Simple RF power indicator with sharp cut-off?

[neazoi]

Hello, I have built a 200mW max @50R HF transmitter (1-30MHz) which has an adjustable output power (1mW-200mW).
From band to band, the output power level changes, so I have to re-adjust the power properly.

I would like to find a way to manually set the output power to 100mW at each band without the need for an oscilloscope each time. The simplest way is a rectifier diode followed by an analogue meter. However these meters are more expensive than my whole circuit, so I would like something that can be done with a LED.

The problem is that I need it to be quite sharp, i.e the led to be switched fully on near 100mW and switched completely off below that.

Can this be done in a simple way? Discrete components is my favourite, if it can be done.

 

[FvM]

It surely can be done. One day you'll start to design elctronic circuits on your own. (I hope so).

 

[neazoi]

It surely can be done. One day you'll start to design elctronic circuits on your own. (I hope so).

Let me rephrase the question.
HOW can this be done in a simple way?

 

[volker@muehlhaus]

The problem is that I need it to be quite sharp, i.e the led to be switched fully on near 100mW and switched completely off below that.

Consider using a comparator (opamp) to detect of your rectified DC is above/below a limit.

 

[neazoi]

Consider using a comparator (opamp) to detect of your rectified DC is above/below a limit.

The only reason I avoid an opamp is to see if this can be done simpler, since only one level is to be detected.
I have seen this zener solution http://www.electroschematics.com/5816/led-volt-meter/ and I wonder how sharp it should be?
Also to minimize circuit loading such a solution could be possible http://www.electroschematics.com/5703/battery-status-indicator/ but with the transistor collector connected to the VCC separately, so that the loading to the input circuit is minimized.
I think that this could be done with a FET as well http://www.circuitsproject.com/2014/06/sound-vu-meter-circuit-diagram.html so as to even more minimize loading.
What do you think?

 

[FvM]

A discrete solution with sharp characteristic spells "use a BJT as level switch".

 

[neazoi]

A discrete solution with sharp characteristic spells "use a BJT as level switch".

Given that a single bjt is not only an on/off switch but it has a "transition" level where it is used as an amplifier, how should I force this transition level to be small and adjustable? A schmitt trigger circuit that can be adjusted to a desired threshold level would be something I need.

 

[FvM]

A schmitt-trigger (a two transistor circuit) is an option, but not necessarily required, I think.

It's useless to attempt a level detector accuracy exceeding the detector accuracy by a large factor. 1 dB accuracy (+/-10 % level) seems a reasonable ballpark figure.

 

[volker@muehlhaus]

The only reason I avoid an opamp is to see if this can be done simpler, since only one level is to be detected.

If "simpler" means to use many components instead of one simple IC, making it larger and more expensive: yes, you can do that.

To me, opamps are basic building blocks that can simply your circuit compared to all-discrete circuits. Just like we use a voltage regulator instead of building it from discrete components.

 

[neazoi]

A schmitt-trigger (a two transistor circuit) is an option, but not necessarily required, I think.

It's useless to attempt a level detector accuracy exceeding the detector accuracy by a large factor. 1 dB accuracy (+/-10 % level) seems a reasonable ballpark figure.

I think this is what I was looking for https://www.johnhearfield.com/Eng/Schmitt.htm
It seems I was looking for a schmitt trigger with a settable threshold instead of a level detector.
A simple diode detector at the imput, shunted with a capacitor would probably do the trick?

Thanks all for your replies

 

[neazoi]

If "simpler" means to use many components instead of one simple IC, making it larger and more expensive: yes, you can do that.

To me, opamps are basic building blocks that can simply your circuit compared to all-discrete circuits. Just like we use a voltage regulator instead of building it from discrete components.

I have tested this circuit and it works very well for the purpose.
The only thing that I notice, is that when the circuit is connected to the transmitter output, there are some low <-35dbc signals induced to the spectrum, probably due to the non-linearity effects of the diodes. In silicon diodes the case is worst, however I have not still tested low frequency rectifier silicon diodes (1n400x would be siotable?).

I think I saw somewhere that a single transistor buffer amplifier before the diodes can solve this problem but I cannot find the link now. I was also thinking if a FET is better, to minimise loading of the transmitter output.
Any thoughts or suggestions on this wouold be appreciated.

 

[volker@muehlhaus]

If you have more level than you need for the detector, how about using an attenuator in front if the detector? This will minimize the distortion caused by the detector.

If I correctly understand your use for the detector, it will be at the transmitter output in parallel to the load (antenna). In this case, add some (large) series resistance between the transmitter and the detector input, and some shunt resistor to ground at the detector input -> voltage divider. Something like 470Ohm series and 47 Ohm shunt. This provides a more correct load to the transmitter output than antenna||detector, and it adds isolation for harmonics created by the detector.

 

[neazoi]

If you have more level than you need for the detector, how about using an attenuator in front if the detector? This will minimize the distortion caused by the detector.

If I correctly understand your use for the detector, it will be at the transmitter output in parallel to the load (antenna). In this case, add some (large) series resistance between the transmitter and the detector input, and some shunt resistor to ground at the detector input -> voltage divider. Something like 470Ohm series and 47 Ohm shunt. This provides a more correct load to the transmitter output than antenna||detector, and it adds isolation for harmonics created by the detector.

Thanks a lot! I will try these ideas.
What if I use a fet as a buffer? It usually includes a 1-10Mohm resistor in series with the gate. I believe the whole setup will be nearly perfect for the purpose, as the fet provides ultimate isolation, but since I am not sure, a second oppinion yould be helpfull.
I am thinking something as simple as this
http://www.techlib.com/area_50/ghostdetector.htm (the first stage of course and without the neon). It may need to be coupled by a very low capacitor value (<5pf?) to the transmitter output so that the effect will be negligible.

