delay time in current biasing circuits due to startup circuit

[Junus2012]

Dear friends,

I am testing the startup circuit operation for my PTAT current generator, I have found at some corners the circuit will take 5 ms to respond and settle to the steady state desirable current.
I would like to ask you hence if 5 ms is normal or I have to consider changing my startup circuit if you think it is unusual.

Just in my mind that startup condition is only happens at powering on the circuit, so in my opinion it should not effect, but this is the first time I face such a slow response

for your information, I am using the dynamic startup circuit shown below, but my PTAT circuit is based on wide swing mirrors

Thank you
Best Regards

 

[sutapanaki]

It's kind of slow, but as you said it is needed at power up only, so if you can tolerate this delay, then it shouldn't be a problem. Maybe your capacitor in the start up takes long to charge. Perhaps Ms1 is too week or Ms2 too strong.

 

[FvM]

Classical start-up circuits in analog design text books are using more clever methods than a big capacitor.

 

[Junus2012]

or Ms2 too strong.

it should respond faster if this transistor is strong because it will inject more startup current (strong means higher W/L)

Classical start-up circuits in analog design text books are using more clever methods than a big capacitor.

Before considering to change the startup circuit, see please the current one,

what are the best node for injecting the startup currents? is there a difference between sinking or sourcing current injection to the circuit?

 

[sutapanaki]

When I said strong transistor, I meant that it will need higher capacitor voltage to shut it off, which means longer time until the current generator starts producing the intended current.