Packet Radio and the UHF Micor
Time has elapsed since I wrote the original of this. New TNCs have appeared on the market. Much has been learned from actual operating experience. I've found that different versions of the PC boards and schematics exist for these units. If I said, "Make sure JU909 is in place." You'd look and discover that your board didn't match the layout and the schematic in your book and you didn't seem to have a JU909! It's time to re-write the piece and incorporate the new while trying to speak more generically yet be more specific. Talk about oxymorons!
One thing hasn't changed - The following comments are my opinions and therefor can be considered dangerous. You are to feel free to disagree.
Verne - 24 October, 1995 W9ZGS @ KB9KKN.#NEIL.IL.USA.NOAM
EMAIL jgalt@iwic.net
If you don't have a fistful of bucks to buy one of the new 9600 baud rigs that are finally appearing on the market you are forced into making something else work. Synthesized rigs are next to impossible to get to "talk" at 9600 baud - including some of the new ones that claim to be built for 9600 baud operation! This leaves crystal controlled rigs. The rig must be frequency modulated to transmit 9600. Phase modulation wont "do" 9600 baud. There aren't many economical solid state choices available. The Motorola UHF Micor is a crystal controlled, frequency modulated rig and it still appears at hamfests although finding the correct version for packet use can be a hassle. The Micor has some excellent receiver characteristics and these can be improved. The transmitter can have some big time problems! The UHF Micors used a ceramic substrate as PC material in the power amplifier stages. This ceramic has wonderful electrical characteristics but terrible thermal and mechanical characteristics! Some units at hamfests have cracked ceramics in them - these are almost worthless! When the substrate cracks, the output of that stage nose-dives. Usually the ceramic strip that is toward the power connections side of the black plastic tray holding the microstrips is the one that breaks. It has soldered connections from it that don't allow the strip to expand and contract. There is also a "lip" of the tray that snaps over the edge of the ceramic to hold it that further enhances the problem.
It is possible to remove the tray lip and replace the connections with a short, wide, thin piece of copper flashing with a slight bend in it. This offers some measure of relief from the stress problem but it is far from foolproof and is a pain to implement. Look your unit over carefully! It can be very difficult to detect the cracks without carefully sliding the edge of a knife across the board while watching with a magnifier. Make sure you get the manual that went with it - you're almost dead in the water without a manual. If you want to listen to channel activity and switch channels, get one with the control head, cable, speaker and channel switch. This piece assumes that you do not plan to monitor with a speaker, or change frequencies from a control head.
These units were made with a variety of frequencies, power output levels and options. You want a unit that covers the 450 to 470MHz range. Do not assume that any UHF Micor in this range is suitable for packet! A particular unit for the 450 to 470 range may be for simplex or duplex. If duplex, it transmits 5MHz up from where it receives. Transmit and receive frequencies for a "channel" are both determined by a single channel element. A channel element is required for each channel. The frequency of the channel element is determined by the receive frequency. The channel element crystal frequency is determined by the formula: Channel element frequency =(Carrier frequency - 11.7MHz)/24. The exciter board in the UHF Micor contains an off set oscillator frequency that is combined with the channel element frequency to produce the transmit frequency. If the unit was produced to work duplex then the off set oscillator operates at 16.7MHz. If the unit was for simplex the off set oscillator operates at 11.7MHz. Some exciter boards contain both off set oscillators. Units were also produced for below 420MHz and from 470 to 512MHz. These may use different off set oscillator frequencies depending upon FCC requirements for commercial users in those segments. Some may even transmit/receive "upside down." Different uses required different off set oscillators. Channel elements were also produced that allowed AFC of the receiver to be used. Different channel elements have a different number. Channel elements for the VHF Micor, and some other Motorola 2 way radios, look identical with the UHF Micor channel elements but have different numbers and are markedly different. The 2 different channel elements that I have seen for the UHF Micor were the KXN1024A (for AFC) and the KXN1029A (no AFC). If you are going to use the Micor on packet, you want to work simplex. To work simplex the UHF Micor exciter must have an off set oscillator crystal of 11.7MHz in it. If it does, and it came with one of the previously mentioned channel element numbers, you should be OK.
