For anyone interested, here's the latest radio we had converted by a certain sydney tech. She called it the EXPANDA board.
I agreed to let my old hornet 2 be the guinea pig for this one and the result was way better than i expected, the final version should be awesome!!!
< START OF QUOTE >
December project - "Expanda" mod board concept / prototype
Over the last few months we've been asked - both in online forums and here at the shop - for an expansion modification that suits a wide variety of radios, but mainly the late 80's onwards Uniden SSB chassis that uses the uPD2824 PLL chip. This is used in many radios, these are just a few:
Uniden PC-122
Uniden AX-144
Uniden Pro640e
Uniden Pro810
Cobra 146
Realistic TRC-451 and TRC-453
Pearce Simpson Super Cheetah / Stalker 9
Hornet 2
First thing to do was draw up a list of requirements, and whether they were just nice to have, almost essential, or must have / non-negotiable.
After asking a bunch of people, the following was decided on:
1 ) Radio must look stock from the outside
This ruled out things like digi-scan boxes and jiffy boxes with C&K toggle switches mounted in them.
2 ) Radio to retain original "look and feel"
No controls that look different to the originals, and no multi-position switches to replace pots such as tone, dimmer, or RF gain.
3 ) Electronic changes such as added boards are OK, but no or minimal mechanical alterations
This rules out changes such as encoder switches to replace the channel switch, changing the display for a different type, or changes to things like front panel control support metalwork.
4 ) No obscure, obsolete, or difficult to obtain (in small quantities) components.
This rules out crystals, "companion parts", and "brand X" unknowns with the part numbers ground off the IC tops.
Also not allowed: overpriced special parts that can only be obtained from one supplier, and "protected" or pre-programmed digital chips like PICs. Programmed parts are OK, as long as the software / firmware is available legally for no cost and the programming hardware is under $10 shipped.
5 ) No channel "lookup" charts needed.
The channel and/or frequency should appear where the user is looking for this information, i.e. on the original channel display digits.
6 ) Good RF performance, at least the same level as the radio was when stock.
This rules out things like crystal oscillator boards (limited clarifier range, too drifty, and difficult to center the clarifier in all modes), multiplexed displays (too noisy), and DDS eBay modules (no FM Tx audio possible, and phase noise too high, causing splatter).
7 ) Simple for the installer to configure the desired frequency coverage
This should ideally be able to be done without changing wiring or components.
The same hardware should be used whether you just want an extra bank of 40 channels or are doing an extreme mod covering 3MHz or more.
The only differences should be the "broadbanding" changes to the front end, output stage, and the mixers in the radio.
8 ) Adaptable into as many different radios as possible with none or minimal changes to the installation process
The initial design should work in any radio, as long as these 3 conditions are met:
- Radio is a PLL type with a VCO (anything made after 1977 should be OK for this one, but most 23 channel sets miss out here)
- Radio stops transmitting when the channel selector is placed "in between" channels
- Radio has a digital readout with at least 2 digits
That should cover most sets. But, for the ones that don't:
Future possibility: an on-board VCO to overcome limitation #1. Only an idea at this stage though
Condition #2 is a non-negotiable for reasons that will be explained down the page a bit.
Only workaround for this one is to change the channel selector switch...
Condition #3 above is not an issue (i.e. the early NDI and Cybernet PLL rigs with a back-lit dial channel readout),as long as you don't mind adding an internal or external LED channel display.
We've since been asked from various places for versions to suit the Cobra 148 (SSB), the Cobra 29 (AM only), an obscure Midland AM/FM set from the 80's the owner wanted putting on the 29MHz amateur band, and an equally obscure "Wagner" SSB base radio made by GRE (the Japanese OEM builder that made the Realistic TRC-448 and many of their VHF/UHF scanners).
Got another PM tonight, asking if it'll work in an 18 channel (old Oz allocation from the late 70's) Cybernet SSB. Yes it will...
Thought about it for an hour or three, then decided on the hardware that would meet all 8 of the requirements above.
The "brains" or core technology - the Arduino Nano clone. Contains a microcontroller, EEPROM, USB port, and on-board 5V and 3.3V regulators.
For under $4 with free shipping, this one was a no-brainer.
If you've never used a micro before, it doesn't get much easier than these. Once you've done one or two it'll be second nature.
