While visiting family in Espanola, the three of us (Julie, Felix and I) stayed with one of Julie’s cousins (Denis), who as it happens is a fellow ham (VE3OOZ). Back in May, Denis showed me some of his radio gear and lamented that he had a nice quad-band radio which had the magic smoke escape. I asked Denis to save the radio from the scrap bin if he was planning to throw it away. In an attempt to save the radio, Denis sent it away to be assessed and if practical, have it repaired.
Work order for repair of FT-8900R
Well .. the radio came back from the repair shop with a ~$100 invoice saying that they were unable to repair the radio as the final amplifier requires a special soldering tool (which apparently the repair shop does not have access to). Looking at the FET, I am unable to imagine what kind of “special tool” is required.
At the conclusion of our visit in Espanola, Denis offered up the “box full of parts” for me to attempt my first modern radio repair!
The problem: the radio would key-up but would not produce any RF at the antenna port. The likely cause: a blown final amplifier.
Rigol DS2202 200MHz oscilloscope using home-made short ground lead and probe tip for high-frequency measurements.
My first order of business was to prepare my oscilloscope for high-frequency probing at an impedance of 50Ω. I rummaged through my box of connectors which a very generous ham (Ted) gave me a few years ago. My quest was for an in-line 50Ω termination. With the scope termination impedance matched to the radio, I turned my attention to the probe itself.
When working with high frequency waveforms, it is useful to keep the ground lead as short as possible which meant fabricating an on-probe ground leg out of some scrap wire.
200MHz probe with small wire wrapped around probe ground to provide short lead.
Short ground legs are commercially available and have the advantage of being made of spring-steel which is a godsend. My poor rendition of a ground spring used a length of tinned copper wire wrapped multiple times around a golf-tee as a form. I then transferred the coil end onto the scope probe and bent up the long remainder of the wire into a lead form as shown in the image.
Having spring steel would have made life a whole lot better as it turns out. Too bad I am cheap!
Basic test setup with radio and a Bird 300W 50Ω dummy load.
With the oscilloscope ready for action, I prepared the radio which meant powering it up and reducing the RF output power to 5W on all bands.
Recall that VHF/UHF is quite bad for the eyes as they absorb the RF energy at a disproportionately higher rate than the rest of your body!
I also terminated the radio antenna port into a 50Ω dummy load to provide a resistive load for the RF to meet with both safety and regulatory requirements. My initial transmission included my callsign and “testing” just to be above-board with the whole process.
Yaesu FT-8900R guts showing primary area of interest for probing 2m and 70cm amplification stages (area shown in red box).
My initial test at 144.400MHz had the probe at the final FET and did not produce a waveform on the oscilloscope screen – it did however show the DC bias on the FET for normal operation. Time to change the probe inputs to AC coupling! (oops)
I then consulted the service manual for the radio in order to determine the signal path for 2m. The radio operates on four bands and so it would be important to check the output for all four bands as there are bound to be at least a few possible paths. The hope is to find a common fault which is ideally less involved and cheaper to repair.
Factory service manual showing 2m and 70cm signal paths.
As per the documentation:
144 MHz Signal
The adjusted speech signal from Q1054 is passed through the transistor switch Q1501 (BU4066BCFV) to varactor diodes D1076 and D1077 (both HVC365), which frequency modulate the transmitting VCO, made up of the VHF-VCO/B Q1109 (2SC5375) and D1078 (HSC277).
Probing the waveform at the gate of Q1139 (also confusingly shown as Q1039 in schematics).
The modulated transmit signal is passed through buffer amplifiers Q1100 and Q1111 (both 2SC5374) and diode switches D1075 (HSC277) and D1128 (DAN235E) to the pre-drive amplifier Q1139 (2SK2596).
5W (intended) 144.400MHz waveform showing 1.5Vpp (500mV vertical scale) when probing the gate of Q1139 (also shown as Q1039).