Units were made with from 1 to 12 channels. Single channel units do not have a channel switch in the control box. They have a jumper(s) for proper operation on the diode matrix board. The diode matrix board is at the back end of the control board. Four channels can be remotely controlled from this diode matrix. Units with more than four channels have either an additional, separate diode matrix board or a universal switching board. The channel elements are at one end of the receiver board. All units that I have seen would accommodate 12 channel elements. You can recrystal the channel elements yourself. Before you order a crystal from a vendor open the channel element and examine the crystal. To open a channel element, carefully slip a thin knife blade under the 2 "dimples" (one on each narrow side of the can toward the bottom), straighten the dimple out and carefully slide out the can contents. You'll see either a large crystal with it's pins soldered to split posts or a small crystal with wire leads. You must state the receive frequency, the crystal size/type, and the channel element number to the vendor when ordering replacement crystals for an element.
Frequency modulation is applied to the off set oscillator and not to the channel element oscillator. This method presents the possibility of replacing the channel element with a synthesizer. Crystal frequencies to cover the 430 to 450 range would run about 17.43MHz to 18.26MHz. When you figure out the synthesizer please share your know how with us!
Remove any PL options that might be in a unit. Replace the jumper on the exciter board that was removed for the installation of PL (JU405 on some schematics). If you plan to use a speaker, replace the jumper on the receiver audio/squelch board that was removed for installation of PL (JU201 on some schematics). If you're not going to listen, remove the audio power amplifier board. Try to obtain the power connector with the heavy gauge power leads on it. Attempt to obtain a unit that has the Motorola Preamp in the receiver - Without the preamp the receiver has approximately .5uV sensitivity for 20dBQ (.35Uv SINAD). With the preamp the stats become approximately .25uV for 20dBQ (.175Uv SINAD). I had previously recommended using the Hamtronics dual GaAs FET pre-amp kit with the Micor. I have recently discovered that Hamtronics has seen fit to remove the dual gate GaAs FET from this kit and substitute a dual gate MOSFET and leave the kit price where it was with the GaAs FET. The performance is not the same. Build one of the GaAs FET preamps from the ARRL Handbook and use it. The FET mixer in the Micor receiver and the selectivity of the helical resonators provide excellent characteristics. Intermod is down about -85dB and spurious/image rejection is about 100dB. This will be reduced somewhat if the I.F. is modified for 9600 baud operation.
Interfacing your TNC to the Micor can be accomplished in several ways. It depends upon whether you are going to operate at 1200 baud, 9600 baud or two modems in parallel. It also depends upon your TNC type and whether operators in your area use standard 6dB per octave pre and de-emphasis. There is an article on pre & de-emphasis on some packet BBSs. It gives some practical examples of part values to use. It is named EMPHASIS.TXT. In my area, "emphasis" is used at 1200 baud but a flat frequency response (no pre or de-emphasis) is used at 9600 baud. Let's take them one at a time.
1200 baud only. Receive: Look at the schematic of the receiver audio/squelch board. Locate pin #2 of IC201 (the output of the emitter follower). Note that the output goes through a coupling cap (C203, .22 Mfd?). Change this cap to 2.2 Mfd and connect the plus side to pin #2. Connect the other end of the 2.2 cap to a 470 ohm resistor. Connect a .15 Mfd cap from the OTHER end of the resistor to ground. The junction of the 470 ohm resistor and the .15 cap is the place to take receive audio from to feed TO your TNC.
(Note: Some TNCs may have a built in bypass capacitor to ground on the audio input TO the TNC - examine. The KPC-9612, for instance, has an .01 cap to ground in the 1200 baud input line. I suggest that you experiment with values if you think your copy isn't optimum.)
1200 baud TNCs that require "squelched" receiver audio. I have one of these rascals, an early Tiny-2. I wasn't aware of its "squelched audio" requirement when I ordered it. This worked for me. Hook things up per the above paragraph on 1200 baud receive. Instead of bypassing the TNC end of the 470 ohm resistor with a .15, I ended up using a .2 Mfd after much experimentation. I connected a 5K ohm trimpot from the 470 ohm resistor, .2 Mfd capacitor connection to ground and the wiper of the pot was connected to the TNC. I was able to find a level by carefully adjusting the trim pot where the Tiny-2 worked very nicely. Experiment with slightly different cap values if you have a problem with yours.