The programmers cost around the $5 mark with a dollar or two for postage. The software is free and open source.
The bottom of the Nano, showing the USB circuitry (that lets it talk to a computer) and the voltage regulators.
A PCF8574 is used as an interrupt generator, no need to know what that means to build the circuit though.
Basically, it sends the Nano to sleep until something happens, such as the user changing channels.
It then wakes up the Nano, which does its stuff for about 1/100th of a second before going back to sleep.
This is done to reduce the power consumption a bit, but mostly to keep noise - that could interfere with weak signals being received on the radio - to a minimum.
The Nano is mounted on a perfboard with the PCF chip and a few extra components (diodes and capacitors) for protection and noise reduction.
The copper side of the perfboard containing the Nano and PCF chips.
The display driver does not use 7447, 7448, or 4511 logic chips, but uses an MC23017 and a pair of ULN2003 drivers instead.
This allows the original 8V or 12V LED power source to still be used (which means the dimmer function of the radio will work the same way it always did). Other advantages are the digits "6" and "9" will look the same as they did before (the older chips will show 6 and 9 with one less segment), and the pinout is user-definable. This means the LED wiring is much neater, just connect the 14 wires in the neatest / easiest way, then define which LED segment belongs to which wire later.
The copper side of the display board.
A new PLL chip - which is only available cheaply in SMT form - was then made with an SMD to through hole adapter board.
I first though I could get away with this because the new PLL is only partially used, to "trick" the non expandable PLL.
A few measurements showed the phase noise level was awful, barely passable on AM and unusable on SSB.
The whole thing was live with RF energy which in turn got into the dividers and the VCO control pins.
We then bought back an old project board from a couple of years back, double sided with ground planes and lots of vias to join them.
Phase noise was now gone, but the board was so big it covered up some tuning slugs we needed access to for tuning. Time for the drill...
First attempt at a new PLL = failure.
New secondhand PLL board, this one worked fine.
A temporary 2 digit LED of the same type was used to allow the initial firmware to be written without the need for a radio.
The LED temp board rear, wire first and configure the pinout second. Much neater than fixed function pins
The radios ANL/NB switch was rewired as a function selector switch.
UP = standard FCC 40 channels
CENTER = 26.215 to 28.805 in 10KHz steps
DOWN = 26.210 to 28.800 in 10KHz steps, i.e. a 5KHz drop from the center position.
Noticed the meter bulb was dark, dim, and had started to melt the clear plastic of the meter.
While it was apart, may as well do a quick 5 x Super-flux SMD LED conversion, which gives a very even light coverage.
PLL trace cuts to disable the section of the original PLL we will no longer be using.
Trace cut needed on the radios LED board to separate segments "A" and "D" on the tens digit.
This is needed so a "7" can be displayed.
13 LED resistors need to be removed, to isolate the channel switch from the display
LED resistors no longer required...
The radios display board with the resistors removed
The new LED wiring. Again, wire first, configure pinout second. It makes for a neat wiring job.
The new channel switch wiring.
How it works:
Four wires are the "BCD" representation of the "units" channel LED, these are:
- Black ( 1 )
- Brown ( 2 )
- Red ( 4 )
- Orange ( 8 )
These are at +5V when active and match the units digit.
For example, on a channel ending in "5" the black and red wires are positive, because (1) + (4) equals 5.
Likewise, on a channel ending in "9" (9, 19, 29, and 39) the black and orange wires will be at a positive voltage level.
The yellow wire connects to the pin that used to stop the radio transmitting in-between channel positions.
It is normally positive, but drops to close to zero volts when the channel selector is in between channel positions.
The PCF8574 is waiting to pounce when the yellow wire changes its voltage level.
When the yellow wire drops from 5V down to 0V, the voltages on the other 4 wires are memorized by the Nano.
When the yellow wire rises from 0V up to 5V, the voltages on the other 4 wires are then compared with what it memorized.
If the BCD number is one higher, or has changed from "9" to "0", the user must have turned the selector to go UP.
If the BCD number is one lower, or has changed from "0" to "9", the user must have turned the selector to go DOWN.
The Nano then adds or subtracts one channel or one 10KHz step as required by the function the radio is currently set to.
The whole process seems a bit haphazard, but it happens in the blink of an eye (about 1/100th of a second).
