Figure 1: The purple-box QRP Tx/Rx combo, almost ready for action (click to enlarge)
Shown in figure one is my very old Bencher paddle, which is itching to make MCU shaped CW dits and dahs with the new QRP rig. But, the transmit filters must be verified first. The Autek research keyer is as old as the paddle, but still works fine. It’s connected to the SCK key line of the purple box TX, and – along with the paddle – is waiting for the go signal. Remember those old CMOS keyers? The MK-1 keyer is bigger than the Qrp rig.
The Pi2 powered tablet is connected to the Progrock and, in spite of the fact that the screen is obscured by glare in the photo, it’s program is ready to set the frequency of the little QRP station. A picture of one of the LPF filter boards is shown below …
Figure 2: The extra harmonic filter is ready for action in this photo.
I’d used a kit for the PA output LPF in the TX box, but when I added an additional filter between the Progrock and the BS170 driver, I just put it on perf board.
I also cobbled together a little coffin for the filter. using double sided copper clad board pieces, and a lot of solder. Note the solder does the most good when it’s used on the inside seams, where there is copper-to-copper contact.
While I was making circuit changes, and adjusting the various things that needed to be adjusted, to guard against the possibility I’d fry the output IRF-510 device (of which I have only one in my possession) – I temporarily soldered an IRF-530 into its place. I have more of those to spare! There didn’t seem to be very much (if any) difference in performance, at least at 7 MHz.
I soldered the IRF-530 directly to the PCB, and found that for only one watt, a tiny clip-on style heatsink seemed to perform well enough for my regular 18 WPM CW transmit duty cycle.
OK, so everything has been great thus far. Seemingly too good. The power modulated raised cosine CW sounded absolutely beautiful. I’d had only one wiring boo-boo on an initial DC checkout, and other than that – everything “just worked.” No project ever stays that way all the way through, does it? I had been running the Si5351A synthesizer board continuously, and keying the driver for the PA, thru the power modulator “CW shaping” feature of the PA kit. This is done simply by attaching a key (or in my case, a 35 year old Autek Research keyer) to the SCK (key_in) pin on the PA kit board, and setting the Progrock to generate the operating frequency continuously via the default register bank (Bank0). That means using the Picocom serial terminal on the Pi2 powered tablet to set register #5 to 7,055,000, (using Clk1 as output – as an example).
To get rid of the contant carrier, the forum moderator at Qrp-labs suggested that I connect the keyer to the key_in line, as I was doing, but then also connect the key_out line (RCK on the PA kit board) – to one of the bank select pins of the Progrock frequency synthesizer. I connected the key_out line of the PA kit to the Bsel0 select pin of the Progrock, which caused Bank1 to be selected on the key-down state, because the other two select pins are kept high (5V) via pullup resistors. At that point, (for key down) the bank select inputs were Bsel0=low, Bsel1=high, Bsel2=high – and this meant bank #1 was selected. There are 8 possible banks, and three separate frequencies can be assigned to each bank. That happens because the Progrock has three clock outputs (Clk0, Clk1, and Clk2), and twenty four registers (plus some more for non-op-frequency data).
After making the RCK connection, I set register #5 to 0 (no RF output), and set register #8 to 7,055,000. So, when the keyer has the key down, for a dit or dah, RCK also goes low, and makes B0sel low, and switches the Progrock to the frequency of Bank1, Clk1 (what was programmed into register #8). Since the default bank registers are all zero (registers 5,6,7) – no output happens when the key is up. When the key is up, the progrock looks at registers 5,6,7 and sees all zeroes (no output). That condition puts Bsel0=high, Bsel1=high, Bsel2=high (Bank 0).
This actually works pretty dang well! BUT – for some reason, at this point, I noticed a little chirp, and a small amount of clicking on the keying. It was not objectionable, mind you, but definitely noticeable. So, did my last change prompt this to happen, or by coincidence did I (unintentionally) do something else to promote the problem?
In another photo somewhere on this site, I showed an old Micronta SWR meter with some RF output indication. Later, I swapped a slightly more trustworthy MFJ wattmeter into the line, and found that the output was five watts, right on the nose. I’ve been running it a three watts, to be conservative, into a dummy load (until I’ve checked the filters thoroughly).
To be continued …
Note: QRP-Labs at https://www.qrp-labs.com is not affiliated with this site or author in any way.
Note: The Raspberry Pi Foundation makes the Pi2 referenced in this post. They are not affiliated with this site or author in any way.