2n7000 as a simple buffer for HF (1-30MHz)

[neazoi]

I have read somewhere that mosfets load the precending circuits much less than JFETs, so I would like to build a small buffer amplifier for HF (1-30MHz) using a mosfet such as the 2n7000. My purpose is not to load the RF input too much, whereas at the same time I will use the mosfet as an intermediate stage to convert the variable power RF input to a DC output (using diode rectifiers for AC->DC )

Can you propose such a circuit to me?
Thank you

 

[betwixt]

The 2N7000 is a small power MOSFET with high gate capacitance, it is intended for small power switching applications and isn't really suitable for RF amplification. I think the type you are thinking of is a dual-gate MOSFET which has lower gate capacitance, typically works up to several hundred MHz and most importantly has good input/output isolation.

Brian.

 

[neazoi]

The 2N7000 is a small power MOSFET with high gate capacitance, it is intended for small power switching applications and isn't really suitable for RF amplification. I think the type you are thinking of is a dual-gate MOSFET which has lower gate capacitance, typically works up to several hundred MHz and most importantly has good input/output isolation.

Brian.

Thanks for the answer.
I think what I was looking for, was a so called "active RF probe". From the circuits I have seen
**broken link removed**
http://www.next.gr/sens-detectors/rf-radiation/rf-probe-l13637.html
http://www.seekic.com/blog/project_...tive_Probe__DIY_Saves_a_Pretty_Penny_（1）.html
Only the 3rd used a dual gate mosfet. However, even not being a mosfet, the second one seems promising too for low frequencies (1-30MHz).
My main problem is the outtput voltage linearity. If the diodes are connected directly to the input, they respond quite linearly intependent of the frequency (1-30MHz). What I mean is that I get almost the same output voltage on the charge pump output if the input RF power level is the same, independent of the frequency.
I worry that the FET/mosfet solution will ruin this effect because it's amplification level would depend on the frequency.
On the other hand I do not want to connect the diodes directly because mixed signals will be induced to the input signal (being monitored).
Any comments/ideas would be appreciated.

 

[betwixt]

You could consider a combination of the circuits. The MOSFET circuit will give least loading effect on the input signal and you can use it to drive a voltage doubling rectifier like the first and second circuit. Note the second one is intended as a modulation monitor, the output coupling capacitor would have to be removed for a DC output. You could consider a small LF/DC amplier after the diodes to increase sensitivity after the rectifier instead of before it. Another possibility if you need more RF gain is to use the MOSFET input then follow it with a MMIC amplifier. They have high gain and enough 'push' to drive the rectifiers yet still only have 3 (or 4) pins.

The ideal scenario would be a buffer followed by a precision rectifier but I doubt you would get one working above a few MHz. I've never tried it myself but I imagine the capacitance of the diodes would seriously upset the feedback paths and hence the linearity as the frequency increases.

Incidentally, the schematic using the MOSFET looks suspiciously like it was copied from Elektor Magazine!

Brian.

 

[neazoi]

You could consider a combination of the circuits. The MOSFET circuit will give least loading effect on the input signal and you can use it to drive a voltage doubling rectifier like the first and second circuit. Note the second one is intended as a modulation monitor, the output coupling capacitor would have to be removed for a DC output. You could consider a small LF/DC amplier after the diodes to increase sensitivity after the rectifier instead of before it. Another possibility if you need more RF gain is to use the MOSFET input then follow it with a MMIC amplifier. They have high gain and enough 'push' to drive the rectifiers yet still only have 3 (or 4) pins.

The ideal scenario would be a buffer followed by a precision rectifier but I doubt you would get one working above a few MHz. I've never tried it myself but I imagine the capacitance of the diodes would seriously upset the feedback paths and hence the linearity as the frequency increases.

Incidentally, the schematic using the MOSFET looks suspiciously like it was copied from Elektor Magazine!

Brian.

Yes, the mosfet schematic is coppied from elektor 4/2004, I have found the actual pdf file on the net.
Ok I got the idea of how this works. I was wondering a last thing.
If I use the mosfet followed by a rectifier (no need for more gain), I worry about the impedance problems that I may cause to the transmitter output, since the mosfet will be connected parallel to it (antenna side).

Practically, is the minimal loading caused by the use of the mosfet, going to cause any problems (i.e. antenna not "see" 50R from the transmitter output?
Remember I am talking about HF, where random wire antennas are not even matched precisely to 50R.

Also, it would be helpful if you could suggest a few non-smd mosfets to try

 

[betwixt]

The input impedance of the MOSFET buffer will be in the 10s to 100s of K Ohms so it will have negligible effect on antenna matching, especially as you correctly noted, the impedance could be almost anything anyway. It would take an extremely high antenna (and therefore improbable) impedance before you noticed it being loaded.

Your bigger problem would be overloading the MOSFET but there is no way to avoid that, you just have to be careful. A buffered probe is normally used in low power circuit rather than at an antenna socket. For example, you might use it to measure LO levels or for measuring IF signals while aligning IF filters. For connection at the antenna socket the buffer isn't really necessary and you can drive the rectifiers directly, that removes the risk of damaging the mosfet as well of course.

Non-SMD MOSFETs are hard to find, the classic types are the 40673, BFR84, 3SK142. The NTE454 is still manufactured but VERY expensive.

Brian.

 

[neazoi]

Your bigger problem would be overloading the MOSFET but there is no way to avoid that, you just have to be careful.
Non-SMD MOSFETs are hard to find, the classic types are the 40673, BFR84, 3SK142. The NTE454 is still manufactured but VERY expensive.

Brian.

I guess BF981, BF966 will do as well (I have them available). The article states that the use of the mosfet is not critical anyway.
The power output of the transmitter is 100mW or so. I am thinking of using a very small value of coupling capacitor to drive the mosfet. Something below 5pF, maybe 1-2pF, so that very little power is coupled to the mosfet. What do you think about it?

At that output level (100mW) I have noticed some -40dbc harmonics appear out of nowhere when I connect the diodes directly to the transmitter output, that is why i would like to use the buffer option.

 

[neazoi]

I guess BF981, BF966 will do as well (I have them available). The article states that the use of the mosfet is not critical anyway.
The power output of the transmitter is 100mW or so. I am thinking of using a very small value of coupling capacitor to drive the mosfet. Something below 5pF, maybe 1-2pF, so that very little power is coupled to the mosfet. What do you think about it?

At that output level (100mW) I have noticed some -40dbc harmonics appear out of nowhere when I connect the diodes directly to the transmitter output, that is why i would like to use the buffer option.

Just a note, I have tried J310 and although it is not a mosfet, I can measure no unwanted signals at the transmitter output, so I guess it is sufficient.

 

[tggzzz]

If you are building an active probe, you might find some useful pointers at https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/