< END OF PART 1 >
I agreed to let my old hornet 2 be the guinea pig for this one and the result was way better than i expected, the final version should be awesome!!!
< START OF QUOTE >
December project - "Expanda" mod board concept / prototype
Over the last few months we've been asked - both in online forums and here at the shop - for an expansion modification that suits a wide variety of radios, but mainly the late 80's onwards Uniden SSB chassis that uses the uPD2824 PLL chip. This is used in many radios, these are just a few:
Uniden PC-122
Uniden AX-144
Uniden Pro640e
Uniden Pro810
Cobra 146
Realistic TRC-451 and TRC-453
Pearce Simpson Super Cheetah / Stalker 9
Hornet 2
First thing to do was draw up a list of requirements, and whether they were just nice to have, almost essential, or must have / non-negotiable.
After asking a bunch of people, the following was decided on:
1 ) Radio must look stock from the outside
This ruled out things like digi-scan boxes and jiffy boxes with C&K toggle switches mounted in them.
2 ) Radio to retain original "look and feel"
No controls that look different to the originals, and no multi-position switches to replace pots such as tone, dimmer, or RF gain.
3 ) Electronic changes such as added boards are OK, but no or minimal mechanical alterations
This rules out changes such as encoder switches to replace the channel switch, changing the display for a different type, or changes to things like front panel control support metalwork.
4 ) No obscure, obsolete, or difficult to obtain (in small quantities) components.
This rules out crystals, "companion parts", and "brand X" unknowns with the part numbers ground off the IC tops.
Also not allowed: overpriced special parts that can only be obtained from one supplier, and "protected" or pre-programmed digital chips like PICs. Programmed parts are OK, as long as the software / firmware is available legally for no cost and the programming hardware is under $10 shipped.
5 ) No channel "lookup" charts needed.
The channel and/or frequency should appear where the user is looking for this information, i.e. on the original channel display digits.
6 ) Good RF performance, at least the same level as the radio was when stock.
This rules out things like crystal oscillator boards (limited clarifier range, too drifty, and difficult to center the clarifier in all modes), multiplexed displays (too noisy), and DDS eBay modules (no FM Tx audio possible, and phase noise too high, causing splatter).
7 ) Simple for the installer to configure the desired frequency coverage
This should ideally be able to be done without changing wiring or components.
The same hardware should be used whether you just want an extra bank of 40 channels or are doing an extreme mod covering 3MHz or more.
The only differences should be the "broadbanding" changes to the front end, output stage, and the mixers in the radio.
8 ) Adaptable into as many different radios as possible with none or minimal changes to the installation process
The initial design should work in any radio, as long as these 3 conditions are met:
- Radio is a PLL type with a VCO (anything made after 1977 should be OK for this one, but most 23 channel sets miss out here)
- Radio stops transmitting when the channel selector is placed "in between" channels
- Radio has a digital readout with at least 2 digits
That should cover most sets. But, for the ones that don't:
Future possibility: an on-board VCO to overcome limitation #1. Only an idea at this stage though
Condition #2 is a non-negotiable for reasons that will be explained down the page a bit.
Only workaround for this one is to change the channel selector switch...
Condition #3 above is not an issue (i.e. the early NDI and Cybernet PLL rigs with a back-lit dial channel readout),as long as you don't mind adding an internal or external LED channel display.
We've since been asked from various places for versions to suit the Cobra 148 (SSB), the Cobra 29 (AM only), an obscure Midland AM/FM set from the 80's the owner wanted putting on the 29MHz amateur band, and an equally obscure "Wagner" SSB base radio made by GRE (the Japanese OEM builder that made the Realistic TRC-448 and many of their VHF/UHF scanners).
Got another PM tonight, asking if it'll work in an 18 channel (old Oz allocation from the late 70's) Cybernet SSB. Yes it will...
Thought about it for an hour or three, then decided on the hardware that would meet all 8 of the requirements above.
The "brains" or core technology - the Arduino Nano clone. Contains a microcontroller, EEPROM, USB port, and on-board 5V and 3.3V regulators.
For under $4 with free shipping, this one was a no-brainer.
If you've never used a micro before, it doesn't get much easier than these. Once you've done one or two it'll be second nature.
