer is in the next cubicle or the next state or even
> in the next country.
>
> Daniel Lomax
> Trace Engineering
> (704) 274-5900, fax 274-0059
> email mar...@ioa.com
Of all the information that has come on this thread, there has been no
discussion of constraint driven design. Where the constraints drive the
designeers and the design. Look at the combination of Unicad( Unisolve
) & CCT ( sppectra ). When I get designs from companies I am able to go
back to the engineer after placement and tell him if the I can meet the
constrains or where my problems are. Layout is no longer an over the
wall process, but one must use the right tools!
Marty Goldgrub
Boardmaster Engineering
Michael R. Kesti
mis-spelt! wrote:
>If you do boards yourself, *you* can choose the tool which you believe
>will be most future-proof, rather than the one which you have been
>using for X years.
And, in my opinion, you run the same risk of being wrong in this attempt
to "tell the future" as does the contract designer.
>And if you do a lot of boards, there is absolutely no argument in
>favour of using an external designer.
Of course there is, and it's the one that drives companies to do so.
It's cheaper, in the long run, as you don't have to pay benefits to
a person whose service you purchase rather than employ. Also, you
never run the risk of having to pay that person during the occasional
slack time during which he has little or nothing to do.
--
========================================================================
Michael Kesti | "And like, one and one don't make
| two, one and one make one."
mke...@netshel.net | - The Who, Bargain
Bas van Rossem
scan...@dial.pipex.com (mis-spelt!) wrote:
>>I completely disagree that it is "inviting" disaster. It is entirely
>>possible to accomplish a design remotely with a layout person. I have
>>done it and seen it done. I do believe that it will take longer.
>IMO the biggest problem in using an external designer is what happens
>in the future.
>Example: one designer I used to use, for about 10 years, had a system
>which became totally obsolete one day. And he had no means of
>extracting his database to another system. So all the old designs were
>lost, with no possibility of doing updates, other than via the
>torturous gerber file route.
>Then he got another system, PC-based this time, called Boardmaker. It
>was very limited, e.g. text height could be only in 0.025" steps - no
>good for dense boards. He did good boards with it, as far as he could,
>but eventually the above limitation became a problem.
>If you do boards yourself, *you* can choose the tool which you believe
>will be most future-proof, rather than the one which you have been
>using for X years.
I do not consider this to be a valid argument. Any chosen system
might eventually become obsolete. Even the computers on which the
software is being used are 'obsolete-prone', and how about the
operating system?
Take the PC, who can predict what the most popular operating system
will be in a few years time? Unix, MSDOS, Windows, Windows 95 or
Windows NT or maybe even an OS2 come-back?
The makers of CAD packages are forcing you into 'file-conversions' on
every major update. Did you ever convert ALL your old files on each
and every update?
Demanding that a certain package has to be used might be a valid
point. Personally I would expect YOU to deliver that software. If you
want a low priced design you cannot expect the designer to go out and
buy the software you prefer without you paying for it.
>And if you do a lot of boards, there is absolutely no argument in
>favour of using an external designer.
True, but even then you sometimes need some extra help or extra
specialized knowledge.
As others have already stated, I think that a PCB designer should
understand the circuit he is designing the PCB for. If not, he could
better be replaced with a super-dooper autorouter; much cheaper in the
end and the results are just as poor.
Just my $0.01
Bas
The beauty
AbdulraHman Lomax
scan...@dial.pipex.com (mis-spelt!) wrote:
>IMO the biggest problem in using an external designer is what happens
>in the future.
>Example: one designer I used to use, for about 10 years, had a system
>which became totally obsolete one day. And he had no means of
>extracting his database to another system. So all the old designs were
>lost, with no possibility of doing updates, other than via the
>torturous gerber file route.
This has nothing to do with using an outside designer vs. doing the
design oneself. Rather it is a problem with the obsolescence of design
software. It is, in fact, more likely to bite the engineer who designs
relatively few boards than it is a designer who is active with many
designs and many customers: the designer is highly motivated to keep
his existing customers and will find a way to solve the problem
In my case, when I went from one design system to another, I wrote a
translator. (This was from Pcpro to Tango). I also wrote a translator
between Protel and Tango.
