The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Analog Switch with small THD
https://designers-guide.org/forum/YaBB.pl?num=1262885649

Message started by rajdeep on Jan 7th, 2010, 9:34am

Title: Analog Switch with small THD
Post by rajdeep on Jan 7th, 2010, 9:34am

Hi all,

Can anyone provide me with some literature of designing analog switch with small THD (total harmonic distortion)?  I got some datasheets. But need some design guidance,

Thanks!
Rajdeep

Title: Re: Analog Switch with small THD
Post by vivkr on Jan 8th, 2010, 2:04am

That is a very vague question. But take a look at the JSSC paper by Dessouky and Kaiser (Mar. 2001).
An excellent review of the subject may be found in the Ph.D thesis of Mikko Waltari (Helsinki Univ. of Technology).

I don't know if you can find that thesis online still, but it contains a comprehensive overview of the subject matter.

Regards,

Vivek

Title: Re: Analog Switch with small THD
Post by rajdeep on Jan 8th, 2010, 3:32am

Vague!! :o
Whatever, thanks for the pointers!

Just to remove the vagueness, I am looking for design methods of such switches
http://www.semiapps.com/High_performance_analog_switches_improve_cell_phone_design.aspx

Ive found some formula relating Ron and THD of a CMOS switch also, from a company  application note. But not sure how they are derived!!

But thanks for the paper link. It talks about bootstrapped switches. A starting point for sure.

Cheers!
Rajdeep

Title: Re: Analog Switch with small THD
Post by bala on Jan 8th, 2010, 3:09pm

Yeah, bootstrapping reduces the input dependence and hence non-linearity. Some good starting points for discussion on switches, I'd say are Razavi's "Principles of Data Conversion Systems". Some good discussions can also be found in EECS 240 lectures. Once this base has formed, i'd go for Vivek's references.

Title: Re: Analog Switch with small THD
Post by vivkr on Jan 11th, 2010, 12:31am


rajdeep wrote on Jan 8th, 2010, 3:32am:
Vague!! :o
Whatever, thanks for the pointers!

Just to remove the vagueness, I am looking for design methods of such switches
http://www.semiapps.com/High_performance_analog_switches_improve_cell_phone_design.aspx

Ive found some formula relating Ron and THD of a CMOS switch also, from a company  application note. But not sure how they are derived!!

But thanks for the paper link. It talks about bootstrapped switches. A starting point for sure.

Cheers!

Rajdeep


There are 2 different types of switches listed here. What app are you targetting? The use of switched bootstrapped switches may or may not be necessary and/or optimal depending on the exact application constraints. You need to weigh the benefits vs. possible issues to look out for.

Here are the benefits of bootstrapping:

1. More linear Ron => you may be able to use a smaller switch with higher Ron because it has lower THD.

2. Due to the above, you also will have lower capacitance,.

Issues to look out for:

1. You cannot use a bootstrapped switch like the one in most of the references I have listed because they are clocked, and it seems to me that you need to pass the signal continuously.

2. You could put 2 switches in || and switch them alternately to maintain a continuous signal path but there will be some glitches, which may or may not affect signal fidelity depending on the clocking rate and the glitch frequency.

3. Power handling may be an issue. Your switch may need to be much larger than what mere Ron or THD constraints dictate because it is carrying several watts to the speaker for instance.

Another option might be (if your signal is not rail-rail or so) to build an a good level shifter that tracks the source end of the MOS switch and drives the gate so as to maintain VGS constant. This has the disadvantages of requiring significant amount of power and will also add more cap loading from the level shifter.

All this is assuming of course that you want to make an on-chip MOS switch. If you are planning to use off-chip alternatives or so, then there may be other possibilities.

Regards,

Vivek



Title: Re: Analog Switch with small THD
Post by rajdeep on Jan 12th, 2010, 1:59am

Thanks a lot!!

Lots of pointers to start with.

And yes, Vivek, for the audio signal the switching should be continuous, but then there
could be some sampling freq. Im not sure. This is something I am just starting to have a look at.

Thanks!
Rajdeep

The Designer's Guide Community Forum » Powered by YaBB 2.2.2!
YaBB © 2000-2008. All Rights Reserved.