The programmers cost around the $5 mark with a dollar or two for postage. The software is free and open source.
The bottom of the Nano, showing the USB circuitry (that lets it talk to a computer) and the voltage regulators.
A PCF8574 is used as an interrupt generator, no need to know what that means to build the circuit though.
Basically, it sends the Nano to sleep until something happens, such as the user changing channels.
It then wakes up the Nano, which does its stuff for about 1/100th of a second before going back to sleep.
This is done to reduce the power consumption a bit, but mostly to keep noise - that could interfere with weak signals being received on the radio - to a minimum.
The Nano is mounted on a perfboard with the PCF chip and a few extra components (diodes and capacitors) for protection and noise reduction.
The copper side of the perfboard containing the Nano and PCF chips.
The display driver does not use 7447, 7448, or 4511 logic chips, but uses an MC23017 and a pair of ULN2003 drivers instead.
This allows the original 8V or 12V LED power source to still be used (which means the dimmer function of the radio will work the same way it always did). Other advantages are the digits "6" and "9" will look the same as they did before (the older chips will show 6 and 9 with one less segment), and the pinout is user-definable. This means the LED wiring is much neater, just connect the 14 wires in the neatest / easiest way, then define which LED segment belongs to which wire later.
The copper side of the display board.
A new PLL chip - which is only available cheaply in SMT form - was then made with an SMD to through hole adapter board.
I first though I could get away with this because the new PLL is only partially used, to "trick" the non expandable PLL.
A few measurements showed the phase noise level was awful, barely passable on AM and unusable on SSB.
The whole thing was live with RF energy which in turn got into the dividers and the VCO control pins.
We then bought back an old project board from a couple of years back, double sided with ground planes and lots of vias to join them.
Phase noise was now gone, but the board was so big it covered up some tuning slugs we needed access to for tuning. Time for the drill...
First attempt at a new PLL = failure.
New secondhand PLL board, this one worked fine.
A temporary 2 digit LED of the same type was used to allow the initial firmware to be written without the need for a radio.
The LED temp board rear, wire first and configure the pinout second. Much neater than fixed function pins
The radios ANL/NB switch was rewired as a function selector switch.
UP = standard FCC 40 channels
CENTER = 26.215 to 28.805 in 10KHz steps
DOWN = 26.210 to 28.800 in 10KHz steps, i.e. a 5KHz drop from the center position.
Noticed the meter bulb was dark, dim, and had started to melt the clear plastic of the meter.
While it was apart, may as well do a quick 5 x Super-flux SMD LED conversion, which gives a very even light coverage.
PLL trace cuts to disable the section of the original PLL we will no longer be using.
Trace cut needed on the radios LED board to separate segments "A" and "D" on the tens digit.
This is needed so a "7" can be displayed.
13 LED resistors need to be removed, to isolate the channel switch from the display
LED resistors no longer required...
The radios display board with the resistors removed
The new LED wiring. Again, wire first, configure pinout second. It makes for a neat wiring job.
The new channel switch wiring.
How it works:
Four wires are the "BCD" representation of the "units" channel LED, these are:
- Black ( 1 )
- Brown ( 2 )
- Red ( 4 )
- Orange ( 8 )
These are at +5V when active and match the units digit.
For example, on a channel ending in "5" the black and red wires are positive, because (1) + (4) equals 5.
Likewise, on a channel ending in "9" (9, 19, 29, and 39) the black and orange wires will be at a positive voltage level.
The yellow wire connects to the pin that used to stop the radio transmitting in-between channel positions.
It is normally positive, but drops to close to zero volts when the channel selector is in between channel positions.
The PCF8574 is waiting to pounce when the yellow wire changes its voltage level.
When the yellow wire drops from 5V down to 0V, the voltages on the other 4 wires are memorized by the Nano.
When the yellow wire rises from 0V up to 5V, the voltages on the other 4 wires are then compared with what it memorized.
If the BCD number is one higher, or has changed from "9" to "0", the user must have turned the selector to go UP.
If the BCD number is one lower, or has changed from "0" to "9", the user must have turned the selector to go DOWN.
The Nano then adds or subtracts one channel or one 10KHz step as required by the function the radio is currently set to.
The whole process seems a bit haphazard, but it happens in the blink of an eye (about 1/100th of a second).
< END OF PART 1 >
