So far, I've found Larry Phipps, N8LP (QEX, Jan / Feb 2006) and Bill Kaune, W7IEQ (ARRL Handbook 2012) the best places to start. Be warned that while N8LP's LP-100A Operating and Assembly Manual gives good general detail on building the coupler, the LP-100A used a different coupler schematic where the voltage and the current are measured independently and the forward and reverse components are calculated by the display unit. That's how they can get away with mounting the one toroid on a standoff that's grounded on the one side.
Anyway. The problem with both of these designs is that the software (source code) is not available (N8LP) or not easily available (W7IEQ, ARRL members only). So if you want to make changes, you're on your own. Also, there's reason to believe (I have not seen the code) that the W7IEQ code is a bit buggy -- G0MGX ended up throwing it all out in favour of an Arduino, and he's even posted the source -- so that's where I'd recommend starting.
I like the idea of having multiple couplers, with one display unit (which does the heavy lifting).
In my opinion the log amplifiers should be housed in the coupler. This keeps all the precision stuff (specifically, the 50 ohm termination / attenuators) in one place, and allows for a single non-critical cable to the display unit. The downside of this is the additional cost of the AD-8307s -- but these are available quite cheaply on eBay, I got 10 for $4.44 (2016).
The N8LP schematic indicates "15 dB Pads" but as far as I can tell they're 20 dB. Not that it matters. N8LP also uses two in-line 10 dB attenuators for a total of 25 or 30 dB, depending on who you choose to believe. W7IEQ uses external 20 dB attenuators.
Of course you also need to factor in the attenuation of your transformers (which is dependent on the number of turns). With 31 turns the transformer output power is 0.00104 x input power or -30dB. For 1kW operation this means the termination (which in this case is also an attenuator) should be able to handle 1W. N8LP's "15 db" pads with 1W resistors for R5, R23 and R26 will do this easily, as those resistors will be dissipating 1/4W each at 1W from the directional coupler.
The maximum input to the AD8307 logarithmic amplifier is about +20dBm referenced to a 50 ohm load, which works out to 2.24V. -50 dB from 1kW is +10dBm or 0.707V which gives 10dB before the log amp runs out of headroom, which looks to be just about perfect.
This is a fairly generic ethernet switch, the "Mecer" is a transparent sticker. It was also buggered, and I've been contemplating re-purposing it for a while (I thought maybe a PIC-based frequency counter, but then I got hold of a Tabor 6010 which in addition to matching my Tabor 8020 signal generator is also a damnfine piece of kit).
Separate mains PSU and a christmas tree worth of LEDs.
Some work with a hacksaw and I've kept the LEDs and the PSU while ditching the ethernet bits.
I had this piece of plywood left over from another project. Nice thing about wood is that I positioned it where I wanted it and whacked it with a hammer and the indentations showed me where to drill the holes. Sorted.
Small problem: it's a 3.3V PSU and I need 5V. A bit of reverse engineering and a different resistor on the voltage reference and we have 5V. The 74HC164 shift registers can handle 5V.
The shift registers are arranged in two banks of sixteen, with a data and a clock input. I could also wire them in series, giving me a whole bunch (technical term) of LEDs with a two-wire interface. No idea what I'll use them for though.
I bought an LM018L 2 x 40 LCD on eBay a couple of years ago, specifically for this enclosure (I have a whole lot of 4 x 20 displays but of course they're too big for this box).
I'm running it in six-wire mode (RS, EN, D4, D5, D6, D7). EN is the only line that's dedicated, the other five lines can be used for other purposes as well.
R/W has to be grounded. Leaving it floating won't work. Ask me how I know. Also, the datasheet loosely indicates that the LCD drive voltage should be between 0 and 5V, so I initially connected it to 3.3V, because 3.3V is right there on the Arduino. Nope, the correct voltage is closer to half a volt. You could probably use the forward bias on a diode if you don't want to fit a pot.
But fit a pot.
Here's a nifty way of connecting up to 24 pushbuttons to 4 I/O pins, while still being able to use those same 4 I/O pins to drive an LCD display.
; The PortB is used as follows: ; RB0 FREE ; RB1 LCD ENABLE ; RB2 LCD R/W ; RB3 LCD RS ; < LCD DATA BUS > ; D7 D6 D5 D4 ; | | | | ; RB7 ---------o------|-------|--|---------|-,--|-,--|-,--|-,- ; | | | | 0|/ 1|/ 2|/ 3|/ ; RB6 --o------|------o-------|--|---------|-,--|-,--|-,--|-,- ; | | | | | 4|/ 5|/ 6|/ 7|/ ; RB5 --|--o---|------|-------o--|---------|-,--|-,--|-,--|-,- ; | | | | | | 8|/ 9|/ A|/ B|/ ; RB4 --|--|---|--o---|--o----|--o---------|----|----|----|--- ; _|_ | _|_ | _|_ | _|_ | C| D| E| F| ; \ / | \ / | \ / | \ / | | | | | ; _V_ | _V_ | _V_ | _V_ | | | | | ; | _|_ | _|_ | _|_ | _|_ | | | | ; | \ / | \ / | \ / | \ / | | | | ; | _V_ | _V_ | _V_ | _V_ | | | | ; | | | | | | |__|_________| | | | ; |__| |__| |__|______________________| | | ; | |__________________________________| | ; |______________________________________________|
The table associated with this ASCII-art schematic is wrong, so I'm not reproducing that here. But basically, you start by pulling RB7 low and reading RB4 to RB6. Pressing "0" will pull RB4 and RB5 low, pressing "1" will pull RB4 and RB6 low, pressing "2" will pull RB4 only low (RB7 is an output, remember?) and pressing "3" will pull RB5 and RB6 low. And then you scan the next line. Problem is you can't detect two switches being pressed at the same time.
1 --/\/\/--o-----------o---------------o--------o-----------o--------
| | | | |
2 --/\/\/--|---o-------|---o---o-------|--------|--o--------|--o-----
| | | | | | | | | |
3 --/\/\/--|---|---o---|---|---|---o---|---o----|--|--o-----|--|--o--
_|_ | _|_ | _|_ | _|_ | _|_ | _|_ | | _|_ |
\ / | \ / | \ / | \ / | \ / | \ / | | \ / |
_V_ | _V_ | _V_ | _V_ | _V_ | _V_ | | _V_ |
| _|_ | _|_ | _|_ | _|_ | _|_ | _|_ _|_ | _|_
| \ / | \ / | \ / | \ / | \ / | \ / \ / | \ /
| _V_ | _V_ | _V_ | _V_ | _V_ | _V_ _V_ | _V_
| | | |___| |___| |___| |__|__| | | |
| | | | | | | | | |
o o o o o o o o o o
| | | | | | | \ \ \
o o o o o o o o o o
4 _________|___|___|___|_______|_______|________| | | |
| | |
5 __________________________________________________________|__|__|
This is a much simplified schematic without the duplexing trick. It will give me seven pushbuttons and three on/off switches using the five lines (RS and the four data lines) that are also being used by the LCD. Can't see myself needing that many though.
The series resistors are required so that lines 4 and 5 are not pulled high by LCD data on the other three lines when the switches (or pushbuttons) are on.
(BTW numbers 1-5 above are arbitrary. In my application they will probably end up being lines 2-7 minus whatever one's EN on the Arduino).
The N8LP article mentions 26 turns on an FT-114-61, while W7IEQ used 31 turns on an FT-82-67 (as did G0MGX). The LP-100A uses 26 turns on an FT-140-61.
I obtained six x FT-82-67 from eBay ($20 in 2016).
I like N8LP's idea of using a VGA cable between the coupler and the main unit, because they're dime a dozen, available in different lengths, and contain at least three screened cables (R, G and B) which can be used for the forward and reverse voltages.
|
|
Back to ZS1KE's Page | (This page last modified 2016-08-05) |