April 2012 Archives

TinyTrak3, Cellphone GPS, and GE PCS

| No TrackBacks
I interfaced a TinyTrak3 to a GE PCS portable VHF radio and a "GPS Battery" for a Nokia 5165.

The GPS battery woke up the GPS off the internal battery and provided GPS sentences out the TinyTrak3 needed. The connector is a 2.5mm stereo plug, one side is TXD, the other is RXD. I never worried with which was which -- I just applied a DC voltage from ground above 1.5V between ground and one of the wires. If it was the receive line, the GPS would wake up, as observed through an LED on the back of the battery next to a button that says "GPS". I experimented with the GPS as much as I could as far as what it would accept. It only provided 1mA of load and "tripped" (turned on) roughly somewhere around +1.3V. The original instructions I had were "give it +12V and it will wake up and spit out NMEA".

To wake the GPS, I installed a 4.7Kohm SMT resistor across J7. This put V+ from the radio connector to the serial connector on Pin 4 (RXD/TXD are pins 2, 3, and ground is pin 5). However, when I connected the TinyTrak to the computer, it transmitted garbage. The RS-232 analyzer showed me nothing useful on it's LEDs, so I was a bit mystified. I finally decided to slap a cable together, assuming the problem was related to excessive capacitance or coupling computer noise into TinyTrak. The cable was a full serial cable at one end and the other was pins 2,3 and 5 -- a safety net. The cable is also wired to be a "null-modem" per the TinyTrak documentation. I believe the actual problem was that the pin on the computer was going to a negative voltage and was dragging the power bus down or otherwise introducing noise. I found out the problem wasn't the cables I'd build when I disconnected the TinyTrak from the computer mid-tone generation and it cleared and sent a normal packet. In the next one, I will be installing a diode to see if it makes a difference.

Annoyingly, each device has an independent battery. On the TinyTrak3, I had to cut J4 to turn the LEDs off, and since they were useful for diagnostics (and I have had a life-long love affair with LEDs) I installed a two-pin header block and a Berg jumper in the case. I wasn't able to find any way to pull power from either the radio or the GPS, since the GPS only had TXD and RXD. The radio only had +4.75VDC on the PTT input, which of course disappears when PTT is brought to ground to transmit. On the receive side, +4.75VDC was present on each of the two speaker leads, but the audio amp is only powered up when the device is receiving. Accordingly, I installed a 9V battery connector on the Radio connector. The TinyTrak draws about 6.6ma in normal operation without the LEDs turned on. A 9V battery can only source about 20mA of current, so this works out. My first attempt at powering the TinyTrak didn't work so well, but it was because the battery only read +6.5VDC before attaching it. The datasheet on the 78L05 regulator used in the TinyTrak requires a minimum of +6.7VDC input to provide properly regulated output. So as long as the battery is new, the TinyTrak is happy.

The GE / Ericsson PCS radio was... interesting. As noted in the earlier Mitrek adventures, the "audio pa enable" line was tapped through an ingenius hack. The rest of the interfacing was pretty straight-forward. J902, the "UDC" (speaker mic connector) provides five signals: GND, PTT, Mic Hi, SPKR+, and SPKR-. However, J902 is also used for programming, and the radio cannot be allowed to turn on with Mic Hi grounded or the radio will go into programming mode -- which does not have a timeout. Another gotcha on the PCS is that if you provide it MORE DC power on the input, it puts out LESS RF power. This radio was designed for 7.5VDC. Don't try to feed it more. Oh, and the power switch is located on the side of the battery. There is no on/off switch on the radio chassis itself.

The GE PCS has more in common with a Motorola Spectra than a Mitrek. The reason why is that the Mitrek has a pair of 500 or 1000 microfarad capacitors between the audio amplifiers and the chassis speaker output. The PCS, and the Spectra both use the same differentially driven audio amplifier -- bridge amp -- without any isolation on either output. The result of this is that on the Spectra, the audio amplifier can be instantly destroyed if one side of the speaker leads touches either the supply voltage (+12V, +A, +B) or ground. Similarly, it's possible to damage the PCS from allowing the same sort of condition to occur.

Using the "audio PA enable", I got lazier than I have been. On the Mitrek, I was able to simply borrow one side of the audio output and route it out of the chassis. I ignored any AC that might have been riding on the signal. In this case, I went a bit brute force. I connected SPKR+ to a 4.7Kohm resistor, and SPKR- to a 4.7Kohm resistor. I then tied the two 4.7Kohm resistors together and soldered them to the Carrier Detect input on the TinyTrak. Of course, like the Mitrek, I had to "Invert CD" for it to work. I also found that -- what I had forgotten -- the squelch level only needs to bet set far enough to trip the TinyTrak's carrier detect input. The squelch is handled by the PCS, and on the PCS it's a digital control.

If there is one thing annoying about the PCS, it's that it has one of the "hottest" receivers in a handheld radio. 12dB SINAD at -124dBm. The TinyTrak won't transmit a packet if the carrier detect is tripped -- random noise or real packet. Since the squelch is managed by the radio, the carrier detect isn't easily adjustable without reprogramming the radio. On the other hand, the squelch action is better than what the Mitrek shipped with.

There isn't much to say about the Mic Hi input on the PCS, other than it's referenced to ground and also used for programming. Connecting that normally worked as expected, and I was able to achieve 5.5KHz deviation just by talking into the mic on the radio. I set the TinyTrak for about 3.7-3.8KHz deviation using "Both" tones.

Power out was somewhat expected. The previous owner identified the radio as putting out about 4.3W, but my testing showed 3.5W, and was repeated with another radio, also making a little north of three watts. On the other hand, these are 150MHz radios operating eight megahertz away from design limits, so I'm not too concerned. The radio is sufficiently complicated to disassemble that I may never try to tweak it.

Finally, I got good decodes of the generated AX.25 packets from the digi (via APRS-IS) with 193 ms of TXDelay. I believe I can set it lower, but I will have to actually get a TNC setup here to confirm.

About this Archive

This page is an archive of entries from April 2012 listed from newest to oldest.

March 2012 is the previous archive.

May 2012 is the next archive.

Find recent content on the main index or look in the archives to find all content.