Selecting a proper pass element of LDO

[quaternion]

I am designing a LDO using umc 0.13.

I have used HGLV transistor as a pass element but the final size exceeds the expected.

So , Which is better to use as pass element is it Low Vth , or native Vth , or regular Vth ?

When I made an error amplifier for the Low Vth pass element ,the output common mode level was high {for proper operation at drop out region} which made restrictions in the design and affected the performance ,so I concluded that a higher Vth transistor is more appropriate and it may have smaller size ??

I want to know your opinion in that issue.

Also as I am not experienced enough in analog design ,I don't know the differences between such devices (HGLV , HGNVT , HG ), Can anyone declare to me these ; such as what is the difference between regular and native Vth , and why they are called High gain rather than high voltage,..,etc.

Thanks in advance.

 

[quaternion]

Re: Pass element of LDO

I have used HG-regular-VT , but the size doesn't decreases a lot [although the design has been simplified ]
Any help please .

 

[keoiuser]

Re: Pass element of LDO

hi
i think that the choise of transistor is also regarding the power supply, what is the value of your VDD?

 

[quaternion]

Re: Pass element of LDO

The drop out is 0.1V; Output of the LDO is 1.2V and input is ≥ 1.3V.

 

[ashish_chauhan]

Re: Pass element of LDO

Your choice of pass device (high vt ,low vt, regular) will also depend on the total leakage current spec you have. Otherwise there wont be much difference in the sizes of pass device supporting same amount of current.

If you need to have lower leakage go for high vt device.

 

[quaternion]

Re: Pass element of LDO

Unfortunately ,I don't have high Vth in my Kit I have (low VT , Native VT , regular VT)

I have asked my friend about Native VT ,and he told me that it has zero VT ??

All of the above types are high gain(high voltage) transistors I think they have low leakage,aren't they ?

 

[ashish_chauhan]

Re: Pass element of LDO

ur friend is right native device rather has negative vt. But I dont think high voltage device means a low leakage device. High voltage to me means capacity to bear higher voltages at gate and drain(vgs & vds).

 

[quaternion]

Re: Pass element of LDO

So , the regular Vth seems to be my best choice as it leads to a simple design and lower leakage and at a little smaller size than low Vth one.

I think in normal LDO design the quiescent current will be larger than leakage ,isn't this true?

Also, I have a little question; why high voltage devices are named high gain ?(is this related to the process?)

 

[ashish_chauhan]

Re: Pass element of LDO

Yes in normal designs the quescient current is greater than the leakage. And this rather is done intentionally ... to make the ldo stable at no load conditions.

As for your second question I am not very sure but maybe its so becoz they support higher current (hence greater gm) for same sizes. correct me if I am wrong.

 

[winsonpku]

Re: Pass element of LDO

I think it has much deffrence between high vt and low vt device.

Your choice of pass device (high vt ,low vt, regular) will also depend on the total leakage current spec you have. Otherwise there wont be much difference in the sizes of pass device supporting same amount of current.

If you need to have lower leakage go for high vt device.

Added after 5 minutes:

I think for the pass device, the low vth maybe a prefered choice
but this will add an extra mask and this will add the cost.

high voltage devices may have large output resistace,
so they have high gain under the normal condition.

BTW: Are you a student or engineer and where are you?

So , the regular Vth seems to be my best choice as it leads to a simple design and lower leakage and at a little smaller size than low Vth one.

I think in normal LDO design the quiescent current will be larger than leakage ,isn't this true?

Also, I have a little question; why high voltage devices are named high gain ?(is this related to the process?)

 

[quaternion]

Re: Pass element of LDO

I am an engineering student.

I am doing LDO project.

Added after 5 minutes:

I think for the pass device, the low vth maybe a prefered choice
but this will add an extra mask and this will add the cost.

How ?

In my design low vth transistor needs a high output DC level from the error amplifier ,that is why I preferred the regular Vth !

 

[ashish_chauhan]

Re: Pass element of LDO

I think it has much deffrence between high vt and low vt device.

I think for the pass device, the low vth maybe a prefered choice
but this will add an extra mask and this will add the cost.

high voltage devices may have large output resistace,
so they have high gain under the normal condition.