1200 baud only. Transmit: Locate pin #4 of IC301 on the exciter schematic. Note that a resistor and a cap are connected in parallel (R304 and C303, 82K and .0047 perhaps?) with one end connected to pin #4 of IC301 and the other end to a coupling cap (C301, .068 Mfd?). The other end of the coupling cap goes to a resistor to ground (R302, 3.3K?) and one more coupling cap (C300, .12?). Replace the resistor to ground with a 68K resistor and remove the "C300" coupling cap. The junction of the "new" 68K resistor and the .068 (?) coupling cap is the place to feed transmit audio FROM the TNC. You're not finished. Note that IC301 has pin #1 connected to the KEYED +9.6 volt line. This is not good - it takes awhile for the caps to charge up when keyed - your packet is half gone before the stage comes alive! Cut pin #1 of IC301 loose from the keyed +9.6 volt line and connect it to P401-1, continuous +9.6 volts. Bypass this connection to ground with a .1 Mfd cap. IC301 is a clipper stage - do not drive it beyond very light clipping! Use the IDC control to set your deviation to 3 to 3.5KHz.
Making it 9600 baud capable will require a little more work. The I.F. amplifier bandpass is too narrow. Remove the 2nd and 4th crystal filters (Y102 & Y104) and replace them with .001 ceramic caps.
9600 baud only. Receive. The best way is to take receive audio directly from the discriminator to the input of the modem. This is identified as P904-15 on my schematic. THIS MAY NOT BE ADVISABLE WITH ALL 9600 BAUD TNCs. If the input impedance of the TNC is low, the discriminator will be loaded so badly that any AFC of the receive crystal is severely restricted. Unfortunately I don't have personal knowledge of any 9600 baud TNCs except the KPC-9612. It's input impedance is 100K and it works great with the direct connection. You'll have to try it to find out. If you experience unacceptable loading you may want to use the method I used - under "Both modems in parallel."
9600 baud only. Transmit: Transmitting 9600 baud audio requires a frequency response from approximately 3Hz (DC is best!) to approximately 8KHz and no phase shift. Locate the base of Q301 on the exciter schematic. Follow it to the left on the schematic until you come to a coupling capacitor (C310, .1?). Remove this coupling cap. Locate P401-11 ("PL" tone) on the schematic. There probably is a resistor (R314, 82K?) connected to it with the other end of the resistor eventually getting to the base of Q301 by way of one or more "JUs" (jumpers). Verify that those jumpers are in place (JU305, JU308?). Replace R314 with a 68uFd, 15 V. tantalum capacitor. The plus side of the capacitor is connected to the base of Q301. Connect a 68K resistor from P401-11 to ground. P401-11 is one of the connectors for the PL unit. P401 is split into 2 segments - P401-11 is the middle one in the group of 3. Next to it is P401-12 and it is at ground potential - perfect for the shield of the lead bringing audio from the TNC. Feed audio from the TNC into P401-11. Also on the exciter board, locate the resistor from emitter of Q301 to ground (R312, 680 ohms?). There is also a cap connected to the emitter of Q301 (C313, .068 Mfd?). Remove this cap and bypass the resistor to ground with a 100 Mfd cap - plus side of cap to emitter of Q301. This defeats the IDC action and raises the stage gain. Exciter boards with 2 offset oscilators will have 2 IDC pots. The "Wide space" IDC pot sets deviation level for simplex. Locate the IDC pot that is used in your exciter. Connected to the "wiper" of the IDC pot is a 15MFd capacitor (C401?). Change this capacitor to the largest value that will mechanically fit in here - at least 100MFd. It can carry a very low voltage rating but make sure you observe the correct polarity. Set the transmit audio level from the TNC short of clipping in Q301. Set 9600 baud deviation to 3KHz with the IDC control.
