The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> What happens if L of all MOSFETs is increased keeping the bias same?
https://designers-guide.org/forum/YaBB.pl?num=1283203449

Message started by kanu on Aug 30^th, 2010, 2:24pm

Title: What happens if L of all MOSFETs is increased keeping the bias same?
Post by kanu on Aug 30^th, 2010, 2:24pm

I have a basic doubt here. For any circuit, what happens if we increase the Ls of all transistors, keeping the bias circuit same. Will it increase the Vgs of the transistors to keep the same previous current in the circuit as per the square law equation ?

Thanks,
Kanu

Title: Re: What happens if L of all MOSFETs is increased keeping the bias same?
Post by AnalogDE on Aug 30^th, 2010, 3:12pm

With current fixed, increasing L while keeping W the same will increase Vgs. In practice this is limited by the headroom on the current source -- at some point you'll choke off the current source then the assumption that I is fixed no longer holds.

Title: Re: What happens if L of all MOSFETs is increased keeping the bias same?
Post by kanu on Aug 30^th, 2010, 9:24pm

Hello AnalogDE,

Thanks for your help. I have further query. We have a level convertor circuit in our project, which is not working across lower temperatures in the range of -25 to 125 C. I understand that threshold voltage, Vth of transistor increases with decrease in temperature. So I tried increasing L and found that the circuit worked fine at lower temperatures as well.

So I wanted to confirm if this behavior is expected as per the explanation of increasing L increases Vgs for same current and drives more transistors in saturation ? If so, what constant biasing should I be referring for my explanation .... is it the input of level convertor which has not changed ?

Thanks,
Kanu

Title: Re: What happens if L of all MOSFETs is increased keeping the bias same?
Post by kanu on Aug 30^th, 2010, 10:27pm

One correction in my previous post.... I meant transistors are going from cutoff to triode as Vgs is increased. We are basically trying to convert subthreshold voltage to superthreshold by the help of level converter.

Title: Re: What happens if L of all MOSFETs is increased keeping the bias same?
Post by AnalogDE on Aug 31^st, 2010, 8:56am

Can you please post a schematic of your circuit? I'm not exactly sure what's going on. Also, what do you mean by "working"

Title: Re: What happens if L of all MOSFETs is increased keeping the bias same?
Post by kanu on Aug 31^st, 2010, 10:57am

Hi AnalogDE,

I have attached a screenshot of the schematic of level convertor.

Further, what I meant by "working" is that if we give pulse input to the level convertor circuit at low temperatures (say -25 C for example), it gives the output as straight line instead of a pulse.

But I found that if I increase L of the transistors, I get a pulse for lower temperatures as well. So I was looking for a possible explanation for this behaviour.

Thanks,
Kanu

Title: Re: What happens if L of all MOSFETs is increased keeping the bias same?
Post by AnalogDE on Aug 31^st, 2010, 11:21am

These circuits are kind of tricky because they rely on contention. Sweeping L is doing something to your P-N ratio strength -- I suggest you narrow down exactly which one of the four level shifters is failing, then you can analyze the fail point deeper. You can see how much drive you get at the gate of the switching PMOS (drain of NMOS pulling down) when switching and compare failing vs passing cases.

Title: Re: What happens if L of all MOSFETs is increased keeping the bias same?
Post by kanu on Aug 31^st, 2010, 11:55am

I did a .op analysis for the circuit and found that the first stage (which is nearest to the subthreshold voltage) has many of its transistors in cutoff which is why the circuit was failing.

So I increased the L by 2 times for all the transistors of the Level convertor to test my theory. Although the circuit strated working fine, I wasn't convinced with Vgs increasing with increase in L for constant current. This is why I just wanted to confirm if this explains the behavior.