Further, it is very easy to import gerber to Tango, and I would think
that it would not be difficult to import it into other programs. You
have to replace the components, but it simply is nowhere as difficult
to import Gerber as it is to redo a design.
I would not dream of using a PCB CAD program that did not have a
database which I could manipulate directly.
>Then he got another system, PC-based this time, called Boardmaker. It
>was very limited, e.g. text height could be only in 0.025" steps - no
>good for dense boards. He did good boards with it, as far as he could,
>but eventually the above limitation became a problem.
This does not encourage me to think that this was a serious designer,
especially when Tango sold for $600, and there are other even cheaper
programs without these limitations.
>If you do boards yourself, *you* can choose the tool which you believe
>will be most future-proof, rather than the one which you have been
>using for X years.
And if *you* are not focused primarily on PCB design tools, what would
make you think that you would make a better decision in this respect
than someone who makes his (or her) entire living from PCB design?
>And if you do a lot of boards, there is absolutely no argument in
>favour of using an external designer.
One designer can serve, in general, I would say, three or four
engineers. Now, can you afford to hire a competent designer? If so,
fine. But you will also have to provide this designer with tools and
space, not to mention supervision and benefits. How many designs is "a
lot?" Is it really enough to justify this expense?
Of course, an engineer may say, "I just design them myself." Fine. And
I can also repair my car myself, and diagnose my own illnesses, and
fix the plumbing, and repair the copy machine, and write all my own
word processing software, not to mention PCB design software. Some
time back I decided that, while I was perfectly capable of doing all
these things, I was not likely to do *all* of them as well as someone
who does it full-time. And even if I could do it as well (or maybe
even better, in some cases), I would certainly not do it with
facility.
I had a friend one time who wanted to publish a few books. He did not
like the quotes he got for printing, and since he had access to quite
a bit of cash, he bought a printing press. Bad move. Printing is a
highly competitive business, and you can buy the services of a highly
competent printer, with state-of-the-art equipment, for much less than
what it will cost you to do the job, *unless* you are prepared to make
it a full-time occupation, and even then you might go belly-up.
There is are actually only a few justifications for *not* using a
design service (or an individual outside designer, which amounts to
the same thing):
(1) You are a startup engineer, and you simply can't afford to pay
what design is worth, and you have the time to battle all the bugs and
irritations of doing it yourself.
(2) You can think of nothing better to do than route a thousand traces
from point A to point B.
(3) Your company does enough design to warrant engaging a full-time
designer.
What often happens is that a company decides that draftsman X, who
already does all their mechanical work, might as well learn to do PC
design. While it is possible that draftsman X might indeed have a
talent for it, so this cannot be ruled out, it is also quite possible
that one will end up with the kind of horror stories that we read
earlier in this thread.
On the other hand, an engineer who turns a designer loose with a
complex, highly-constrained design, without specifying the constraints
and without looking at a placement before the board is routed, has
asked to be presented with a nightmare. If this is an in-house
designer, your company eats it. If it is an outside designer,
hopefully you asked to see a placement before it was routed, hopefully
you specified the constraints. If you did, and the designer ignored
them, you should not have to pay to fix or redo the board.
AbdulraHman Lomax
no.jun...@thanks.com (Peter) wrote:
>>Demanding that a certain package has to be used might be a valid
>>point. Personally I would expect YOU to deliver that software. If you
>>want a low priced design you cannot expect the designer to go out and
>>buy the software you prefer without you paying for it.
>This I agree with, but I have found that the man refuses to learn new
>tools.
The problem may well be "the man," not the principle of in-house vs.
outside design, or this package or that package.
I'm pretty comfortable with Tango PCB Plus, but if a customer offered,
for example, to set me up with PCad, or some other reasonable package,
I'd go for it. I might even not charge for the time spent learning a
new package; after all, it would increase my potential customer base.
I did, in fact, do this when a customer requested Tango design and
offered me enough work to justify the change. Once I used Tango (which
I paid for, not the customer), I never looked back. I did write a
translator when I had some time.
On the other hand, if I was overwhelmed with work, using a package
with which I was comfortable -- which ought to translate to
"efficient" -- then I would not be too eager to waste my time.
