can i set the current mirror ratio to 20:1

[foreverloves]

If i need a steady but not an accuracy current, can i set the current mirror ratio to
20:1?

for example, I have a current source of 200uA, but I need a current of 10uA(10.5, or 9.6 are also ok, accuracy is not required, I need it to be steady), can i
set the current mirror ratio to 20:1?

can there be any other problems ?

 

[MSSN]

as far as i know since accuracy is not required then there is no problem

 

[ebalboni]

With CMOS you should be ok scaling down currents. Scaling up can be an issue because any input noisae will alsi scale up.

 

[mirja]

definitely you can. In fact, with this ratio you can even scale up, with proper care in layout, quite accurately.

 

[wpchan05]

In terms of layout (or reliability) issue, I would recommend 200uA to 50uA, then 50uA to 10uA.

 

[eng_Semi]

sure u can, but assure current mirror matching in the layout

 

[davidwong]

Hi wpchan05
why u recommend 200uA to 50uA, then 50uA to 10uA in term of layout.
better matching can be got with low current ratio ?

 

[husseinadel]

i am using ratio of 100!!! ,, accuracy idnt concern me much as long as i am within the specified limits,,,as in my application i depend on the current gain ,,i.e. as it is bigger it is better,,,,if anyone can tell me the problem of that it would be great,,,

 

[wpchan05]

Hi wpchan05
why u recommend 200uA to 50uA, then 50uA to 10uA in term of layout.
better matching can be got with low current ratio ?

I always limit all my current mirror ratio up to 5 or down to 1/5 so as to reduce parameters variation of the transistors within a chip. Although the first author mentioned that a high precision current gain is not required in his design, I have no confidence on expecting a current mirror with a ratio of 20:1 can achieve 95% accuracy, from a statistical point of view.

 

[husseinadel]

Hi wpchan05
why u recommend 200uA to 50uA, then 50uA to 10uA in term of layout.
better matching can be got with low current ratio ?

I always limit all my current mirror ratio up to 5 or down to 1/5 so as to reduce parameters variation of the transistors within a chip. Although the first author mentioned that a high precision current gain is not required in his design, I have no confidence on expecting a current mirror with a ratio of 20:1 can achieve 95% accuracy, from a statistical point of view.

well , i f i want a certain current , then can i make sweep to get it regardless the ratio of current mirror ? is this will create a problem??,,