I won,t disagree to you but as I said in my previous post... choice depends on ur specs. If you are targetting a lower standby current... high vt shall be the preferred choice... keeping the stability at "no load " in mind.

by the way , who are you Engineer or student?

 

[winsonpku]

Re: Pass element of LDO

If you do the transient simulation by giving a step load current, you will find the Pass transistor work in the triode region and the low vth transistor will give larger current than the high vt counterpart with the same area.

Leakage current maybe not a problem compared with the quiescent current consumed by other part in the LDO.

If you do the analysis,you will find the stability is more serious in the large current load condition.
.

I think it has much deffrence between high vt and low vt device.

I think for the pass device, the low vth maybe a prefered choice
but this will add an extra mask and this will add the cost.

high voltage devices may have large output resistace,
so they have high gain under the normal condition.

I won,t disagree to you but as I said in my previous post... choice depends on ur specs. If you are targetting a lower standby current... high vt shall be the preferred choice... keeping the stability at "no load " in mind.

by the way , who are you Engineer or student?

 

[ashish_chauhan]

Re: Pass element of LDO

I think you are not getting my point...
I agree that in case of large current load transients a low vt device will help in quick recovery...

But, say when I need the reg to be functional and regulating for loads which can be below 100nA... i think you low vt device will make the error amp saturate and thus killing the loopgain and hence the complete regulating function.

In this case a high vt will be better as it will get shut before ur amp saturates.

I ur case , if you check stability at no load you will not be able stablize the regulator without bleeding some extra current through ur pass device , which can be more than 1uA even. And when you target a spec of 2uA TOTAL current consumption u cannot afford 1uA thru pass device alone!!!

hope I made my point this time.

 

[quaternion]

Re: Pass element of LDO

But, say when I need the reg to be functional and regulating for loads which can be below 100nA... i think you low vt device will make the error amp saturate and thus killing the loopgain and hence the complete regulating function.

In this case a high vt will be better as it will get shut before ur amp saturates.

That is what I meant .

 

[winsonpku]

Re: Pass element of LDO

Under your 'specified' condition, you are right!
I did not find the application in my life and it's very difficut to get your opinion in your replies!
LDO has a feed back loop in the normal application, and the feedback loop consumes a few uA under most condition.


BTW: Would you explain how the error amp is saturated with the low vt device in details? Thanks!

I think you are not getting my point...
I agree that in case of large current load transients a low vt device will help in quick recovery...

But, say when I need the reg to be functional and regulating for loads which can be below 100nA... i think you low vt device will make the error amp saturate and thus killing the loopgain and hence the complete regulating function.

In this case a high vt will be better as it will get shut before ur amp saturates.

I ur case , if you check stability at no load you will not be able stablize the regulator without bleeding some extra current through ur pass device , which can be more than 1uA even. And when you target a spec of 2uA TOTAL current consumption u cannot afford 1uA thru pass device alone!!!

hope I made my point this time.

 

[ashish_chauhan]

Re: Pass element of LDO

There are uP products in market which need such specs of ldo in their standby mode.

for a low vt device , with sizes being larger... when the amp tries to cut off the device it effectivey moves up the voltage at gate of pass device(pmos say) and in this case one of the transistors of output stage of amp may go out of saturation get vds as low as 25mv ... Now to fix this either you try to fix the amp with more current or the other way.

 

[quaternion]

Re: Pass element of LDO

I have used low Vth but its function didn't accomplished except when I have added an up level shifter(diode connected loaded nmos common source). And this is the design complexity I have meant in the past posts.

Now , I have another issue that is related to the same topic (so I will ask here):

How is the size of the pass element related to the gain of the error amplifier, I have modeled the error amplifier and found that increasing the gain decreases the required size - some how -, Can anyone explain please.

 

[ashish_chauhan]

Re: Pass element of LDO

That shall be a case only when you are talking of a Pmos pass device , which itself provides you some gain.(common source config)

See for non-inverting amplifier vout =vin(1+R2/R1) ... This assumes that gain is infinite. but actually its finite. So higher the gain the better.

When you have larger gain for error amp you dont bother if ur pass device is saturation or not and hence a smaller device also works. but for smaller amp gains u try to compensate for the same by keeping pass device in saturation(thus providing some 15db to 20 db gain)

hope this gives some clue.