Don Yuniskis
In article <326999e7...@news.u-net.com>,
Peter <no.jun...@thanks.com> wrote:
>>There is are actually only a few justifications for *not* using a
>>design service (or an individual outside designer, which amounts to
>>the same thing):
Sheesh! Obviously we work on quite different types of projects!
I've yet to meet a "design service" that is clever enough to look
at a GAL and be able to decide which pins *could* safely be swapped
with a minor tweek to the internal logic. Or, when it might be
preferable to put a group of 1/4W resistors in place of a big
2W resistor to avoid a mechanical interference problem.
Note I'm already *assuming* the service bureau can handle things
like gate swapping from package to package, scrambling the address
lines to an SRAM or EPROM (!), etc. *And* I'm assuming all of the
mechanical constraints, keep outs, etc. have been *diligently*
documented before the service bureau gets their hands on the layout...
Unfortunately, not all of us work in nice cookie-cutter environments
where we can walk away from a design during layout. There is often quite
a lot of "value added" to a design that is rooted in the *designer's*
head and would be quite difficult to put on paper. Indeed, it would be
quite difficult to even prescribe *what* should be written down for
conveyance to the service bureau.
>>What often happens is that a company decides that draftsman X, who
>>already does all their mechanical work, might as well learn to do PC
>>design. While it is possible that draftsman X might indeed have a
>>talent for it, so this cannot be ruled out, it is also quite possible
>>that one will end up with the kind of horror stories that we read
>>earlier in this thread.
>
>But this is usually what happens, here in the UK.
However, even a "draftsman" laying down tape and rubilyth (sp?)
can get the job done if the designer is readily available. If,
on the other hand, the designer has to drive across town to deal
with the service bureau, it quickly becomes far too expensive
to farm out layouts (since a trip *anywhere* has to take at least
an hour plus bullshitting time, etc.). Bosses tend to frown
on designers that "have to go solve a problem at the service bureau"
>>On the other hand, an engineer who turns a designer loose with a
>>complex, highly-constrained design, without specifying the constraints
>>and without looking at a placement before the board is routed, has
>>asked to be presented with a nightmare. If this is an in-house
>>designer, your company eats it.
>
>This is quite true, but at least you can see problems *while* they are
>developing, not when the board is finished and the invoice from the
>designer is in the post.
I think the far more significant benefit of the in-house approach
is that you don't lose all the TIME before realizing you're in
trouble (I assume these comments were focussed on the MONEY)
--don
AbdulraHman Lomax
Peter,
no.jun...@thanks.com (Peter) wrote:
[I wrote:]
>>On the other hand, an engineer who turns a designer loose with a
>>complex, highly-constrained design, without specifying the constraints
>>and without looking at a placement before the board is routed, has
>>asked to be presented with a nightmare. If this is an in-house
>>designer, your company eats it.
>This is quite true, but at least you can see problems *while* they are
>developing, not when the board is finished and the invoice from the
>designer is in the post.
It is apparent that Peter has had some bad experiences with outside
designers. In fact, there are quite a few of us who got into this
business without much study and qualification; sometimes it seems that
any beginning draftsperson with a PC and a copy of a CAD program is
ready to set him or herself up as a "design consultant."
However, the difference between in-house and outside design is pretty
artificial. One can check the progress of an outside design pretty
easily. Whenever there is the slightest doubt about a placement, for
example, I send the board to my customers before going ahead with
routing. I have never charged for changing the placement. When I do RF
designs, sometimes the board goes out several times during the
process, with one section being routed and then another.
I have known many engineers who did their own board design. I even
know the president of one fairly large company who does it. My guess
is that those are about the most expensive designs in their class in
this country; this guy's time is probably worth ten times mine, or
more.
There really is very little special about P.C. board design such that
a competent electronics engineer should not be able to do it, and do
it well, *if* he or she wants to spend the time. The only little
exception is that electronics engineers often don't know much about
manufacturability, but, then again, judging from some of the layouts I
have been given to clean up, neither do many designers, and, of
course, an engineer can learn about this pretty quickly if it is
desired.
I am in business for those engineers who either lack the time or
inclination to do their own work, or who can't afford an in-house
designer (or the in-house designers are too busy), or who need my
specialized experience.