The amplified transmit signal from Q1139 is passed through the diode switch D1139/D1140 (both HSC277) and the driver amplifier Q1137 (2SK2975) to diode switch D1133 (RLS135), then finally amplifier by power amplifier Q1134 (RD70HVF1) up to 50 Watts of power output. These three stages of the power amplidier’s gain are controlled by the APC circuit.
The 50-Watt RF signal is passed through a low-pass filter network to the antenna switching relay RL1001 (G5A237P), then passed through a high-pass filter network and another low-pass filter network to the ANT jack.”
The results shown indicate that the modulated audio is making it as far as Q1139 (the pre-drive amplifier) which is taking a 1.5Vpp signal and effectively amplifying it to 28Vpp.
Probing the waveform at the drain of Q1139 (also confusingly shown as Q1039 in schematics).
Checking the drain of Q1139, the 144.400MHz signal was observed to have a peak-to-peak voltage of 28.2V
This means that we can feel reasonably certain that the pre-driver amplifier is fine and is passing the appropriate signal level on to the next amplification stage.
5W (intended) 144.400MHz waveform showing 28.2Vpp (5v vertical scale) when probing the drain of Q1139 (also shown as Q1039).
Between the stages are a series of filter elements and diode switches. The easiest circuit element to check are the inductors as they should have a low in-circuit resistance and in the correct signal path, should pass the RF signal with some ease.
The diodes can be checked with a VOM in-circuit by using the (often) built-in diode check function. Being sure to test the diodes in both possible orientations, my first check would be for complete opens (that is no conductivity in either direction).
Before starting on the individual components, it is worthwhile to check the input side of the next amplification stage (the gate of Q1137) to see if we have a reasonable signal at that pin.
5W (intended) 144.400MHz waveform showing 28.6Vpp (5v vertical scale) when probing the left side of L1141
Upon scoping the gate of Q1137, it was found that there was no appreciable signal at that pin. Time to go back through the signal path and take some measurements at various points.
The first verification was the left side of L1141 to ensure that the same signal leaving Q1139 actually lands at the inductor. Sure enough, we have a good waveform and so to ensure that the
5W (intended) 144.400MHz waveform showing 28.0Vpp (5v vertical scale) when probing the right side of L1141
inductor is passing the signal, the right side was scoped and the output captured.
The voltage on the right side of L1141 was 28.0Vpp and the waveform looked very good. C1727 was the next component in the signal path, so it made sense to continue proving the signal as it progressed through the filtering / switching stage.
The results were interesting to say the least.
5W (intended) 144.400MHz waveform showing 27.2Vpp (5v vertical scale) when probing the bottom side of C1727
At the bottom side of C1727, a nice clean 27.2Vpp signal was observed. On the other side of C1727, there was a significant drop in voltage.The signal was measured to be 9.76Vpp. The voltage dropped across C1727 was 17.44V
5W (intended) 144.400MHz waveform showing 9.76Vpp (2v vertical scale) when probing the top side of C1727
Although my oscilloscope is rated (and calibrated) to 200MHz, it has been tested to be within the 3dB range beyond 350MHz and so I am fairly comfortable in the measured waveform voltage.
Does this decrease in voltage across C1727 concern me? Time to have a look at the value for C1727.
The service manual lists the capacitance to be 33pF with a 50V rating.
The expected current through a 50Ω resistor with a voltage drop of 17.44V is 0.35A (I = 17.44 / 50Ω) and thus, using this current, we can calculate the expected voltage drop across the capacitor and compare it against the measured value. First, we will need to calculate the impedance (capacitive reactance) of the capacitor.
Recall that capcitors improve in their ability to pass AC signals as capacitance and frequency increase, therefore, the reactance decreases. This informs us to use the following equation:
Xc = 1 / (2 π f C) therefore Xc = 1 / (6.28 * 144E6 * 33E-9) = 0.033Ω
According to Ohm’s Law, the expected voltage drop should be: 0.01V (V = 0.35A x 0.033Ω) versus the observed 17.44V. With a voltage drop of 17.44V across an (expected) 0.033Ω the current would be 528A (I = 17.44v / 0.033Ω) which is impossible for this tiny capacitor. Looks like this capacitor is suspect!
… more to follow
