Shorting an AC signal to ground

[evandude]

For an RFID project, I need to be able to selectively short an AC input signal to ground (this is to short the input coil terminals as well as possible, which is how the tag transmits data back to the reader)

I use the term "short" loosely here, as there will of course still be a voltage drop over the diode and transistors, however it's not significant compared to the amplitude of the input signal it is "shorting" (which has been 80-100Vpp in testing)

so far the best i've come up with is this. the diodes make it so each transistor shorts the input signal for half its cycle (positive or negative side)

it seems like a usable design, the only problem is that A) it doesn't seem very elegant, and B) it requires a negative rail for the PMOS gate.

I can have a negative power rail, but that's just about all it would be used for so if I could do this differently and eliminate it, that would be preferable.

I am looking into the use of triacs, as they seem like they could do what I need. Specifically, I'm looking at opto-triacs so I can drive them with a logic output... But I'm not familiar with triacs at all, so maybe you guys can shed some light on whether these would be of any use to me. I'm worried about the amount of current it takes for the output to "function", as the amount of current coming from this coil is going to be pretty weak (haven't quantified it yet though)

 

[Optikon]

For an RFID project, I need to be able to selectively short an AC input signal to ground (this is to short the input coil terminals as well as possible, which is how the tag transmits data back to the reader)

I use the term "short" loosely here, as there will of course still be a voltage drop over the diode and transistors, however it's not significant compared to the amplitude of the input signal it is "shorting" (which has been 80-100Vpp in testing)

so far the best i've come up with is this. the diodes make it so each transistor shorts the input signal for half its cycle (positive or negative side)

it seems like a usable design, the only problem is that A) it doesn't seem very elegant, and B) it requires a negative rail for the PMOS gate.

I can have a negative power rail, but that's just about all it would be used for so if I could do this differently and eliminate it, that would be preferable.

I am looking into the use of triacs, as they seem like they could do what I need. Specifically, I'm looking at opto-triacs so I can drive them with a logic output... But I'm not familiar with triacs at all, so maybe you guys can shed some light on whether these would be of any use to me. I'm worried about the amount of current it takes for the output to "function", as the amount of current coming from this coil is going to be pretty weak (haven't quantified it yet though)

You dont mention duty cycle or speed so...

Why not a small mechanical relay? You can get them rated for well beyond 100V and the parasitics in the "off" condition will be much better than fets.

 

[evandude]

You dont mention duty cycle or speed so...

Why not a small mechanical relay? You can get them rated for well beyond 100V and the parasitics in the "off" condition will be much better than fets.

sorry, you're right, that would be important :lol:

the tag needs to be as small as possible, first of all... and secondly it needs to be able to switch at 15.625 KHz at 50% duty. Believe me, if I could use a relay I'd be all over it.

 

[Optikon]

You dont mention duty cycle or speed so...

Why not a small mechanical relay? You can get them rated for well beyond 100V and the parasitics in the "off" condition will be much better than fets.

sorry, you're right, that would be important :lol:

the tag needs to be as small as possible, first of all... and secondly it needs to be able to switch at 15.625 KHz at 50% duty. Believe me, if I could use a relay I'd be all over it.

well the fets dont seem a bad idea.. Sounds like you need a relatively fast, 100V analog switch... two FETS seems like a good place to start.

 

[Nigel Goodwin]

the tag needs to be as small as possible, first of all... and secondly it needs to be able to switch at 15.625 KHz at 50% duty. Believe me, if I could use a relay I'd be all over it.

Have you considered the use of PIN diodes?, they are commonly used for TX/RX switching in amateur radio equipment. Google will find suitable examples - and I'm sure it's come up once before in a thread?.

Also, why that exact frequency? - particularly for someone in the USA?.

It's the exact line frequency of 625 line PAL TV's.

 

[evandude]

the system i'm making uses a 125khz carrier, and I'm using FSK modulation, so 0 or 1 is represented by attenuations at a frequency of 1/10 or 1/8 the carrier frequency, so 12.5KHz and 15.625KHz respectively.

 

[audioguru]

Optically-isolated Mosfets are available. They use a fairly high-voltage solar cell to feed enough voltage to their gate. Their input is just a 1.3V LED. You must check that their max voltage rating is high enough and speed is fast enough, then two could be wired back-to-back for your AC signal. :lol:

 

[evandude]

I found an opto-isolator with a FET output, and from everything i could tell it was bi-directional so I could use it directly between the AC signal and ground, nothing else needed! the only problem is that the breakdown voltage is 30 volts, so i'd probably need to protect it with a zener diode, and i'm not entirely sure how well my detector circuit in the reader would deal with that. but it's a possibility...

 

[audioguru]

I found an opto-isolator with a FET output...

H11F1? No, I was talking about an opto-Mosfet made by International Rectifier I think. I think it has a fairly high voltage rating and high power rating in a TO-220 case. Maybe it is also available as a tiny surface-mount.

I tried the H11F1 to control the level of a Wien-bridge low-distortion oscillator. It is bi-directional but behaves strangely when its output voltage exceeds only about 60mV. It produced the opposite of crossover distortion, the peak voltages of the sine-wave were expanded! I even tried one in series then another as a shunt, hoping that their distortions would cancel. They didn't. I tried modulating the LED current but gave-up and used an ordinary FET.

 

[evandude]

Well I think i've solved the problem. I set it up with only the NMOS side of that circuit, so it only attenuates the positive half of the wave... and my reader circuit on the other side gave me an output that was almost the same as if i just manually shorted out the coil with a wire. So, I guess I don't need to be attenuating the whole wave, so I can avoid having to use a PMOS and a negative supply rail to go with it.

Love it when things work out the easy way :lol:

 

[Roff]

Another option is to tap the coil, and short the tap. It's an autotransformer. You trade voltage for current - lower voltage, higher current. Your voltage probably isn't high enough to bother.