The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Precision Poly Resistor - # of contacts
https://designers-guide.org/forum/YaBB.pl?num=1273512631

Message started by jiesteve on May 10^th, 2010, 10:30am

Title: Precision Poly Resistor - # of contacts
Post by jiesteve on May 10^th, 2010, 10:30am

I am laying out a precision poly resistor... I would like to have the most flexibility in selecting tap points so I have filled the entire length with contacts (each contact is 0.24um apart) with overall length 300um so more than 1000 taps. Will having so many taps alter the R/sq of the poly ?

Thanks

Title: Re: Precision Poly Resistor - # of contacts
Post by wave on May 10^th, 2010, 7:13pm

Are you planning on spinning metal masks until you get lucky?
And hoping the process never shifts?
I would think analog trim links would be easier to get in somehow.

I'm hopeful there is a good story here.

8-)

Title: Re: Precision Poly Resistor - # of contacts
Post by loose-electron on May 10^th, 2010, 8:48pm

"precision poly resistor"

No such thing.

+/- 20% variance on process needs to be accounted for.

even if you are trying to get matched resistors you will have matching errors and contact resistance vaires widely.

What is the problem you are trying to solve with this magic resistor? There are probably better ways to solve the problem.

Title: Re: Precision Poly Resistor - # of contacts
Post by jiesteve on May 11^th, 2010, 10:42am

This is my nifty VDD/2 rail splitter reference/resistor divider. It's a poly resistor between VDD and GND with a large NMOS switch at the bottom.

I want to have the flexibility to get right at the 1/2-way point in terms of voltage and cancellating out any systematic offsets (say from package routing, for example, or anything else). I really only need enough tap points around the middle, but since this is a matched resistor I need to have the contacts over the entire length. I'm worried about messing up my R/sq or anything else by having so many contacts over it.

Title: Re: Precision Poly Resistor - # of contacts
Post by jiesteve on May 11^th, 2010, 10:44am

Also, I'm not changing the length, just the tap point on the divider. Contact resistance doesn't matter because there is no DC current flowing through the contacts/tap point.

Title: Re: Precision Poly Resistor - # of contacts
Post by love_analog on May 14^th, 2010, 4:16pm

why don't you just put 2 equal resistors between Vdd and ground and take the mid-point as Vdd/2.

Title: Re: Precision Poly Resistor - # of contacts
Post by ywguo on May 19^th, 2010, 4:39am

Hi jiesteve,

What are you working on? A resistor ladder or just a divide-by-2?

What type of resistor do you use? Poly with silicide or without silicide?

What is the resistance of the all resistors?

What's the accuracy/error you are seeking? If you need a 10+ bit DAC, there is an good example in VERMA and RAZAVI's paper. JSSC Nov. 2009.

Best Regards,
Yawei

Title: Re: Precision Poly Resistor - # of contacts
Post by jiesteve on May 19^th, 2010, 9:19am

It's just a divide-by-2 with metal options to adjust the tap point.

We're using poly with silicide (~16ohms/sq typical)

Total resistance is 1.6kohm

I need something like 10-11bit accuracy

Title: Re: Precision Poly Resistor - # of contacts
Post by rf-design on May 19^th, 2010, 6:43pm

Taking your numbers

16 Ohms/sq & 1.6KOhm and 11 Bit accuracy of the 0.5 divider ratio makes 50sq+50sq but the tap should be stepped at 50/2048 sq =1/40 sq

You should have a very wide poly to make tap adjustments by 1/40sq with single contacts. I would also layout two interwinded snakes of poly because otherwise the poly and silicide thickness gradient are violating 10-11bit. If you want to have a selectable tap by less than a contact size I propose to use a second resistor connected at 0.525 and 0.475 positions and having a higher resistor and being longer to place the taps.

Title: Re: Precision Poly Resistor - # of contacts
Post by ywguo on May 20^th, 2010, 8:48am

Hi Jiesteve,

What's Reiner is correct. It seems very difficult to fullfil your job using the low sheet resistance poly resistor only.

Yawei