AbdulraHman Lomax
mar...@ioa.com
AbdulraHman Lomax
d...@rtd.com (Don Yuniskis) wrote:
>Peter <no.jun...@thanks.com> wrote:
{Bad rap. I actually wrote what is quoted:]
>>>There is are actually only a few justifications for *not* using a
>>>design service (or an individual outside designer, which amounts to
>>>the same thing):
>Sheesh! Obviously we work on quite different types of projects!
From what is said below, this is not true. But we *do*, obviously work
in very different ways.
>I've yet to meet a "design service" that is clever enough to look
>at a GAL and be able to decide which pins *could* safely be swapped
>with a minor tweek to the internal logic.
I have done many projects where there was communication on these
points. No, I don't decide what can be moved. Usually, I just design
these kinds of parts as they are given to me. But if there is a reason
to move them, such as the customer wants a two-layer design and there
is not a lot of space, I ask, and am told, what I can move. In one
case I was told to give the customer a preliminary pin assignment and
they would see if it could be routed internally. They did this; as I
recall, they came back with one small change.
>Or, when it might be
>preferable to put a group of 1/4W resistors in place of a big
>2W resistor to avoid a mechanical interference problem.
Obviously, a designer can only work with what he or she is given. If
there are mechanical restraints, I need to know them -- preferably in
advance! This is true for any designer, outside or in-house. As to the
example given, I have often suggested such changes to customers. I
fail to understand why this would be a particularly difficult problem.
It seems that these engineers have been the victims of some poor
choices in design consultants.
>Note I'm already *assuming* the service bureau can handle things
>like gate swapping from package to package, scrambling the address
>lines to an SRAM or EPROM (!), etc.
Yes, we routinely do that. Except for the EPROMs. Because EPROMS are
generally programmed on another piece of equipment, they should not
have nonstandard address and data lines. I do remember the first time
I suggested to an engineer that we might scramble the address and data
lines to a RAM. It took him a few minutes to realize the implications;
of course, he did give permission once he thought about it.
>*And* I'm assuming all of the
>mechanical constraints, keep outs, etc. have been *diligently*
>documented before the service bureau gets their hands on the layout...
If we have those constraints, then we are capable of making those
resistor changes. All these problems are really the problems involved
in working with other people, not the problems of inside vs. outside.
Actually, working with outside designers can have the effect of
forcing a little discipline in what is communicated. It is a little
easier to imagine that the designer in the next office is a
mind-reader than the one across town, or across the United States, in
my example. At the present time, my closest customer is 3000 miles
away.
>Unfortunately, not all of us work in nice cookie-cutter environments
>where we can walk away from a design during layout. There is often quite
>a lot of "value added" to a design that is rooted in the *designer's*
>head and would be quite difficult to put on paper. Indeed, it would be
>quite difficult to even prescribe *what* should be written down for
>conveyance to the service bureau.
Or to the in-house designer. So the company ends up paying the
designer *and* the engineer to do the same job. If it is truly as
difficult as Don is saying (and, yes it can be difficult), then it
would actually be better for Don to ask the designer to set up all the
parts and let him place and route them, and then -- maybe -- let the
designer check over the design for the kinds of things that a designer
might catch that the engineer would miss.
>>>What often happens is that a company decides that draftsman X, who
>>>already does all their mechanical work, might as well learn to do PC
>>>design. While it is possible that draftsman X might indeed have a
>>>talent for it, so this cannot be ruled out, it is also quite possible
>>>that one will end up with the kind of horror stories that we read
>>>earlier in this thread.
>>
>>But this is usually what happens, here in the UK.
>However, even a "draftsman" laying down tape and rubilyth (sp?)
>can get the job done if the designer is readily available. If,
>on the other hand, the designer has to drive across town to deal
>with the service bureau, it quickly becomes far too expensive
>to farm out layouts (since a trip *anywhere* has to take at least
>an hour plus bullshitting time, etc.). Bosses tend to frown
>on designers that "have to go solve a problem at the service bureau"
I'm glad that Don mentioned bullshitting time. I have found that my
productivity -- and the ultimate satisfaction of my customers -- has
*increased* with increased distance from the customers. When I lived a
few minutes away from my customers, there was occasionally the
temptation to visit them, or for them to visit me. But even then, when
I was close, I learned that a modem was much faster than my car, and
when we worked by phone and fax and modem, and, more recently, e-mail,
there was much more of a tendency to stay to business.
