Kenwood TM-D710GA Low Deviation_Troubleshooting
I went out to Quartzfest in late January to support the ANSR balloon launch, and spend some time camping with friends of ours.
While there, I was presented with the opportunity to purchase a Kenwood TM-D710GA APRS 2-meter/70-cm dual band transceiver for a price considerably lower than market value.
The ham selling the radio stated that it had low audio, but otherwise worked well. I had been looking for one of these on the used market for some time, as it is widely considered to be the cream of the crop in APRS transceivers.
I picked up the radio at a convenient meeting place the next week. I learned then that the actual reported problem was low transmit deviation, no low receive audio. No big deal, just a different troubleshooting path.
When I got the radio home, I powered it up and confirmed the problem as low deviation. It didn't seem to matter if I was using the microphone input or the audio in through the accessory jack (intended for echolink, according to the service manual).
A visual inspection revealed that the microphone cord was worn and possibly not making good contact. I substituted another cord that I had, but it did no affect the deviation level at all. So that was not the problem.
I then made up a test microphone cord from an old ethernet cable (had the RJ-45 installed on one and, and loose wires on the other.
I injected a 5mv 700 hz tone into the microphone input wire and started tracking down the problem.
There are two boards in the TM-D710 main body. Unfortunately, all of the points that I needed access to were on the bottom board, and inaccessible without removing the top board. The ribbon cable connecting the two boards for operation is not long enough to simply fold the top board out of the way. I had to remove both boards from the chassis in order to troubleshoot this problem.
I started by taking off the top cover.
Than I had to remove the speaker and the speaker mounting bracket. There are a few screws around the outside of the speaker mounting bracket, and a hidden screw that is revealed under the speaker to remove the bracket.
Then two plugs are disconnected to remove the speaker. They can be plugged back in when the boards are laying open on the bench.
To remove the top board, there are several screws, mostly around the outside of the board to remove.
The SO-239 socket needs to be unsoldered from the board.
There are four screws holding the final RF power blocks to the chassis. These must be removed as well. The RF power blocks will lift out with the top board.
The ribbon cable connecting the top to the bottom board must be disconnected by lifting up on the black tabs that lock it in place.
There is a 3-terminal 8-volt regulator next to the fan cable that must be unscrewed from the chassis. The fan cable must also be disconnected.
With all of this done, the top board may be carefully lifted out of the chassis, and flipped over on the bench.
For the bottom board, there are two active devices (a 10-volt 3-terminal regulator IC and an audio amplifier IC) that need to be unscrewed from the chassis. Then all of the screws holding the board down can be removed, and the board lifted out of the chassis.
With both of the boards out of the chassis, I was able to connect the two boards together with the short ribbon cable that I had unplugged earlier. Adding the control head and test microphone cable allowed me to power up and begin troubleshooting.
The first thing I noticed was that the 3-terminal device on the bottom board was getting pretty hot when powered on, even without transmitting. I put a screw through it with several nuts and a large wing nut to act as a heat sink for the device. Then I was careful to only power up the rig while I was actively taking a measurement, and power down as soon as possible to prevent any damage.
I used the control head to set the radio up to transmit on 2-meters at the lowest power level. I soldered a wire onto the RF out connection and connected it to a dummy load.
With my signal generator set to send a 5mv 700 hz tone into the microphone input line, I began tracing the signal through the path of the block diagram.
I would carefully place the tip of my scope probe on the point of interest in the transmit audio path. Then I would reach over and use an alligator clip to ground the ptt pin of the microphone cord. Then glance back up to check the waveform on the scope. Then I would remove the ptt signal as quickly as possible, as the RF Final blocks were not heat sinked during this troubleshooting.
All the while, I would listen on another radio to the transmit audio and note the audio level.
I found the 5mv from the signal generator at the microphone input.
I found the output at IC802-1 to be several hundred millivolts rms, as indicated on the scope. So the first stage was amplifying and appeared to be good.
The output of the next stage (IC804 pin 11) dropped down to a lower level, as expected from the diagram, bit was still in the range of 50mv or so. I didn't transmit long enough to get a more precise read on most of these measurements, as I was being careful to limit my transmit to very short periods.
Also, it was very difficult to keep the scope probe steady on the tiny leads of these components, without slipping off and potentially causing an inadvertent problem from shorting the wrong things together.
So anyway... The output of the ALC Amp at IC802 pin 7 was back into the several hundred millivolt range, so that stage was good, as well.
The output of IC806 pin 19 was also several hundred millivolts. Again, good.
The output of the preamp (IC803 pin 1) was around the 1 volt rms range, as expected. So that stage was also good.
I followed the signal to the base and emmiter of Q805 in the SPL filter circuitry. Both input and output was around the 1 volt rms expected level.
The next reading is where things went bad.
The base of Q808 has a value around 1 volt, but the emmiter had a value under 50 mv rms. It should have been at about 950 mv.
I suspected a problem with Q809, the mic mute fet. Why? well it was a fet...
I got lazy at this point, and maybe a little worn out from carefully holding the scope probe. I measured the resistance from source to drain of Q809, and it was not shorted, but had somewhat less than the 1.8M ohm that I was expecting there. However, I don't have a lot of experience with analyzing a fet switch, and was confused as to how it works with no drain voltage applied in this circuit. I go lazy and decided to try replacing that fet first.
I found that the 2SK1830 FET was no longer available through any reliable supplier. I did some research and settled on a SSM3K37FS from Mouser. It can handle a little heavier current, and is a little faster than the original, but looked like it would work. $1.53 for 10 of them, and $8.00 for shipping!
While I was awaiting the parts, I was trying to come up with a better way to test if the mute switch was working. I could see the gate go from 5 volts to near zero volts when I went to transmit, but the signal level on the drain remained the same.
I had some 12 volt white LEDs that had leads on them, as well as current limiting resistors. I hooked the positive side of one of these to my 13.8 volt input of the radio, and careful touched the drain of Q809. The LED glowed brightly on receive and dim on transmit. But it did not go all the way out. So it looked like it might be OK, but still wasn't sure.
After the new fets arrived, I masked off the Q809 area with Kapton tape to prevent my hot air gun from affecting any nearby components. Then I applied a little flux to the leads of Q809 and used the hot air at 350 degrees Celsius to slowly heat up the area and move closer to the fet until I saw the solder melt. I was able to lift it off with my tweezers at that point.
I then used solder braid and little more flux to clean the pads for the new fet. After cleaning with alcohol, I placed a tiny dab of low temperature solder paste to the pads, and picked up a new fet with the moistened end of a toothpick, using surface tension to stick it to the toothpick.
When I placed the fet onto the solder pasted pads, it did not stick, as I expected, but rather remained stuck to the toothpick. It was just too tiny.
I have done some smd soldering before, but the component sizes that I had worked with were in the 1206 or 0804 sizes, considerably bigger than what I was working with here. So this is a new experience for me.
I tried again by picking up the new fet with the moistened end of a piece of 24 gauge solid wire, instead of a toothpick. With that, I was able to get the fet placed and pushed it around to get it properly centered on the pads.
Then I applied heat with the heat gun again, this time at a lower temperature, since I was using low temperature paste (I think I used 250 degrees C). The solder flowed but not as well as I had hoped. I touched up the leads with my soldering iron and the smallest tip I had. The results were barely acceptable. But upon looking at zoomed in pictures of other components in the radio, I decided that I didn't do that bad of a job. I cleaned up the area with alcohol, and was ready to test.
There was no difference in the deviation level, and a quick check with the scope confirmed that it was still broken in the same place.
So next, I thought... Well, maybe that 1000pf cap from the drain to ground on Q809 drain is leaky. So I removed that cap from the circuit, figuring that if it was bad, the short would be removed, and I would see larger signals in the area. Nope... That wasn't it either.
Finally, I decided to go back to real troubleshooting.
I backed up to the prior stage and started measuring DC voltages. I found that the collector voltage was not present on Q808. Further reading confirmed that 8 volts was present on the high side of the collector resistor R852, but not on the other side.
I measured R852 and found it to be in the 450k-ohm range. It should have been 33 ohms. There's your problem!
I didn't have any 0402 size resistors to replace it with, so I bridged a 1/4 watt through hole resistor around it on a couple of wire legs. Not very pretty, but I was able to confirm that it fixed the low deviation problem. Signal levels returned to normal in the circuit, and I could clearly hear a much louder audio in my monitor receiver.
So Off to Mouser I went with a second order. I bought a hundred 33 ohm 0402 chip resistors and a hundred 1000pf 0402 chip capacitors on this order. $1.50 for parts, $8.00 for shipping! Lesson learned...
When the parts came back in, I was able to replace both. Surprisingly, the 0402 chip capacitor was the one that was the most difficult to solder in. Too much solder paste, too light and blew away with my hot air gun, etc. I finally held it in place with a toothpick and dabbed a little solder on it with the tip of my soldering iron. That finally worked, although again, no very pretty in my eyes.
After getting the parts reinstalled, I fired up the rig and noted full audio from the monitor receiver when I spoke into the microphone.
I put the boards back into the chassis, after applying some fresh heat-sink paste to the components against the chassis. I buttoned everything up and gave it a final test. All is working well now.
Comments
Post a Comment