Let There Be Electrons!

Spoiler alert: There aren’t a lot of pictures in this installment. There just isn’t a lot to show other than a dodgy test cable and a chicken-scratch checklist of RC-96 signals to test.

Now that the soldering operations are finished, the sockets and fuses needed to be populated and then ENERGIZED. The design has already been proven, with all of the (likely) hundreds of RC-96s that have entered service in the past. Thus, I decided that only an ohmmeter test of the power rails was needed. That checked good so I connected a speaker and applied 13.8V to the coax jack.

I was rewarded with the RC-96 reset banner message. Then, I was rewarded again… with the same message. And then a third time… perhaps you see where this is going. No BUENO! I used my PhakePAL interposer for the build, and I had always had a notion in the back of my mind that eliminating the U34 chip select signal was going to get me into trouble, even though nothing was installed in that socket (the connection was there, but only the DAC address space was being decoded). I was convinced enough that there might be trouble that I went ahead and ordered a new PCB of a redesigned interposer that addressed the board offset issue and also added logic to unify the U34 and DAC address selects to match the original PAL logic. I just had to assemble one and install it. Reviewing the schematic confirmed this, BTW, as the DAC/U34 chip select is clearly connected to the DS1232 watchdog reset chip. Duh.

This is a gander at the new PhakePAL:

After installing the new PhakePAL, things went swimmingly. I did a cursory test focusing on DTMF decode (chassis keypad and RX input), COS, and PTT function and the front panel LEDs associated with those functions. At that point, I turned my attention to the Tone Panel. Here, the results were not great. No encode or decode. At this point, I was not sure I understood the configuration commands, so it wasn’t clear if there was a hardware problem or if it was operator error. Checking the divide-by-4, 1 MHz, oscillator output on the tone panel revealed that it was at least hardware related since I was only seeing about 8 KHz. I used the 22pF loading caps that were used on the original, but the crystal was a mystery. I used what I could get at Mouser. I adjusted the loading caps and was able to get the oscillator stable at 1 MHz.

I still wasn’t sure I had the configuration figured out. I was able to encode but decode was still not behaving. After some probing and head scratching, I figured out the error, which traced back to my schematic re-draw. I had done a copy paste of the tone channel chips (I have 5 of them) and forgot to go back and delete the D6 connection (RX/TX) from all but CH 1. This meant that all of the chips were set as encoders (since CH 1 was set to be an encoder). So, no decode. Fortunately, a single trace cut under the CH 2 chip fixed the problem. After that, the decoders and encoder worked! This also meant that I now had some notion of how to configure the tone panel, assuming I could remember that at a later time.

Testing… TESTING!!!

This was the test plan:

Not much more than a list of connector pins/signal names and the adjustment pots. If I ever have to do this again, I’ll make a formal test plan, but this is good enough for this effort.

This is my dodgy test cable:

The PCB with the Kenwood microphone attached is my “speaker/mic” RX/TX simulator that I use for all of my controller testing. Saves me from basically having to have a repeater sitting on the shelf to do non-RF testing. The breadboard captured the miscellaneous connections such as S-meter (note the pot for this purpose) and CTCSS I/O. Since J7, J6, and J5 have similar pinouts with respect to the receiver/transmitter pins, this one cable can serve to test all of the signals on those three connectors. For the remaining signals, I decided to clip onto the ferrite beads on the PCB. This skips a few solder connections, so it isn’t optimal, but I did ohmmeter checks to validate the connections between the DIN connectors and the ferrite beads to cover that.

The only thing un-checked on the “test plan” is the phone balance pot. I couldn’t come up with a simple procedure for this, so I rigorously inspected the R108 circuit for soldering.

For the DVR, the serial data signal was the most important to verify. Here is screen capture of the logic analyzer DVR serial data stream:

This shows a logic analyzer capture of the 1200 baud data sent to the DVR. There is a lot of other “traffic” on that signal, but it is at a higher rate and apparently ACC determined that none of it would be mistaken for a pair of CR/LF characters which start every DVR control message. The “0x50 0x15 0x0d 0x0a” (hex for “P21″+CR/LF) message that follows the dual CR/LF “attention” message is the contents of the RC-96 message that was programmed into BB1 to trigger the DVR request. This allowed the DVR control to be tested without having a DVR.

With no fanfare or ticker-tape parade, I was able to check-off everything on my test-list and declare victory (yeah).

The Last Tidbits

With a working system, all that was left was to address some of the last cosmetic issues. Chiefly, this focused on the back-panel labels but also on S/N labels for the PCBs. I have had my eye on a Brother P-Touch label printer for some time now. I have one of the early models which prints text labels, but I wanted something that could do graphics and also save labels for future reprint. The PT-P900 could do these things but was a bit pricey. This opportunity was a chance to at least provide an above average justification for the cost. It came with its own set of issues, but that is way out of scope for this venue. Suffice to say, I now have a new label printer that can do a lot more than what I was able to do previously.

One of the nifty tricks I could now pull-off was to include QR codes in my labels. So, I added QR codes for this blog, and for the github repo I started for my ACC support data. Actually, I think I might have gone a bit overboard on the QR codes.

The Fond Farewell…

Thus, the journey has come to an end. An unpopulated RC-96 PCB arrived in my lap one day, and now it is a real assembled thing, ready for the world: S/N LST96-001. With just a few embellishments it is, in great measure, a factory-stock RC-96 produced over 2 decades after the last one left the ACC factory floor. A few final photos are posted below. It has been an interesting trek down some dusty old paths that I once tread often, and some that were completely new to me.

I want to thank Tracy (N5LUY), Jeff (WN3N), Ed (WA6AXX), Cort (N0MJS), and all those who offered help. I hope you enjoyed this little trip down memory lane.

It is possible that there may be a post-script post at some point in the future, but I have no specific ideas for such an endeavor at the moment.

Cheers & 73,

Joe
KE0FF

Image Gallery:

OK, maybe there are a few images this time after all…

Some random notes regarding the speaker and power jacks from early in the process:

The AND gate in the bottom left is the rev B add-on to unify the DAC and U34 chip select spaces into one connection to the “PAL” pin.

The Assembled RC-96 PCB, top and bottom:

Front and back views of the chassis (see if you can count how many QR codes there are… there also is one on the back of each of the PCBs):

End Of Line…