The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> Bias switching
https://designers-guide.org/forum/YaBB.pl?num=1328258102

Message started by DDC on Feb 3^rd, 2012, 12:35am

Title: Bias switching
Post by DDC on Feb 3^rd, 2012, 12:35am

Hello,

I have a problem. I try to switch on/off my bias networks. I use a second BJT to perform a current mirror with the input BJT of a cascode and a transmissiongate to short circuit the base of the input cascode BJT. Discharging is performed very well in a short time (some ps) but charging takes too long because of my DC decoupling cap (10pF). The problem is if I use a very small cap (some hundret fF) I can lower the settling time but my impedance matching is completly mismatched because my frequency ist only 900 MHz. Is there somebody who has an idea how I can speedup the charging of the cap or how I can bias my cascode without any DC blocking cap ? I think that I could also bias the BJT with a simple DC voltage at the base but this is not usefull due to bias jitter and so the collector current will change exponentially. The main problem is that other solutions shuoldn't cause any further power dissipation. Thanks for response.

Title: Re: Bias switching
Post by aaron_do on Feb 3^rd, 2012, 2:03am

Hi,

I suggest you post a schematic. Or take some time to describe your circuit better. For instance...

- Is it an LNA?
- Is it a cascode LNA?
- Is it differential or single-ended?

regards,
Aaron

Title: Re: Bias switching
Post by DDC on Feb 3^rd, 2012, 6:16am

Yes it should be a LNA. I need to switch off the bias but the DC decoupling cap limits the time for charging until the circuit can reach his bias point.

Title: Re: Bias switching
Post by aaron_do on Feb 3^rd, 2012, 6:36pm

Hi,

based on your schematic, you could simply switch on/off the cascode transistor and leave the bottom one permanently on. But I think your cascode transistor probably needs a large cap to ground too unless Vb is really ideal. One thing you could consider if you want to speed up the charging is to replace Rbias/10 with a switch, and turn it ON during the charge up phase and OFF once the node is fully charged.

Also, I doubt you need a full transmission gate for the switching. You can just use an NMOS.

regards,
Aaron

Title: Re: Bias switching
Post by DDC on Feb 6^th, 2012, 12:14am

Hello Aaron,

thanks for response. I have further the problem, that the DC decoupling cap (C_DC ) has to be very high (> 100pF ) that it has less influence on matching but it has to be small for a short settling time. Do you have an idea how I can adjust the operating point without using a DC blocking cap ? I could also use a common base stage but there is the same problem. Do you have any idea ?

Title: Re: Bias switching
Post by loose-electron on Feb 6^th, 2012, 3:11pm

Consider a non linear circuit that forces the bias upon startup?

Comparator methods to a reference circuit, or
boost the current for a brief period of time.

That is a rather large capacitor for inside a chip,
so is that the case for what reason? KT/C ?

Title: Re: Bias switching
Post by aaron_do on Feb 6^th, 2012, 5:28pm

Hi DDC,

you could do a few things.

1) use a differential topology with transformers at input and output to avoid large blocking caps. Or use differential with no gate cap for the cascode and simply switch the cascode.

2) As previously suggested, briefly short the biasing resistor during start-up.

3) Switch the gate voltage of the cascode transistor to either ground (OFF) or bias (ON). Make sure the bias is properly AC grounded, and use a large enough switch to minimize the ON resistance. Of course in this case, your biasing circuit would have to be ON all the time.

regards,
Aaron