The ramp function from my ramp generator is an input to a comparator(vin-) and another input is vref = 0.45V(vin+), the resulting wave is a square wave with 50% duty cycle.
Now when I made a schematic of bandgap voltage reference(attached image) and simulate it I have difficulty having an output of Vref = 0.45V, I can only get 0.8V-1V vref no matter how much I change the resistors.
My source voltage is 1.2Vdd.
I try using the formula(attached image), but I don't really know what is Vt and n. I assume it is my source voltage and bjt ratio of areas. Please enlighten me.
Also if you can recommend some better bandgap voltage reference I would really appreciate it.
Thank you very much for taking time reading this:thumbsup:
Bandgaps have a fix 1.2V output voltage, so your circuit won't operate with 1.2V Vdd. Simply not possible.
Bandgap requires OPAmp not a comparator. Comparator is instable in a feedback loop.
Vt is the thermal voltage (k*T/q), n is the multiplier of the BJT device, which is actually yes, the ratio of the areas.
Thanks for the info. Can anyone recommend a voltage reference circuit that can output 0.45V at 1.2vdd? I need a temperature independent voltage reference. What are my options other than BGP?
Can I use a simple voltage Divider(resistors and supply voltage)?
Dear ZoOneR,
It is really interesting, I have never heard about similar. My only concern about it is how does it work when the resistors or BJTs have small mismatch?
I am not sure but the shifted characteristics of the BJT devices can introduce maybe higher curvature differencies over temperature which come from asymmetrics.
Maybe you have got more experiences about this, I am realy interested, I couldn't figure it out from this article.
If you are not really pressed for area and current a simple
resistor divider off the Vref point can give you any lesser
output you want (provided that the pass transistor MP17
is up to the task, in its min/max gate drive and sizing).
The source voltage is the problem.
It seems to me (though I have never personally put it
to the test) that the bandgap could be "folded" so the
TCR resistors are not stacked on top of the PTAT pair
but PTAT match-point current could be put onto (say)
a divided-down Vf and (say) a half-value TCR resistor
to get a half-bandgap voltage. The Vf/2 would need an
amplifier or a properly tuned resistance (wonder if the
divider resistor, and the TCR resistor, could have a
"magic point" that allows them to be one and the same?).
You might chase through "sub-bandgap reference" art
and see what's been going on; it has been somewhat
popular ever since Vcore dropped below 1.8V.
Hi frankrose,
Sorry for the very late reply. I haven't used this circuit yet, since a simple bandgap satisfied my needs until now. If you try it share with us the results :thumbsup: