View Full Version : Homebrew RF milli voltmeter
11-02-2011, 08:16 PM
Wonder if there is any recommended schematics or kits, be they digital, analogue or test meter adapters?
I'm aware of the measuring capabilities of 'rf sniffers' and it's diode 'knee' voltage... for some reason I always seem to get better results with germanium types even than a biased 1N5711. I was looking from say, 50mV upwards throwing rf sniffers out of the equation.
Take a look at www.kangaus.com. He has a kit to build the Wes Hayward W7ZOI RF Meter that was introduced in June 2001 QST. The meter has range of -70dbm to +13dbm. I think kit goes for around $60.00. Terry K9TW
11-03-2011, 12:21 AM
There are a couple of ways to do this.
One would be to use your HF radio's S-meter. You would have to know what each division represents. You could then make an attenuator for it that you could step through known levels and translate that to the signal level under test. This method requires good shielding and proper calibration at all frequencies that you want to test.
Another method would be to make a wide band flat response rf amplifier and use a germanium diode for obtaining the signal level. As before you could use a known attenuator to get the step levels you want. Again this method requires that you have a low noise rf amplifier, have a known response versus frequency and good shielding.
All of these methods would, if you want the tests to be proper, have some way to properly calibrate them. A good quality signal generator that covers the spectrum you desire is a good way to check it.
There are other methods and perhaps someone will drop in and share them.
Hope this helps
11-03-2011, 12:48 AM
Thanks for the inputs.
I wanted little more elaborate unit rather than use a S meter, the link is interesting we have a Kanga UK here. Something I've never tried is biasing a Ge diode, guess the problems with 1N5711 diodes is the guard ring. The kits which seem relevant they don't work so they mustn't be ready yet.
Using diode biasing looks fine and dandy on paper, they all look the same there but in practice the characteristic is variable with each diode taking the calibration with it, and temperature no doubt - am I splitting hairs here?
I built a couple W7ZOI hybrid IF amp (the modded version), complete with AGC, S mtr, SBL1 and af pre-amp.... works a treat and very quiet and stable even on single sided pcb.
Thanks for the replies,
11-03-2011, 04:28 PM
I think the trick is to put the amp in front (one of the minicircuits monolithic broadband amps comes to mind) thus taking the small signal characteristics of the diode out of the picture. The early '90's ARRL handbooks have a high accuracy SWR meter project that goes into great detail around how you might do this.
11-03-2011, 06:07 PM
There are several manufacurers that make ICs that will convert an RF input level to a voltage level. These chips usually run on 5v at a few mA of supply current. The output is a voltage that is linear with respect to the db change of the input signal, and are pretty accurate.
Here is a datasheet for one family of these devices. One of the models works from -45dbm to 0 dbm and has a 2K ohm input impedance. Frequency is from 2MHz to 1.6GHz. A simple meter could be constructed by use of this part, a 5V regulator, and a few capacitors. The output could be read by a simple DVM, and that could be converted back to dBm or volts. Since the response is logarithmic, the output would not change like a normal meter. The minimum usable signal detection would be around 5mV or so. If signal levels below that were needed, a simple preamp could be used to get down to the .5mV range, or better.
These devices are available from DigiKey for under $3 in single unit pricing.
There are other devices similar that cover other freq ranges and levels, this is just used to give an example of what is available.
11-03-2011, 08:36 PM
Thanks for the MAX9930 chip link, will give it a 'chew over' shortly, using a monolithic amp crossed my mind as well. I just thought there may have been some proven design or kits out there, anything with 'amateur' accuracy would be fine say from about 50mV.