The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> How long does analog circuit design take?
https://designers-guide.org/forum/YaBB.pl?num=1328864489

Message started by philcorb on Feb 10^th, 2012, 1:01am

Title: How long does analog circuit design take?
Post by philcorb on Feb 10^th, 2012, 1:01am

This may be slightly broad, or off topic, but how can one estimate the time a circuit will take to design? I have heard various estimation methods, such as 1 week per schematic. What do you do to estimate design time? Is this something that can be learnt from experience? I'm really looking at the most abstrate ways to estimate this.

kind regards,

Phil

Title: Re: How long does analog circuit design take?
Post by Frank Wiedmann on Feb 10^th, 2012, 5:52am

You should base your estimate on the design time of earlier projects of similar size and difficulty. A commercial tool for doing this is described at http://www.numetrics.com/numetrics-icpa-analog-rf-capabilities/. As I have not used it, I don't know how well it works in practice.

Title: Re: How long does analog circuit design take?
Post by loose-electron on Feb 10^th, 2012, 11:39am

heavily dependent on the difficulty of the design, the experience
of the designer, whether it is a new thing, or a variation on an old design.

Very tough to generalize.

Title: Re: How long does analog circuit design take?
Post by philcorb on Feb 17^th, 2012, 1:09am

Thanks very much for the responses

Frank Wiedmann,

Thanks for mentioning the numetrics tool. This is a company I hadn't heard of before. It sounds interesting and will be certainly contacting them. Has anyone else used this tool?

loose-eletron,

I agree, it is very hard to estimate or to generalise. However, I think it is an important aspect of design. It is important to know how long projects are likely to take, and estimates can vary greatly, from close to completely wrong. What are the main reasons for circuit design not meeting schedules? Are their practical steps that can be done to improve time scale estimates?

Title: Re: How long does analog circuit design take?
Post by loose-electron on Feb 17^th, 2012, 1:25pm

philcorb wrote on Feb 17^th, 2012, 1:09am:

What are the main reasons for circuit design not meeting schedules? Are their practical steps that can be done to improve time scale estimates?

Big issues:

Poorly defined target-specifications for the design

Feature creep (adding new features after design starts)

Unrealizable performance criteria

Redefinition of design in the middle of the process.

Learning curve and experience issues.

Those are what come to mind.

Title: Re: How long does analog circuit design take?
Post by carlgrace on Feb 29^th, 2012, 4:23pm

loose-electron wrote on Feb 17^th, 2012, 1:25pm:

philcorb wrote on Feb 17^th, 2012, 1:09am:

What are the main reasons for circuit design not meeting schedules? Are their practical steps that can be done to improve time scale estimates?

I would add:

1. Inadequate design resources

2. Poorly set up tool flow (leads to a lot of time wasted getting simulations to run on brittle systems)

Title: Re: How long does analog circuit design take?
Post by Lex on Mar 1^st, 2012, 1:09am

philcorb wrote on Feb 17^th, 2012, 1:09am:

...
What are the main reasons for circuit design not meeting schedules? Are their practical steps that can be done to improve time scale estimates?

Taking aside the nonideal behavior of customers (everybody that received a sudden spec change, raise your hand ;) ).

The design process involves iterations. Up front you cannot judge how many iterations you need.

The number of iterations depends on
-talent of the designer
-knowledge of the designer
-experience of the designer
-organization of the designer (structured schematic/layout entry, database organization, documentation)
-verification skills of the designer
-tools available
-support software/tools to prevent losing time with a wrong setup
-support technical - discussion on technical matters with a third person
-...
-...

This means that the time estimate for every project will be different. In practice this results in managers pushing the designers maximally, accepting delays rather than not pushing and having them idle. Hence these optimistic time schedules are not met. This is my opinion.

Title: Re: How long does analog circuit design take?
Post by Vladislav D on Mar 1^st, 2012, 2:03am

Lex wrote on Mar 1^st, 2012, 1:09am:

philcorb wrote on Feb 17^th, 2012, 1:09am:

...
What are the main reasons for circuit design not meeting schedules? Are their practical steps that can be done to improve time scale estimates?

You are talking like a manager, Lex...
I agree with Jerry....

Title: Re: How long does analog circuit design take?
Post by RobG on Mar 1^st, 2012, 8:09am

Changes. Then again, if paying by the hour, change all you want ;D ;D

Title: Re: How long does analog circuit design take?
Post by Lex on Mar 2^nd, 2012, 5:03am

Vladislav D wrote on Mar 1^st, 2012, 2:03am:

...
You are talking like a manager, Lex...
I agree with Jerry....

Ouch that hurts.

