AM/ASK Envelope Detector in-output voltage relationship

Test circuit, the sensitivities of this test receiver unit is about -73dBm, a low gain configuration.

The signal generator specifying output power in dBm, in this way the raw data for input is in dBm

X-Axis converted to RMS voltage

Converted to mV for input, the plot is quite linear at first and later saturated after input raised to 1mV , corresponded to output voltage of 2.9V. The saturation is caused by the voltage supply of final opamp stage is limited to 7.4V battery voltage.

The linear range of the IF amp-detector circuit is about 25dB. AGC circuit must control the signal level fall within this range.

Long range challenge 2

 In the previous long range challenge in Plover Cove Reservoir main dam, the transmitter was using two parallel PN2222A as the final stage, that give 500mW output, this time, a new lot of BLT50 transistor was brought from another online store, it's rating is 1.5W, but can it really gives 1.5W? let's try to see the result.

with class B bias (= no bias) and power supply of 10V, 27MHz, the output power could be obtained is 31dBm. (28dBm showing in spectrum analyzer and counting -3 dB additional attenuator), this is a little bit more than 1W, but certainly less than the rated 1.5W as datasheet mentioned. May be the input power is not enough or output matching not good enough.

more than 1W output power, it's quite satisfying.

adjusted the instrument's stop frequency to ~250MHz, it shows quite many higher harmonics, well of cause, it's class B operation and not much output filtering.

Anyway, the transmitter is now upgraded from 500mW to 1W,  it's time for outdoor range test. The selected transmitter location is Ma Liu Shui Ferry Pier, and receiver location in Ma On Shan, they are about 800 meters apart. The previous attempt in this site was failed, because the receiver sensitivities was  just about -60dBm, obviously not enough. In this time the receiver sensitivities is increased to -90dBm. And the antenna of both side is 10cm longer than last attempt.

The bicycle trip

Transmitter is firstly placed near the pier, then I took the receiver and ride along the path shown, when riding along to the target receiving site, some signal (about 3 RSSI LEDs) could still be detected, even on the bridge, when riding closer and closer to the target receive site, the signal become very strong.

Mission accomplished. All 16 RSSI LEDs is on, showing full signal strength. The transmitter is on the opposite of the harbor. Which is more than 850 meters apart.

Sadly, I have to ride back to take back the transmitter.

Lesson learnt

Before deploying a radio, the output power and receiving sensitivities must be measured clearly using equipment.

Aluminium foil shielding box

In recent 2.4GHz measurement, I suspect the WIFI around here is interfering, to get rid of these annoying WIFI or other 2.4GHz source, a shielding box is made, using aluminium foil, it is found that the top cover need to be in tightly contact with the lower box body, otherwise it's not shielded enough for this bluetooth speaker to shut up. 

Bluetooth speaker as a test sample.

Measuring antenna is insert from the top little hole. Speaker and mobile phone inside.

Active RFID tag spectrum shape and burst time casual measurement

The ADVANTEST R3463 spectrum analyzer is just arrived, it frequency range reached 3GHz and it's a synthesizer type of spectrum analyzer so it get rid of many problem for example frequency drift off the view, ofcause a 3GHz analyzer enabled measuring of abundant 2.4GHz radio systems, including WIFI, bluetooth, zigbee and also this one: 2.4GHz Active RFID Tag.

This little 2.4GHz radio tag is powered by CR2302 battery, firing a short burst of signal in 1 sec time interval. May be it could be used to locate a pet cat.

Because there are many WIFI routers nearby as interference and the tag is very low power (~0dBm), the tag have to put closely to the antenna.

The spectrum view of tag signal, using long time peak hold, because it firing a burst every second, which is quite infrequent. The center frequency is measured 2.457GHz and power -13.5dBm.

As we can see the SPAN setting of this screenshot is 10MHz, and every grid equal 1MHz, the -6dB bandwidth of the spectrum is roughly 1MHz (for compliance to FCC rule for non-frequency hoping system in 2.4GHz ISM band, the -6dB bandwidth should be wider than 500KHz).

