| Author |
Message |
lhlbluesky
Joined: 30 Mar 2007
Posts: 152
|
 10 May 2008 18:30 about layout of mirror |
|
|
|
| to mirror current exactly,the ratio of W/L of the mirrored and mirroring transistor should be exact,too;i want to ask ,for W/L=50/0.5 with finger number=10 (mirroring transistor)and W/L=10/0.5 with finger number=2 (mirrored transistor),how is the layout?if i use the technique of sharing source and drain,the parasitic cap may be different,and the ratio can be not exact;how to solve it.pls.
|
|
| Back to top |
|
 |
gte582w
Joined: 01 Sep 2006
Posts: 96
Helped: 7
|
| 12 May 2008 22:43 about layout of mirror |
|
|
|
your ration is a little too large: 500/20
try unit element and cc techniques etc.
|
|
| Back to top |
|
|
jecyhale
Joined: 19 Feb 2008
Posts: 396
Helped: 39
Location: China
|
| 13 May 2008 4:44 Re: about layout of mirror |
|
|
|
| lhlbluesky wrote: |
| to mirror current exactly,the ratio of W/L of the mirrored and mirroring transistor should be exact,too;i want to ask ,for W/L=50/0.5 with finger number=10 (mirroring transistor)and W/L=10/0.5 with finger number=2 (mirrored transistor),how is the layout?if i use the technique of sharing source and drain,the parasitic cap may be different,and the ratio can be not exact;how to solve it.pls. |
Usually the parasitic of Current mirror is not critical.
And is 50/0.5 is the total size?
if it is 500/0.5 and the other is 20/0.5, the ratio will be too large,
you'd better use the high swing cascode mirror.
And 0.5u length is too small for current mirror.
Beside, which circuit does the current mirror use?
|
|
| Back to top |
|
|
sgoat
Joined: 30 Nov 2006
Posts: 2
|
| 18 May 2008 19:00 Re: about layout of mirror |
|
|
|
Usually the parasitic of Current mirror is not critical.
And is 50/0.5 is the total size?
if it is 500/0.5 and the other is 20/0.5, the ratio will be too large,
you'd better use the high swing cascode mirror.
And 0.5u length is too small for current mirror.
Beside, which circuit does the current mirror use?[/quote]
I think 50/0.5 is the total size,If it is 500/0.5,the parameter would be M=10,but here it is nf=10,the finger width is 5/0.5
I have met the same poblem when designing a current mirror with chartered 0.13um process. When I used the 'w=10u ,M=5' ,total width is 50u, the current will be well matched,but when I use 'w=50u,nf=5',it has a great mismatch about 10%.
I think the layout is like 'w=50u,nf=5'.
I also want to kown,which one I shoud use,thanks
|
|
| Back to top |
|
|
quaternion
Joined: 12 Nov 2006
Posts: 197
Helped: 11
Location: Cairo , Egypt
|
| 18 May 2008 22:04 Re: about layout of mirror |
|
|
|
| sgoat wrote: |
I have met the same poblem when designing a current mirror with chartered 0.13um process. When I used the 'w=10u ,M=5' ,total width is 50u, the current will be well matched,but when I use 'w=50u,nf=5',it has a great mismatch about 10%.
I think the layout is like 'w=50u,nf=5'.
I also want to kown,which one I shoud use,thanks |
I think for good matching of current mirrors they should have same W , finger no.s , & length .While varying the no. of units (M) to vary the mirroring ratio.
|
|
| Back to top |
|
|
rosaeidi
Joined: 19 Dec 2006
Posts: 97
Helped: 5
|
| 22 May 2008 10:19 about layout of mirror |
|
|
|
Hi
Using large values for w and l would result in better matching. The parasitic caps can restrict your BW not your matching.
|
|
| Back to top |
|
|