 

[jiripolivka]

Thanks a lot! I will try these ideas.
What if I use a fet as a buffer? It usually includes a 1-10Mohm resistor in series with the gate. I believe the whole setup will be nearly perfect for the purpose, as the fet provides ultimate isolation, but since I am not sure, a second oppinion yould be helpfull.
I am thinking something as simple as this
http://www.techlib.com/area_50/ghostdetector.htm (the first stage of course and without the neon). It may need to be coupled by a very low capacitor value (<5pf?) to the transmitter output so that the effect will be negligible.

In all above discussion I failed to see the most important point: First you need a calibrated detector, with a response "perfectly" flat over 1...30 MHz. I am afraid such detector is difficult to find or make.
Diode detectors used as RF power meters are on the market but expensive. Their response may vary by more than +/- 1 dB, which translates to +/- 12% of power in mW.
Then with such detector plus a power divider (to relieve detector load on generator output), including temperature variations, only then one can add a level comparator to indicate a power level exceeded a preset limit. Or two comparators to indicate a power level within a window.

I think a very simple solution as required does not exist.

 

[BradtheRad]

Thanks a lot! I will try these ideas.
What if I use a fet as a buffer? It usually includes a 1-10Mohm resistor in series with the gate. I believe the whole setup will be nearly perfect for the purpose, as the fet provides ultimate isolation, but since I am not sure, a second oppinion yould be helpfull.

I used an fet for the input when I built an rf detector (for 49MHz). The signal was picked up by a normal antenna and tunable LC tank. Then came the fet. This drove a transistor. I connected my VOM to take readings.

To turn off the fet, it needed a negative signal at the bias. By adjusting a 1M potentiometer attached across 2 batteries, I managed to get a zero reading when no signal was present. Then when a signal was broadcast, the fet responded only to the positive half of the waveform. No diode was needed.

With stronger rf signal strength, my meter reading went upscale.

An led would be just as good for an indicator. It has a built-in voltage regulator, so that provides your threshold.

 

[neazoi]

I used an fet for the input when I built an rf detector (for 49MHz). The signal was picked up by a normal antenna and tunable LC tank. Then came the fet. This drove a transistor. I connected my VOM to take readings.

To turn off the fet, it needed a negative signal at the bias. By adjusting a 1M potentiometer attached across 2 batteries, I managed to get a zero reading when no signal was present. Then when a signal was broadcast, the fet responded only to the positive half of the waveform. No diode was needed.

With stronger rf signal strength, my meter reading went upscale.

An led would be just as good for an indicator. It has a built-in voltage regulator, so that provides your threshold.

Yes the FET is very sensitive and that is why it is used in these electrometers. However in my application the FET does not need to turn off ever. It will continuously buffer the AC to a diode for rectification as shown in the schematic posted. It is all about loading of the RF circuit and isolation from the detector. I think a single LED will load it too much and it may behave in a non-linear way as the diodes do and I do not want that.
Your point about recification without even using diodes is what I have thought as well, but I do not know if it can be done with the same isolation FET.

- - - Updated - - -

In all above discussion I failed to see the most important point: First you need a calibrated detector, with a response "perfectly" flat over 1...30 MHz. I am afraid such detector is difficult to find or make.
Diode detectors used as RF power meters are on the market but expensive. Their response may vary by more than +/- 1 dB, which translates to +/- 12% of power in mW.
Then with such detector plus a power divider (to relieve detector load on generator output), including temperature variations, only then one can add a level comparator to indicate a power level exceeded a preset limit. Or two comparators to indicate a power level within a window.

I think a very simple solution as required does not exist.

It does not need to be that accurate or linear.
My experiments with the circuit posted, show great results. The LED switches on at 6.2-6.4Vpp and I have done this test at all interested frequencies. Again, experimentation showed that.

 

[neazoi]

Just as I though it, the FET before the circuit is just too sensitive! How can I limit it's sensitivity?

 

[jiripolivka]

Try use an attenuator, or loose coupling to the main line.

I like the technical term "too sensitive"! Have you MEASURED the RF power by a RF power meter?

- - - Updated - - -

Try use an attenuator, or loose coupling to the main line.

I like the technical term "too sensitive"! Have you MEASURED the RF power by a RF power meter?

 

[D.A.(Tony)Stewart]

Your input rectifier capacitance ratio is too high and attenuates the RF!!

100mW*50=V^2, V=2.2V which enough to drive a Red HB LED with a Vf of 2.2V and threshold of ~1.85V , while a Blue or white LED has a Vth of 2.85 and Vf=~3~3.2V use a 10R series in each LED and draw 50mW from your excess power via rectifier to drive both LEDs if too high. Of course this may generate harmonics, so a double emitter follower PNP > NPN can buffer the peak detector and drive the LEDs off a 3.7V LiPo using higher series R's for limiting current. Red dim indicates 80mW, bright =100mW, and very bright with White indicates 200mW.

 

[neazoi]

Your input rectifier capacitance ratio is too high and attenuates the RF!!

100mW*50=V^2, V=2.2V which enough to drive a Red HB LED with a Vf of 2.2V and threshold of ~1.85V , while a Blue or white LED has a Vth of 2.85 and Vf=~3~3.2V use a 10R series in each LED and draw 50mW from your excess power via rectifier to drive both LEDs if too high. Of course this may generate harmonics, so a double emitter follower PNP > NPN can buffer the peak detector and drive the LEDs off a 3.7V LiPo using higher series R's for limiting current. Red dim indicates 80mW, bright =100mW, and very bright with White indicates 200mW.

I made it with a J310 and an input coupling cap of only 3.3pF. You were right the smallest the input cap, the less loading of the transmitter. A point has to be found where enough signal is coupled to the fet but only enough.
Thank you all for your replies