Not even sure why you're saying that, since I just reflect how things are done (in several companies I've been, at least).

Maybe you have different experience(s)?

Title: Re: How long does analog circuit design take?
Post by loose-electron on Mar 2^nd, 2012, 1:12pm

I actually think we are all
saying sort of the same thing.

As for consulting gigs dragging out due to
customer changes?

Me, I hate that.

I just wrapped something up where
the customer was requested to provide
feedback 4 times along the way.

All 4 times, no feedback until we
were halfway into the next step.
Redirect and change, toss what you
have been doing for the last few weeks.

I hate that, even when
I am getting paid to do it.

Title: Re: How long does analog circuit design take?
Post by loose-electron on Mar 2^nd, 2012, 1:17pm

Lex wrote on Mar 1^st, 2012, 1:09am:

This means that the time estimate for every project will be different. In practice this results in managers pushing the designers maximally, accepting delays rather than not pushing and having them idle. Hence these optimistic time schedules are not met. This is my opinion.

Setting of schedules is an interesting thing.

Impossible to keep schedule (that slip) vs. realistic schedule with
"the schedule does not slip" mantra?

I've done both. I prefer the latter, and when I am doing PM stuff
that's how I fly it. Get better results, and often sooner that way.

Title: Re: How long does analog circuit design take?
Post by philcorb on Mar 9^th, 2012, 1:17am

Thanks everyone for your great contributions

Thanks loose-electron, carlgrace and Lex for coming up with some great reasons why schedules are not met. For many of the points I feel I have been there and seen that and it is nice to see them in comprehensive lists.

RobG, thanks for the poster image -- very nice :)

Frank Wiedmann, I've tried following up the numetrics tool that you mentioned. Not sure if the company is still in business. I've tried to contact them but haven't had a response. Anyone know if they still exist?

Loose-electron, you make a very good point about the 'impossible schedule' versus 'no slip schedule' approches. Is is ever possible to have a 'no slip' schedule? Does this approach work? What needs to be in place for it to work well? Is there benefit from the 'impossible schedule'? Can it ever be helpful, for example increasing speed/productivity?

Title: Re: How long does analog circuit design take?
Post by loose-electron on Mar 9^th, 2012, 8:01am

philcorb wrote on Mar 9^th, 2012, 1:17am:

Loose-electron, you make a very good point about the 'impossible schedule' versus 'no slip schedule' approches. Is is ever possible to have a 'no slip' schedule? Does this approach work? What needs to be in place for it to work well? Is there benefit from the 'impossible schedule'? Can it ever be helpful, for example increasing speed/productivity?

The no schedule slip methods require that you know what you are doing as a manager, assign realistic deadlines, well defined and frequent milestones, allow some revision/respin in the schedule and meet those milestones.

You meet all the milestones in a series of smaller goals.
No exceptions, no excuses, no "it can wait"
Since the goals are smaller, then its more realistic.

To do that requires somebody in the room that knows what they are doing (did I say that already? :) ) and is the rare exception and not the usual course of action.

Oh, and you got a milestone due on Monday, and you finished and everyone is happy with it on Thursday? Enjoy your three day weekend.

Oh, and you got a milestone due on Monday and its not looking ready on Friday afternoon? Looks like you are working Saturday (and Sunday) if needed.

It works, but its not how most of the world does things.

Title: Re: How long does analog circuit design take?
Post by boe on Mar 9^th, 2012, 12:10pm

philcorb wrote on Mar 9^th, 2012, 1:17am:

Is there benefit from the 'impossible schedule'? Can it ever be helpful, for example increasing speed/productivity?

If you find one, please let me know... YMMV but it tends to decrease productivity: You are behind schedule, after all, so some manager might try to give you more people to get you back on (the impossible) track. Unfortunately, adding more people to a late project makes it even later (and the team less efficient).
- B O E

Title: Re: How long does analog circuit design take?
Post by loose-electron on Mar 10^th, 2012, 12:50pm

boe wrote on Mar 9^th, 2012, 12:10pm:

philcorb wrote on Mar 9^th, 2012, 1:17am:

Is there benefit from the 'impossible schedule'? Can it ever be helpful, for example increasing speed/productivity?

Well Said!

:)

Title: Re: How long does analog circuit design take?
Post by philcorb on Mar 13^th, 2012, 1:07am

philcorb wrote on Mar 9^th, 2012, 1:17am:

I've tried following up the numetrics tool that you mentioned. Not sure if the company is still in business. I've tried to contact them but haven't had a response. Anyone know if they still exist?

Just as follow up, I'm made contact with Numetrics and they are still in business. Looks like I was having a bit of email trouble. I'll be interested in what they can provide.