The Designer's Guide Community Forum https://designers-guide.org/forum/YaBB.pl Design >> RF Design >> How to block RF signal? https://designers-guide.org/forum/YaBB.pl?num=1384836710 Message started by shaikss_rf on Nov 18th, 2013, 8:51pm

Title: How to block RF signal? Post by shaikss_rf on Nov 18th, 2013, 8:51pm Hi, I have designed two rectifier topologies. One topology works well and provides good efficiency for voltage range 600mV to 800mV. The other topology provides better efficiency for 850mV to 1.2V. Now, I want to integrate these two topologies by sensing the input/output voltage/current. If I sense the output voltage, I can know the input voltage range. So, depending on this, I can enable either of the above topologies. When I enable any of the above topologies, I should be able to block/cut the RF signal going to another topology as the same RF signal is going to both topologies. I have attached two topologies. How can I do this?

Title: Re: How to block RF signal? Post by aaron_do on Nov 18th, 2013, 11:52pm use a switch... Aaron

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 1:02am I used a switch (NMOS and PMOS). It should be either one of the topologies is ON. But in my case, the other topology is also ON.

Title: Re: How to block RF signal? Post by aaron_do on Nov 19th, 2013, 1:28am You mean both are on? How did you design your switch? Read up on SPDT switch design. You need a series and shunt branch, and you shouldn't need any PMOS. Depends on your operating frequency, but in the low GHz range you should be able to get around 30 dB isolation... regards, Aaron

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 1:42am I tried with PMOS switch. I connected PMOS switch to the cap where RF input is going. In total I have used 4 PMOS switches: 2 switches for each topology. I have given RF input to its source and the gate is enabled/disabled. Can you pls throw more light on the technique you had suggested.

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 1:42am aaron_do wrote on Nov 19th, 2013, 1:28am: You mean both are on? How did you design your switch? Read up on SPDT switch design. You need a series and shunt branch, and you shouldn't need any PMOS. Depends on your operating frequency, but in the low GHz range you should be able to get around 30 dB isolation... regards, Aaron I am using 865MHz.

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 1:43am shaikss_rf wrote on Nov 19th, 2013, 1:42am: aaron_do wrote on Nov 19th, 2013, 1:28am: You mean both are on? How did you design your switch? Read up on SPDT switch design. You need a series and shunt branch, and you shouldn't need any PMOS. Depends on your operating frequency, but in the low GHz range you should be able to get around 30 dB isolation... regards, Aaron I am using 865MHz. How can I simulate in cadence?

Title: Re: How to block RF signal? Post by aaron_do on Nov 19th, 2013, 2:01am Why are you using PMOS? NMOS is better for this application. Here's an example: http://www.rfmd.com/cs/documents/CommATombakIMS12.pdf Note the floating gate. You may need a floating body too. Make sure your transistor is properly biased. Aaron

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 2:11am I am designing a rectifier for passive tag applications. I don't have any voltage source/battery to provide gate voltage to switch ON NMOS. So, I selected PMOS switch. Whatever dc voltage I get, it is the rectified dc voltage.

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 2:47am aaron_do wrote on Nov 19th, 2013, 2:01am: Why are you using PMOS? NMOS is better for this application. Here's an example: http://www.rfmd.com/cs/documents/CommATombakIMS12.pdf Note the floating gate. You may need a floating body too. Make sure your transistor is properly biased. Aaron Aaron, I tried to use SPDT switch. Here also, I see the same problem. The other topology is sucking some current and I do see some negative voltage at the switch output (which is acting as input to other topology).

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 3:39am Hi, Using SPDT NMOS switch, I am able to see enabling and disabling topologies to some extent. But I am unable to replicate the same using PMOS swicthes. Request you to please help me on this

Title: Re: How to block RF signal? Post by aaron_do on Nov 19th, 2013, 6:38am Hi, Quote: I am designing a rectifier for passive tag applications. I don't have any voltage source/battery to provide gate voltage to switch ON NMOS now your question is starting to become more clear. Quote: So, I selected PMOS switch. Whatever dc voltage I get, it is the rectified dc voltage. I don't understand why you can use PMOS, but not NMOS. Both require a DC drop from the gate to the source/drain. Quote: The other topology is sucking some current and I do see some negative voltage at the switch output The negative voltage is either caused by your rectifier, or the SPDT. It could be due to diode (body diode) turn on in the SPDT. Normally you need to bias the gate/body negative wrt the drain/source in order to properly turn off the switch. Otherwise a large input swing could turn on the switch when its supposed to be off. The same phenomenon could be causing the current draw. Instead of biasing the gate/body at a negative voltage, some designs ground the gate/body, and pull up the source/drain. If you show your waveforms I might be able to provide better insight into whats happening. Aaron

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 8:55pm aaron_do wrote on Nov 19th, 2013, 6:38am: Hi, Quote: I am designing a rectifier for passive tag applications. I don't have any voltage source/battery to provide gate voltage to switch ON NMOS now your question is starting to become more clear. [quote]So, I selected PMOS switch. Whatever dc voltage I get, it is the rectified dc voltage. I don't understand why you can use PMOS, but not NMOS. Both require a DC drop from the gate to the source/drain. Quote: The other topology is sucking some current and I do see some negative voltage at the switch output The negative voltage is either caused by your rectifier, or the SPDT. It could be due to diode (body diode) turn on in the SPDT. Normally you need to bias the gate/body negative wrt the drain/source in order to properly turn off the switch. Otherwise a large input swing could turn on the switch when its supposed to be off. The same phenomenon could be causing the current draw. Instead of biasing the gate/body at a negative voltage, some designs ground the gate/body, and pull up the source/drain. If you show your waveforms I might be able to provide better insight into whats happening. Aaron [/quote] I have attached the circuit and the corresponding waveforms. In 3.png, I could see that the rectifier topology which I want to enable is enabled. But how can I avoid those negative voltages in 1.png and 3.png. 6.png is the SPDT switch using PMOS. Please tell me how to modify the circuit when I use PMOS as I don't have any voltage source to design SPDT switch using NMOS. 2.png is the circuit diagram of 1.png waveform. 3.png is the output waveforms when I integrate two rectifier topologies. 4T_out is the topology which i had enabled. 5.png is the waveforms of 6.png circuit.

