Design Con 2015
Home    Bloggers    Blogs    Article Archives    Messages    About Us   
Tw  |  Fb  |  In  |  Rss
Scott Elder

Analog Design on 2 Different Planets

Scott Elder
SunitaT0
SunitaT0
10/31/2013 2:55:38 PM
User Rank
Master
Re: IC design rules and tips
 The other thing I see that is easily done is resistor ratios and capacitor ratios.

@RedDerek, one more important issue is  Matching. Matching is very much needed if we want build accurate systems hence matching of devices is absolutely necessary if we want to build robust systems.

50%
50%
SunitaT0
SunitaT0
10/31/2013 2:51:57 PM
User Rank
Master
Re: New process for IC designer
 Stick with guaranteed, universally accepted methods and one will avoid alot of headaches.

@Scott, completely agree with your opinion. I think iys very important to remember that having shortcut like trick might cause some issues in future hence its always better to stick to universally accepted methods.

 

50%
50%
SunitaT0
SunitaT0
10/31/2013 2:47:39 PM
User Rank
Master
Re: IC design rules and tips
You design on ratios and not on accurate numbers.

@goafrit2, thats the beauty of analog design, always work on ratios and hence makes the implentation easy.

50%
50%
goafrit2
goafrit2
7/4/2013 6:44:00 PM
User Rank
Master
Re: IC design rules and tips
>>  Though the value may change slightly due to process and wafer locality, the ratio, unless damaged cells, is very accurate to generate.

That is the holy grail of matching especially in unstable process. You design on ratios and not on accurate numbers.

50%
50%
goafrit2
goafrit2
7/4/2013 6:40:03 PM
User Rank
Master
Re: New process for IC designer
>> But in the future, it might impact IC designer. Manger and VP could be educated by a new information.

@Deaj, I do hope it is not the VP that drives your design strategy. They can setup the big picture but at circuit level, they must stay away. My VP thinks that all designs are the same - all you need is modify old circuits. Yet, every 2 years we have a new process and things just get harder.

50%
50%
Scott Elder
Scott Elder
7/3/2013 8:06:11 PM
User Rank
Blogger
Re: New process for IC designer
@DaeJ -- From time to time one sees proposals for clever tricks.  My concern is always that every one in the process of making the IC needs to agree with the tricks.  Otherwise, one day someone in the fab will change something to "make it better" and will completely KILL your trick.  Stick with guaranteed, universally accepted methods and one will avoid alot of headaches.

50%
50%
Scott Elder
Scott Elder
7/3/2013 7:49:45 PM
User Rank
Blogger
Re: IC design rules and tips
@RedDerek - Thanks for commenting!  Let me think about a good blog angle on design rules/tips.  It is always the unwritten things that get designers into trouble.  Just like the unwritten contract between designers and fabricators to use guaranteed wafer results rather than recognize some trick that could be played.  Tricks are great as long as they don't come at the expense of another party (i.e. the fab guys didn't know you depended upon TC tracking!)

50%
50%
RedDerek
RedDerek
7/3/2013 3:57:01 PM
User Rank
Master
IC design rules and tips
I think a blog on IC Design Rules and Tips would be helpful. I am aware of the bandgap understanding, but never really implemented into silicon myself. The other thing I see that is easily done is resistor ratios and capacitor ratios. Though the value may change slightly due to process and wafer locality, the ratio, unless damaged cells, is very accurate to generate. Thus if one knows what 2.5000 volts is based on the bandgap, a 1.2500 reference can be made by the resistor divider.

50%
50%
eafpres1
eafpres1
7/3/2013 2:16:59 PM
User Rank
Blogger
A difference is a difference?
@Scott--"Another important aspect of ΔVbe is that it is proportional to absolute temperature (PTAT) with a temperature coefficient of about +3300ppm. As it turns out, the aluminum metal interconnect used on an IC also universally has a TC of about +3200 ppm. So now one can build a current sense resistor out of aluminum interconnect and compare the current sensed voltage against the ΔVbe voltage with a near zero TC drift of an absolute over current detection threshold."

I like your article becuase, although I'm not an IC designer (even remotely) the idea of working with and taking advantage of everything a process gives you seems inherently right to me.

However, in your example, you use one resistor plus the other voltage sense to correct for temperature.  The author you cited was taking advantage of two resistors in the process space that would also result in low TC drift.  Is your main case here that his method requires adding an extra resistor?

One concern is how stable are the two coefficients you cite across process runs?  There are lots of "off data sheet" applications at many levels in electronics, but they can come back to bite you if something changes the thing you are taking advantage of.  Again, I'm essentially a caveman vs. you on IC design, so this may be nonsense.  But is there a risk that process changes could fiddle with your version of temperature compensation?

50%
50%
DaeJ
DaeJ
7/3/2013 10:49:42 AM
User Rank
Master
New process for IC designer
I am wondering if there is more article or journal paper by Jean-Francols Debroux.  I am not sure how it is practically, unless understanding all process. But in the future, it might impact IC designer. Manger and VP could be educated by a new information.

50%
50%
More Blogs from Scott Elder
Even without the hand holding of expensive, sophisticated simulation software, you can design an analog IC. Let's proceed.
When designing an analog custom IC, be sure to look at the big picture, rather than simply transferring a discrete design into an ASIC.
Fewer companies fabricate their own ICs; more are going fabless. Companies now just need the software to design their ICs and that is a growing trend.
Precision analog components are expensive -- everyone knows that. They are expensive for good reasons -- or are they? Could they actually be significantly cheaper?
We would benefit from a hardware description language tailored specifically for analog design.
flash poll
educational resources
 
follow Planet Analog on Twitter
Planet Analog Twitter Feed
like us on facebook
our partners
Planet Analog
About Us     Contact Us     Help     Register     Twitter     Facebook     RSS