Designer’s Guide Community

sos

Community Member

Offline

Posts: 32

MOS input bias current with esd diodes
Mar 18^th, 2010, 12:44pm

I'm a bipolar designer doing an MOS amplifier chip. The input pad connects to a fet gate plus the usual reverse biased esd diodes to each supply rail.

MOS amplifiers are expected to have picoAmp input bias currents. This seems like a lot to ask of the leakage in a relatively large esd diode. The foundry simulation models for a single diode show leakage of about 20pA with 5V across it, so they think it works, just a lot lower than I expected. I will appreciate any explanations to improve my intuition.

Steve

Back to top

IP Logged

wave Senior Member Offline Posts: 117 Silicon Valley	Re: MOS input bias current with esd diodes Reply #1 - Mar 18^th, 2010, 5:29pm Steve - Nothing wrong with your intuition. For generic Op-Amps CMOS, standard ESD is fine and most people can neglect the leakage. �Basic pad cells can have nA or uA at temperature. For a precision devices, you'll obviously need custom ESD that meets your leakage specs. Good luck, Wave �:D
Back to top	IP Logged

sos

Community Member

Offline

Posts: 32

Re: MOS input bias current with esd diodes
Reply #2 - Mar 19^th, 2010, 5:53am

wave wrote on Mar 18^th, 2010, 5:29pm:

Steve - Nothing wrong with your intuition.

For generic Op-Amps CMOS, standard ESD is fine and most people can neglect the leakage. �Basic pad cells can have nA or uA at temperature.

For a precision devices, you'll obviously need custom ESD that meets your leakage specs.

Good luck,
Wave �:D

Do you think my foundry's models are reasonable at room temperature? I agree that the leakage should go up a lot with temperature.

Thanks,
Steve

Back to top

IP Logged

love_analog Senior Member Offline Posts: 101	Re: MOS input bias current with esd diodes Reply #3 - Apr 22^nd, 2010, 4:37pm You probably got your answer by now. I have worked with some BCD foundries before and their pad leakage was way higher than 20pA - more like 1uA or so. Need a custom ESD cell with low junction area. Would be interested in what you find out.
Back to top	loveanalog.blogspot.com The Power of Analog IP Logged

sos

Community Member

Offline

Posts: 32

Re: MOS input bias current with esd diodes
Reply #4 - May 3^rd, 2010, 12:00pm

love_analog wrote on Apr 22^nd, 2010, 4:37pm:

You probably got your answer by now. I have worked with some BCD foundries before and their pad leakage was way higher than 20pA - more like 1uA or so.
Need a custom ESD cell with low junction area. Would be interested in what you find out.

Thanks for the reply, sorry for my lack of attention. So far we've decided that the leakage is something less than 1nA, and we have board leakage issues. We're waiting for a board turn (putting ground rings around the inputs).

I'm on xfab xb06. My esd diodes are about 40% larger than a pad.

Steve

Back to top

IP Logged

love_analog Senior Member Offline Posts: 101	Re: MOS input bias current with esd diodes Reply #5 - May 5^th, 2010, 12:41pm Fabs tend to give you larger ESDs than required. Helps them to sell more wafers
Back to top	loveanalog.blogspot.com The Power of Analog IP Logged