The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Doubling lenght in pseudo cascode
https://designers-guide.org/forum/YaBB.pl?num=1207433628

Message started by thechopper on Apr 5^th, 2008, 3:13pm

Title: Doubling lenght in pseudo cascode
Post by thechopper on Apr 5^th, 2008, 3:13pm

Hi All,

I have a simple question that usually makes me doubt and for which I got different answers:

In current steering DACs or any other structure comprising MOS current mirrors if doubling the length of the unit cell is required a pseudo cascode (the pseudo cascode involves two MOS in series with their gates tied together to the same potential) is usually the option. The question is: where should the bulk of the cascode device be connected to in order to get an effective length of 2L (L is the length of the both devices) , to its own source or to the source of the main current mirror?

I'm interested in your comments.

Best Regards
Tosei

Title: Re: Doubling lenght in pseudo cascode
Post by vivkr on Apr 7^th, 2008, 11:11pm

Hi Tosei,

I would imagine that the bulk connection ought to remain the same. Then, you can remove the name "pseudo-cascode". The
combination looks like a single transistor.

Perhaps you can give the arguments for the other option since you mention having received different answers.

Regards
Vivek

Title: Re: Doubling lenght in pseudo cascode
Post by thechopper on Apr 15^th, 2008, 6:53pm

Hi Vivek,

I would tend to think the bulk of the cascoded transistor should go connected to its own source since, in that way and assuming same W/L that the other one, woould have the same Vth. Having the same threshold voltage guarantees the cascoded device will work in the saturation region will the other one will be in triode. If you force the condition of same drain current on both the triode and saturated devices having same W/L you'll get I=Iref/2, thus looking like a 2*L device.

Do you agree with this?
Thanks and regards

Tosei

Title: Re: Doubling lenght in pseudo cascode
Post by loose-electron on Apr 17^th, 2008, 3:38pm

What is the motivation for doing this? You wont get better matching, since you are still using the shorter channel devices. If you just want to get a 2*L effect, then (for NMOS, with the bottom transistor grounded) then the two devices should share a common substrate node and that gets tied to ground.

But with the above, why not just make the transistor have a L = 2X what it was to do the same thing?

If you are trying to do something where the top transistor (NMOS example still) has less body effect, then tie the body to it's source. However, if it is P-substrate, N-well CMOS, you dont have that freedom. They are all on the same substrate.

If the example twists to a PMOS device where you DO have control of the substrate, then you can tie the well to the source. BUT when you start doing layout on this, the design rules associated with wells wil make the layiut get big very quickly, and you will probably want to put them in a common well to save space.

Lots of possibles here.

Title: Re: Doubling lenght in pseudo cascode
Post by thechopper on Apr 20^th, 2008, 8:27pm

Hi loose-electron,

The main idea behind using equally sized devices for getting 2L effect is not to break the symmetry of - for example - an N bit current steering DAC. If differentail and integral non linearity specs are stringent, then you will care about matching even for the LSB. Assuming the LSB is the 2L device I'd rather use two in series with L to get such effect than using one of 2L...layout wise is much more symmetrical and allows you for performing a better scrambling (for averaging errors) among current mirrors.

Again, tying both bulks to same common substrate is not exactly equal to @L (body effect). On the other hand I agree with you that tying the body to source will end up generating more space between wells, therefore affecting matching somehow.

So, what option should be chosen for optimizing matching?

Tosei