Both modems in parallel: The KPC-9612 has the ability to run both the 1200 baud and the 9600 baud radio ports in parallel on one transceiver. I paralleled the 2 ports and gave it a try. It worked, sort of, but it was obviously "sick!" I took a good look and found that the input impedance of the 1200 baud port was 600 ohms. This placed a tremendous load on the discriminator. I considered using 2 separate connectors on the back of the Micor; One for 1200 baud and the other for 9600 baud. However I wanted to be able to use the Micor with a "variety" of TNCs - low and high impedance. I had to bring the high impedance of the discriminator down to a stiff, low impedance. I tried several methods with varying degrees of success. I finally used a TL081 JFET Op-Amp. It is not possible for me to draw the schematic here so I'll describe the circuit. I wired it as an inverting amplifier with a 1 meg feedback resistor and a 100,000 ohm input resistor connected to the inverting (-) input. The input cap is a 15 Mfd tantalum with the plus side connected to the .1 meg resistor and the other side of the cap is connected to the wiper of a 100,000 ohm trimpot - high side of the trimpot connected directly to the discriminator and the low side to ground. The output side of the op-amp is connected to the plus side of a 100 Mfd cap. The other side of this cap has a 82,000 ohm resistor connected from it to ground. (The 82K resistor is needed when this stage is used with a TNC that already has a DC blocking capacitor on its input - The resistor provides a discharge path for the two capacitors that will end up in series.) The junction of the 100 Mfd cap and the 82K resistor feeds receive audio to the 5 pin DIN jack on the back of the Micor. I connected the +V terminal of the op-amp to plus 13.8 volts and the -V terminal to ground. I then placed two 100,000 ohm resistors in series - one end to +13.8 volts and the other end to ground. The junction of the two resistors was bypassed to ground with a 47 Mfd cap. A 100,000 ohm resistor was connected to the junction of the other two resistors and the 47 Mfd bypass cap. The other end of this 100,000 ohm resistor was then connected to the non-inverting (+) input of the op-amp. The stage has approximately 20dB of gain. It was built on a narrow scrap of "perf" board and mounted with a 4-40 screw in one end of it to the receiver PL mounting bracket that is adjacent to the receiver audio board. I suggest you initially set the input 100,000 ohm trimpot to about 1/4 open. You now have a Micor radio with a "generic" receiver audio output - no de-emphasis and with the capability of high level output. I took advantage of this high level output. I wanted as much isolation as possible between the 1200 baud and 9600 baud receive audio at their respective ports. Resistors were used in series with the actual receive audio to each. About the highest resistance that can be used in series with the 1200 baud port is 6800 ohms due to the input capacity of the port. A 6800 ohm resistor plus the input capacity give about a 6dB per octave de-emphasis. Considerable audio voltage is dropped across the resistor as it is in series with the 600 ohm impedance of the 1200 baud port - this is an approximate 11 to 1 voltage reduction. The input capacity of the 9600 baud port also restricts the resistance size used on the input of that port. I wanted to use a higher resistance then I ended up using but I experienced too much high frequency roll off - a "no no" at 9600 baud. I ended up using a 27,000 ohm resistor. I connected a 8,200 ohm resistor from the "port" end of the 27K resistor to ground to reduce the high audio voltage to that port These resistors were placed within the DB-9 and DB-15 connectors that mate with the two ports. Setting the level of the trimpot is easy. Type the command CD INTERNAL/INTERNAL to your KPC-9612. Adjust the trimpot on the op-amp so that the RCV LED on the KPC-9612 lights. Turn it just a "tad" higher than the point where it just lights. This is the correct point. Type the command SOFTWARE/INTERNAL when finished. The transmit audio was handled as in the 9600 baud transmit paragraph above - with one addition: A 470 ohm resistor and .15 Mfd cap were placed in parallel and this combination was placed in series with the audio FROM the 1200 baud port. Further experimentation is ahead to determine if better isolation of transmit audio between the two ports would be helpful. I recommend that the internal jumpers in the KPC-9612 be set in this manner: J1 no equalization (until you determine otherwise by the calibration routine). J5 high range. J6 DC coupling. J7 high range. J9 high level. Set the command ONERADIO to ON and type the command CD SOFTWARE. Follow the calibrate routines for the two ports per the KPC-9612 manual. Be prepared to readjust the level control on the op-amp stage. My K-net node now operates at 1200 baud AND 9600 baud. The 9612 also provides me with an intelligent mailbox and allows me to provide a weather information server from what would normally be considered the "keyboard" side.