>>>On the other hand, an engineer who turns a designer loose with a
>>>complex, highly-constrained design, without specifying the constraints
>>>and without looking at a placement before the board is routed, has
>>>asked to be presented with a nightmare. If this is an in-house
>>>designer, your company eats it.
>>
>>This is quite true, but at least you can see problems *while* they are
>>developing, not when the board is finished and the invoice from the
>>designer is in the post.
>I think the far more significant benefit of the in-house approach
>is that you don't lose all the TIME before realizing you're in
>trouble (I assume these comments were focussed on the MONEY)
If an engineer is truly concerned about the progress of a design, it
is quite feasible for the engineer to see the design several times a
day or even more often. Once again, if a company has the money for
in-house staff, it may be quite wise to have an in-house designer. But
sometimes this designer is busy on another project, or, sometimes, the
design would benefit from, say, RF design experience.
If time is truly one's concern (and it usually is a prime concern),
then I would recommend a competent in-house designer PLUS several
outside consultants. You want several because when you need your
project, if you have only one, he or she may be busy with something
else. The point is to identify and qualify these outside designers
*before* you have that rush project.
In fact, if a project involves more than one board, I have seen many
companies split up the work among several designers. Somebody, of
course, has to coordinate the work, but this is, in fact, the fastest
way to market, though it can be a bit more expensive.
AbdulraHman Lomax
mar...@ioa.com
Carl Wuebker
AbdulraHman Lomax (mar...@ioa.com) wrote:
> However, the difference between in-house and outside design is pretty
> artificial. One can check the progress of an outside design pretty easily.
I support the PC design process & Mentor S/W tools for one of the HP
divisions here in Roseville. We have an in-house PC Design process. I agree
with AbdulraHman except for two things:
+ I've noticed that engineers often sit down with their PC
Designers for a few hours at the beginning of the design and
give them a general idea of what the board constraints are.
Given the Internet, it's possible for an engineer to do this
with an outside design house -- but it's a little tougher. A
good workaround would be for the PC Designer to copy their
"live" display to the engineer over the Internet & have a phone
conversation... We use HP's SharedX to do this inside HP, but
HP's firewall prevents us from sourcing X windows outside the
firewall.
+ Anther nice thing about an in-house PC Design group is that a
single copy of the design data can be shared between the engineer
& PC Designer. Changes happen quicker when you don't have to copy
data back & forth.
> I have known many engineers who did their own board design. I even know
> the president of one fairly large company who does it. My guess is that
> those are about the most expensive designs in their class in this country;
> this guy's time is probably worth ten times mine, or more.
> There really is very little special about P.C. board design such that a
> competent electronics engineer should not be able to do it, and do it well,
> *if* he or she wants to spend the time. The only little exception is that
> electronics engineers often don't know much about manufacturability, but,
> then again, judging from some of the layouts I have been given to clean up,
> neither do many designers, and, of course, an engineer can learn about this
> pretty quickly if it is desired.
I think AbdulraHman is being too kind to EE's. It's true that a good
EE can become a good PC Designer, but they have to take the time to learn
the design rules, how to use the tools, what tricks to use to make the board
space work for them -- but relatively few EE's have the patience to spend
their time learning & practicing these things. As an example, I spent an
hour laying out a simple PCB that had to be small (and I tried) -- one of our
PC Designers spent 10 minutes working on my layout & cut the board area in
half. Probably the best way to say it is that PC Designers spend years
practicing their ability to "see" the PCB & signal flows in their minds &
quickly solve the associated problems -- the engineers similarly spend years
understanding circuit topology, digital design, etc., but are usually not as
good at the PC Design problem as PC Designers. Of course, inexperienced PC
Designers may not be as good as experienced engineers... it works both ways.
Thanks,
Carl Wuebker * HP Roseville * c...@f.rose.hp.com * (916) 785-4296