Magic Voltage - Supply Independent Voltage Reference

What? What will 430 mvolts do? Also, don't really need the second dc power supply that is one volt.

The circuit requires 1v minimum to operate, in this case a 1v power supply has been utilized so that when the potentiometer is at its minimum setting, there is still at least a 1v supply. The same effect could have been achieved by placing a 36ohm resistor in the 1v power supply position. While this circuit design is a good idea, is there away to obtain a higher value of reference voltage from the circuit.

Guys i think @asdasda like to demonstrate a bandgap voltage reference. Far from perfect but somehow it does work http://everycircuit.com/circuit/5657748327104512

@justinmh Assume that you have a rechargable battery, 3.3v, and you want to produce 1.8v supply for your circuit. You will use buck converter or something. However, Rechargable battery voltage always changes according to charge level. In that point, you should have at least 1 supply independent voltage reference or current reference to be sure about that your circuit has exactly 1.8v supply. This circuit is taken from 'Analog CMOS Integrated Circuit', Behzad Razavi.

I want to restrict supply range between 1v and 30v because this circuit does not work at low voltages. Another problem about this circuit is 'startup'. However, detailed circuit analysis is needed to observe startup problem

Yes, except for usually to charge a battery you have a higher voltage potential, not lower like you're insisting. You don't use 1.8 v to charge 3.3, do an analysis of this and you will see the battery will over come your charger. A good example of this is your phone battery that is more than likely a 3.7v lithium ion or lithium poly battery, the charger outputs a nice 5 volts dc.

Also, if you were able to boost the voltage through a converter to get it to a nice voltage high enough to charge, remember ohms law takes effect here. Same thing with a transformer with ac. If you have a one volt ac power supply that is capable of 10 watts, that means it can produce 10 amps at 1 volt, if you step that up to 10 volts then your current is only 1 amp while the voltage is 10 volts because p=I've, and energy is conserved. That why we usually step down voltages because power stays the same, and voltage drops, that means current has to increase in order to satisfy the equation.

So you would the voltage, just not enough current, it would be a trickle charge at best.

This is not for battery charging. For example If you use 180nm cmos process your supply voltage is probably 1.8V. If it is portable like phones there should be a battery and you should design down converter to supply your analog circuit with 3.3v or 5v battery. Voltage and current reference is needed to stabilize output voltage of your converter. Wish to be more clear now.

@justinmh, please lookup what a reference voltage is. Anyway, @asdasda this circuit wont work in reality, cuz pA are much to low. Keep it up.

@hurz check 'Analog Integrated Circuit Design, Carusone-Johns-Martin' 2012 pub. chapter 7.2.1. Show me the correct one and I will delete this circuit.

This circuit works. Only startup problem occurs after production. It is solved by adding another transistor related with design. But in computer simulation, current is higher, you are right. I thing it is related with everycircuit. It is not so sensitive in analog design.

And what issue you see for this circuit with EC?

EC has problems. Mosfets can produce their gate voltages by current in EC. It is false. Check that circuit. Gate voltage occurs although it is floating. http://everycircuit.com/circuit/5621014813999104

Put a giga ohm and it will droop to zero. This is cuz there are capacitances inside mosfet model. While sim startup the DC operating point is calculated. I wont say its an EC problem. The DC OP has to be calculated to something clever. If you have a better way to calc DC OP values we might start a new topic for this.

I wont expect any different voltage here than supply / 2 http://everycircuit.com/circuit/6230931065274368

It is totally false approach. First of all, Cgs and Cgd capacitances are directly related with operation region of Mosfet. They are not always equal. Secondly, I have simulated the same circuit in 3 different programs LTspice, Pspice and Proteus. Ltspice and Proteus shows that Gate node has about 1 pV voltage. Pspice rejected to make simulation because of floating node. After that I have tried to make biasing with capacitances. The results were same.

Check that circuit. EC contradicts with itself. Simulate circuit when the switch is on. result does not change with the switching. After that, Simulate circuit when the switch is off. When you change the switch, result is totally different. http://everycircuit.com/circuit/6143434931044352

You should reject circuits with floating nodes by yourself, dont you? And about Ciss, its the model what EC uses and its not to bad. If you want to change the model you have to use spice on a pc. Again what is the voltage of a floating node YOU expect? I wont expect a proper value.

This new example is more interesting. As far as I understand one can call it "bootstrap" for the open start and then close case. The closed start is boring and for sure different cuz its done by DC OP calculation. The circuit itself doesn't make sense, connecting a mos source to a mos gate and nothing else, I hope you agree. But its definitive not contraticting cuz both are different cases!! Start-Open-and-close and closed-start. Simulators are quite complex

Floating node cannot be calculated because it is out of circuit rules. Last circuit is not bootstrap. Both cases are false. Maximum gate voltage of last mosfet is 5v. And we know that Vgs-Vt<0 for cut off. So maximum source voltage is Vdd - Vt. It is 4.57 for this circuit. In circuit, it is 3.71v for first case. It thinks that output is Vdd - 3*Vt. In second case, it says output is Vdd, 5v, and gate voltage is higher than Vdd! Both cases are false.

Its a little hard to have a discussion with you, cuz you coming again and again with new cases. Maybe you can comeback to the original circuit and simplifiy it at that point were you see problems. Open new fields doesn't help us a lot.

Call a spade a spade. EC has errors. I dont know who you are. But I know that If I were a designer of this program, I would be grateful for users who find problems about EC. We paid for this program and it has errors! My profession is analog VLSI circuits and EC has problems even if you do not accept. I am not trying to help. I just want to use a reliable android circuit program. Errors should be checked.

EC has errors, but more importante EC does have limitations. As any simulation environment have! Cadence Pspice does have also limitations to reality. But EC is probably the most simple and useful compromise in between. I dont expect Pspice results but I can see EC is much ahead of all other sim applications on android for people who want to start and continue learning electronic. BTW, Cgs in one of your examples does cause a bootstrap behaviour, but you dont agree to this which makes it a little difficult in this discuss. Again, lets break down your original circuit and see why its not proper simulated.