voltage reference problem

[liletian]

Hi Guys
I found a paper as I attached, it seems to be a good voltage reference in the beginning, but when I used a short time step (1e-12 s) for a longer time, 100ns, then the volltage reference begins to oscillator, do anyone know what's the problem and how to fix it?
Thanks a lot

 

[ytliang]

not enough phase margin

 

[dumbfrog]

interesting paper....
(if I replaced those two mosfets with BJT....looks like the latchup circuit)

sounds like unstable?!

 

[quanble]

Why do you use so short time step ? Try use longer time pwl singal ( uS) .

Hi Guys
I found a paper as I attached, it seems to be a good voltage reference in the beginning, but when I used a short time step (1e-12 s) for a longer time, 100ns, then the volltage reference begins to oscillator, do anyone know what's the problem and how to fix it?
Thanks a lot

 

[liletian]

because if you use a longer phase margin, it can not caught high frequency characteristic. Anyway, it should converge at any time step, right? I suspect this voltage reference will be an oscillator instead of voltage refernece.
Any suggestions?
Thanks

Why do you use so short time step ? Try use longer time pwl singal ( uS) .

Hi Guys
I found a paper as I attached, it seems to be a good voltage reference in the beginning, but when I used a short time step (1e-12 s) for a longer time, 100ns, then the volltage reference begins to oscillator, do anyone know what's the problem and how to fix it?
Thanks a lot

Added after 10 minutes:

not enough phase margin

I thought it is unstable too, is there a way to fix it? Any suggestions?

Added after 36 seconds:

interesting paper....
(if I replaced those two mosfets with BJT....looks like the latchup circuit)

sounds like unstable?!

I would think it is unstable too, is there a way to fix it?
Thanks

 

[eegchen]

it likes a latchup circuit. Why do you use this one? any advantage ?

 

[liletian]

it likes a latchup circuit. Why do you use this one? any advantage ?

no advantages, maybe only it can give a very stable voltage for a big range input voltage. I am just curious if there is a way to fix it.

 

[saro_k_82]

This is a nice circuit and there is no need to fear linking it to SCR kind of latchup. The one in Fig 1 is same as a wilson mirror with the diode connected device replaced by a resistor. As we know wilson mirror can be unstable because of two poles., but just a bit of work on stability can give you a nice reference.

 

[niezimei]

there are two feedback loop:a negative(N2,P3,P1,R,N1) and a positive (N2,P3,P2).The loop can work only when negative feedback is dominant (R*gmn1>1).Add a cap to split the two poles for stability.