One feature of this R3463 spectrum analyzer is Transient analysis, which is a power vs time measurement mode, by using this mode, we can obtain the power envelop of the signal burst, the burst time is 180uS, that's translate to 1/5555 duty cycle.

HP 8590D spectrum analyzer CRT repair attempt

Well, the CRT of HP 8590D spectrum analyzer has been malfunctioned for a few months, a while before i was trying to hook up an external monitor to the machine, hope that i can still make use of this machine for a little while more, however i found this way inconvenient because the context button next to the original screen is far away to the external monitor and I also have to turn on the external monitor and video converter whenever i want to use the spectrum analyzer.
As I know the CRT is actually working quite independently to the spectrum analyzer core, i hope that i can open up the analyzer to check if there are some burnt component to replace or some tuning knob i can adjust so to save back the CRT.
May be it's time to start this exploration journey.

This CRT have been looking like this for a few months, too bad.

Amazing looking, what a sophisticated high tech stuff in 90s, 

as a conventional electronics testing instrument design, it's heavy but, not difficult to take off the cover. As I expected the internal of the machine is highly modularized.

CRT, mixer and various of RF modules are shown, many SMB connectors, RF modules are heavy shielded and connected with semi-rigid cables, the CRT shielding could not be open, may be need to tear apart more module before i can take out the CRT.

The A5 second converter module, and a stuff in the middle of a rigid cable looks like common mode filter.

The bottom side contains an unshielded logic board, may be it's the computer, (90s computer)

Marking on the RF modules shows "LOG AMPL", "BW FILTER", "AMPLITUDE CONTROL" and "THIRD CONVERTER", exciting!

Carefully removed the front cover and power supply module, so the CRT could be extract out.

The extracted cute looking CRT, shielding was removed. I can see some knob showing "Vertical Line", "Horizontal line" and so, may be I could try to adjust these knob.

There we go! after tuning the "Vertical line" knob, the screen is backed to normal. I could also tuning the "Vertical size" knob to make the graphics fully using the whole screen size.

Reassemble the stuff back to the position.

It looks better than before, say goodbye to the external monitor.

Breakthrough of the year 2014 - Long range challenge

Breakthrough of the year! Plover Cove Reservoir main dam long range challenge completed!

It has been a long time goal for me to making a wireless range that could cover the main dam, using 27MHz band and just a monopole antenna. Previous attempts was not successful because the power of transmitter was not enough and sensitivity of receiver is also insufficient. 

It's not easy to get to a good test environment in urban area because the buildings and terrain blocked the signal. The only testing site having straight open channel is the Plover Cove Reservoir main dam.

The simplest form of communication transmitter is to merely feed a carrier to the antenna, no modulation, the key point is big power. Previous attempts was an crystal oscillator stage and one stage of power amplification. In this attempts, one more stage of power amplification is added, coupled by 4:1 step down transformer.

The transmitter with 650mm monopole antenna on ice-cream stick.

27.76 dBm output power, which is more than 500mW

(Note that CRT of this analyzer gone crazy just after boot up for a while)

The transmitter was firstly placed on the beginning of the dam and receiver walks towards the end of the dam. 

The receiver showing the carrier is still detected, at the end of the dam. The sensitivity of this receiver is measured as -90dBm. I guess if need better sensitivity, a double conversion architecture is required.

The receiver signal strength indicator LEDs gradually show lower and lower signal strength until the end, 3 of the total 16 LED still remain shining, what a relief.

Frequency response of 3 types of 455KHz ceramic filter

There are many types of 455KHz ceramics filters out there, what is the different between them? the 455KHz usually comes with suffix A/B/C/D/E, indicating the bandwidth of the filter. what is the filter response actually looks like? In this test we picked 3 ceramics filter, They are -A -D and -E.

The A type filter having an impedance of 600 ohm for both in-out, D and E type having an impedance of 1500 ohm. The filters are loaded with resistor and output RMS voltage are measured using oscilloscope.

The responses are normalized and the result turns out like this. So now I get some concept of what the shape of their frequency response looks like.