Fun with a KPC 9612 and a DVR4-10 !

( A little note on some pitfalls experienced )

I help look after a Packet Node and TCP/IP BBS in West London. It lives at the work QTH, and is active on 70,144,433 and 439 MHz, the latter being a 9K6 link making use of a Baycom 4 port card in the PC.

A major bugbear is a one hundred foot run of feeder from the kit to a roof where our aerial farm is situated. In a bid to `cut our losses', we thought of siting a cabinet just below the roof, and installing both 70cm radios and a TNC there, with a wire link to the 386 in the radio room.

433.65 is a 1200 baud user and light mail port, 439.825 is a 9k6 network link. The Kantronics KPC 9612 looked like just the job, equipped as it is with a 1200 and a 9600 port. At the same time, a DVR 4-10 was purchased to replace an aging Storno 5000 series mobile on 9k6 duty.

In making the replacement system work, I came across a few obstacles and a bug. These notes refer to the specific circumstances we found ourselves in, and you may experience better or worse !

The Radio

The DVR 4-10 radio looks the biz, outside and in ! Have the soldering iron ready to make up the TNC lead - plus you may need a connector for the tnc's end of the cable ! Two TNC ports are provided, an analogue for links up to and including 9600, whilst a TTL port caters for 19K2 speed fiends. If you're going to use the radio with a KPC 9612 - wire up the external DCD line, and you want the analogue port !

The Radio came set up on 430.55 MHz and we needed the gear to work on 439.825 MHz. If, like us, you dont' wish to work on the `factory frequency', the Kantronics manual provides an exhaustive spec of the crystals the 4-10 needs.

The first major challenge is opening up the radio. Best if you can find a few Judo mats, a beanbag or two, and a friend. Undo every screw and pull at both ends 'till you're red in the face. The first to hit the floor buys the drinks !

Crystals installed and radio connected to a PMR test set, I followed the alignment instructions. The first major caveat is the adjustment of the phase locked loop voltages on tx and rx. Anytime I got near the quoted value, the loop lost lock. A trusty analogue AVO 73 on the 10 volt range isn't good enough - it's best to have a FET VOM or a scope with a x10 probe for this bit, or at least something in the 500,000 ohms per volt area. It's as well to point out that when the manual tells you to adjust the loop pots slowly - SLOWLY is the word !!!

Once the crystals are on frequency and the loop volts are OK, the battle is won ! A few trips round the rx front end helical filters, and a `follow the nose' on the tx PA and drive trimmer caps saw the radio seemingly happy on 439 megs. When adjusting the tx for maximum smoke, do remember that the PA transistor isn't bolted to the case, and will start to shut down after more than a few seconds.

The KPC 9612

I'd had the KPC 9612 at home for a few days, so I could put it through its paces in the various kiss modes. The use of kiss mode is mandatory for G8BPQ and NOS users. There's a bog standard kiss mode, plus extended kiss with polled and checksum extensions available.

Bog standard kiss is fine for those who run `one tnc, one com port'. Polled and checksum modes are for those who wish to connect more than one TNC, via a matrix, to one com port.

For startup from cold, you'll need a basic terminal program. The KPC 9612 first comes up in a `find the *' mode. i.e. once you see a sensible message coming from the TNC, send it a * . The Tnc then prompts you for your callsign.

KISS mode chaos

One of the nicer features to be found is CWID even when the TNC is in KISS mode. The 9600 port even ID's in keyed audio tone ! On the nasty side, the TNC is in `newbie' mode, and INTFACE TERM is a must for the serious user wanting to do some serious configuration !

`One TNC - one com port' users can safely set up their CWID parameters, DCD mode etc. and then use INTFACE KISS followed by RESET to enter boggo kiss mode and skip the next caveat. This basic kiss mode is fine for using both radio ports in one KPC 9612 together with BPQ or NOS.

Extended kiss users who wish to make use of polled mode in conjunction with checksum mode should note the following:

In polled mode, the BPQ / NOS or WHY individually asks each port in each TNC in turn to send it any data it may have. A TNC should not send to the computer unless it has been asked.

In checksum mode, a checksum byte is appended to each KISS frame. In the unlikely event of frame corruption, the checksum should identify a naff frame which should be discarded. In JNOS at least, POLLED and CHECKSUM modes are enabled together at the JNOS end, so they must both be enabled at the TNC.

Here's the fun !:

In CHECKSUM mode, early releases of version 5.2 software in the KPC 9612 will append a checksum byte to all KISS frames destined for your computer.

This value is always zero.

If you are lucky, BPQ or NOS may accept one in every two hundred and fifty six frames as being OK.

Thankfully, Kantronics say this bug has been fixed in later versions of 5.2, though why they couldn't be bothered to append a revision letter to the software version number (like everyone else who is on THIS planet) is totally beyond me !

It don't work !

Now's the time to configure your NOS / BPQ or WHY to accept the serial link from the TNC and wire up the necessary. Hopefully you're set to a channel where there's already some 9k6 activity and you've got a NOS trace, or BPQ monitor going.

If you're running JNOS, here are the relevant attach lines:

# This is the normal attach line for straight kiss to com 1
# 9600 baud computer <> TNC
# Port called tnc0 is the only one
# attach asy 3f8 4 ax25 tnc0 2048 256 9600

# KPC 9612 : tnc0 is the 1200 port, tnc1 the 9K6 port

# Following two for polled kiss
# attach asy 3f8 4 pkiss tnc0 2048 256 9600
# attach kiss p0 1 tnc1

# Following two for non-polled kiss
attach asy 3f8 4 ax25 tnc0 2048 256 9600
attach kiss tnc0 1 tnc1

Plug a speaker in the back of the DVR, hear the packets hit the Tnc, look at the trace screen, ..... Sweet Fanny Adams !

Open the squelch on the DVR, and you may notice that the DCD lamp goes OUT whilst a frame is being received. External DCD derived from the DVR is a must. Re-set the squelch, and reconfigure the TNC for External DCD if necessary.

The next hurdle awaiting the unwary is the receive equalisation. This is best set up with the meter on a test point as per the instructions, hoping that a long enough frame comes along from a station being monitored. Assuming the fixed equalisation doesn't give any output, take the lid off, move the rx eq jumper to variable, and move the eq pot slider towards the front of the Tnc (this is the pot nearest the front). At endstop we found that we could copy some, but not all frames from two other nodes on the channel. A 680 ohm resistor soldered across the 620 ohm round the back of the pot made the system come good. Doubtless we invalidated the warranty, but what the hell ... it works now !

N.B. Going anywhere near the 9612 with a soldering iron, whether earthed or unearthed bit, causes a reset to `virgin values' - you're back to finding the * again !

After a back-off of the receive eq pot, we found that the position of the receive input level (Hi/Lo) jumper didn't make that much difference

On TX, we had to jumper select LOW level audio O/P, and the level pot slider was set to 9 o'clock when viewed from the side nearest the pot. The TX side just seemed to work without any hassle.

Before I wish you well in your endevours, let me just add one more little experience. It's directed to anyone who might wish to operate a DVR 4-10 on 439.825 in conjunction with another tx on 433.650.


The DVR Rx Local oscillator crystal frequency is 439.825-433.650

Cheers and good luck !

Gareth Rowlands,