Title: Re: How to block RF signal? Post by aaron_do on Nov 19th, 2013, 9:15pm Hi, your switch isolation isn't good because you only have a series branch. Adding a shunt branch should make a big difference. The negative voltages are probably coming from the transistor partially turning on when the input voltage goes below zero. This appears to have a kind of charge-pump effect. The way to avoid this is to bias your gate and body -ve wrt the input (i.e. -1 V for example). This might not be a viable solution in your case. An alternative would be to bias the source and drain of the switch's off branch at +1V and the gate/body at 0V. There is another solution I know, but its quite difficult to explain, so I suggest you read about switch design. There are plenty of papers on IEEExplore. As for whether you use NMOS or PMOS, it makes no difference, Both cases require a supply voltage. You probably need some kind of system solution to manage your power supply. Aaron

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 9:32pm aaron_do wrote on Nov 19th, 2013, 9:15pm: Hi, your switch isolation isn't good because you only have a series branch. Adding a shunt branch should make a big difference. The negative voltages are probably coming from the transistor partially turning on when the input voltage goes below zero. This appears to have a kind of charge-pump effect. The way to avoid this is to bias your gate and body -ve wrt the input (i.e. -1 V for example). This might not be a viable solution in your case. An alternative would be to bias the source and drain of the switch's off branch at +1V and the gate/body at 0V. There is another solution I know, but its quite difficult to explain, so I suggest you read about switch design. There are plenty of papers on IEEExplore. As for whether you use NMOS or PMOS, it makes no difference, Both cases require a supply voltage. You probably need some kind of system solution to manage your power supply. Aaron For PMOS, I need -ve voltage. So, I need to go for negative voltage converter. Is there any way to generate negative voltage within the system using minimal number of components? Can you pls give me some hint on the other idea?

Title: Re: How to block RF signal? Post by aaron_do on Nov 19th, 2013, 10:18pm Hi, you can use a -ve charge pump. It requires a ring oscillator an the charge pump itself. Why do you need a minimum number of components? You're in the wrong design age if you're worried about complexity. Problem is it needs to be on as long as you want to enable/disable, so power consumption is an issue. I still don't get why you're so keen on using PMOS... Aaron

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 19th, 2013, 10:27pm aaron_do wrote on Nov 19th, 2013, 10:18pm: Hi, you can use a -ve charge pump. It requires a ring oscillator an the charge pump itself. Why do you need a minimum number of components? You're in the wrong design age if you're worried about complexity. Problem is it needs to be on as long as you want to enable/disable, so power consumption is an issue. I still don't get why you're so keen on using PMOS... Aaron I am not worried about circuit complexity. I am worried about power consumption. Since it is  a passive tag and it harvests energy, whatever power I harvest, some of the harvested energy is utilized for Chargepump and other devices. Even I am not keen about PMOS. As u said, it is PMOS or NMOS, I need voltage source. The only Q is who wil provide the voltage. During energy harvesting, Rf input signal is in the order of few mV and the control voltage for NMOS SPDT switch is more than the peak of the RF input. I am thinking of using a start-up circuit to generate required voltage for SPDT switch. Once I get some voltage, then I can enable/disable the switching scheme. Pls correct me if I am wrong.

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 20th, 2013, 1:50am Hi, I have one basic query. For NMOS, bulk is connected to Vss and PMOS is connected to Vdd. In case of ac signals, RF signal goes negative. In such scenarios, the most negative voltage is the negative peak of RF signal. How should the bulk connected in such scenarios?

Title: Re: How to block RF signal? Post by aaron_do on Nov 20th, 2013, 1:57am I think you should try reading my earlier replies. I told you the answer several times. Aaron

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 20th, 2013, 2:26am aaron_do wrote on Nov 20th, 2013, 1:57am: I think you should try reading my earlier replies. I told you the answer several times. Aaron Sorry for repeatedly asking the same. I understood your point - Pull-up or down to either S/D terminal. But though I do, pull-up /down, I see bulk current. Earlier, when I connected bulks to ground, I could see bulk current of 200uA. So, I thought of nullifying it. I tried to do pull up/down, but no significant result. Not getting how to solve.

Title: Re: How to block RF signal? Post by shaikss_rf on Nov 20th, 2013, 2:40am shaikss_rf wrote on Nov 20th, 2013, 2:26am: aaron_do wrote on Nov 20th, 2013, 1:57am: I think you should try reading my earlier replies. I told you the answer several times. Aaron Sorry for repeatedly asking the same. I understood your point - Pull-up or down to either S/D terminal. But though I do, pull-up /down, I see bulk current. Earlier, when I connected bulks to ground, I could see bulk current of 2uA. So, I thought of nullifying it. I tried to do pull up/down, but no significant result. Not getting how to solve. This is how bulk currents are...

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