A 9600 baud note. If you are using a TNC other then the KPC-9612 for 9600 baud operation, note that many other brands of TNCs have an input impedance around 600 ohms whereas the KPC-9612 input impedance on the 9600 baud port is 100,000 ohms. If you strap both ports of the KPC-9612 in parallel, and don't use any "isolation" resistance in series with each port, you will be looking at an input impedance of 600 ohms. Some brands of TNCs ALWAYS have both ports in parallel. Some of the coupling cap values that I have used for the KPC-9612 MAY not give you FLAT response down to 3 Hz. Be prepared to increase some values. If your brand of TNC has both ports hardwired in parallel, you may want to consider internally disconnecting the 9600 baud port and running separate leads out of the TNC for 9600 baud use only. Hanging a 600 ohm load directly on a discriminator can pose some problems.
Carefully align the unit per the manual when you have finished. Make certain that the 11.7MHz frequency (the off set oscillator) in the exciter board is on frequency. Do not attempt to correct transmit frequency by adjusting the 11.7MHz frequency - use the frequency trimmer on the bottom of the channel element. If the channel element is on frequency and the 11.7MHz oscillator is on frequency then the transmit frequency should be correct.
A UHF Micor delivering 100 watts draws approximately 30 amps. If you didn't get the power plug with your unit you're faced with supplying power to the unit. Connect pins 1 and 11 of the power connector (front of unit) to chassis ground. Connect pins 3, 8 and 22 together then through a 3 amp fuse to the big feed through cap carrying the +13.8 volts to the power amplifier stages. You can jumper these connections on the end of the PC board where the power connector socket is connected. Soldering to the large minus and plus pins (13.8 volts from the power supply) would be a difficult job - I haven't tried it. Soldering a pair of heavy, flexible leads to the INSIDE of the unit would be preferable.
Read the manual carefully so you know where to place the jumper(s) for correct single channel operation. This becomes particularly tricky if your unit has both off set oscillators and a universal switching board. You'll have to deal with channel switching if you plan to use more than a single channel.
Give serious consideration to running the power level down from maximum. If you use one on a BBS or a node you might want to add a "low" temperature thermostat directly to the middle of the heat sink to control a muffin fan blower. Ceramic substrate cracking is a serious problem with these units and the ceramics are no longer available from Motorola. A very large up/download could put you out of business! The VHF Micors don't have the cracked ceramic problem. They have a different problem in their power amplifiers. They also use phase modulation and wont transmit 9600 baud. They make great packet rigs but modifying them is another story for another time.
I built my control box from a miniature "Bud" box, miniature volume/squelch controls, channel switch and speaker. Even a little toggle switch on the PTT line to assist in tuneup. If you try this be careful: The speaker line is 2 separate leads - do not ground either in the control box. The squelch lead has a separate shield. The audio from the detector, the line from the volume control to the receiver audio and the microphone lead all share another shield. Use the squelch shield for the "ground" return on the squelch control and use the audio shield for the "ground" return on the volume control. Do not ground them in the control box. Use one of the actual chassis ground leads to ground your homemade control box. An LED can indicate power. Connect a 680 ohm resistor from plus 13.8 volts with the other end connected to the positive side of an LED. The negative side of the LED goes to ground. Another LED can be added to indicate "transmit" - This LED should have the positive side connected in the same manner as the "power" LED but the negative side of the LED should be connected to the PTT line.
Dan (WB9TPN), Bruce (WA9JQE) and Tom (WD9HSH) have been very helpful to me. They are professional service technicians in the business of making communications "stuff" work. Conversations with them have brought me back on track several times after stumbling and falling into the weeds. Every time I thought I had the final handle on the beast, one of them would point out another "oddity". Of course I plan on blaming them for anything that isn't correct in this piece but until then, I'll collect my 30 pieces of silver!
JAN Crystals (813) 936-2397
International Crystals 1-800-725-1426