Making the µBITX transceiver “CW-worthy”

Some years ago I built a µBITX, version 3, and set out to make it useable on CW. It presented multiple challenges: lots of BCI from a local station, an audio IC that literally blew up, a problematic sidetone, lack of AGC and an impossibly wide passband, among others. I did manageable to make it useable, but just barely.

This fall I decided to try the latest version, V6, as a winter project. I was curious how much it had improved and whether I could make it “CW-worthy.”

V6 is much improved over the V3 I had earlier, but it still is not a CW rig. The Arduino-based firmware that comes loaded on it is problematic on CW, though that is easily fixed with other open-source options. However, the passband is still very wide.

The following are my fixes and mods.

AVC is mounted on the acrylic platform, left. Hi-Per-Mite is center. Zero beat indicator is right. S-meter is back right and the RF sensor is the small vertical board to the right of that.

First, I installed the more sophisticated software designed by N8ME and updated by VU3GAO. In addition to refining CW operation, if offers additional user options through the touchscreen. It also makes it unnecessary to use the encoder push-button feature. I often found it difficult to push the encoder without moving the knob and had added a push button in parallel with it that I no longer have to use.

An acrylic platform inside the case supported by nylon standoffs solved the problem of where to put any added circuit boards. I mounted a DC connection block there to make DC connections easy.

This receiver is extremely susceptible to BCI. I have a broadcast station nearby and even the BCI filter I have installed in my antenna coax was not enough to keep it out. I added the BCI filter from qrpme.com as described in my contribution to ubitx.net. This requires cutting and scraping a trace on the PCB. Initially, though, I used the qrpme.com filter (with the toroid options provided with the kit) as an external filter, attached directly to the rig’s BNC connector. It seemed to handle the RF power OK.

Next I installed a Hi-Per-Mite audio filter. I’ve used these filters in numerous rigs and they are quite amazing. In addition to narrowing the passband, they seem to act as a kind of peaking filter. I had thought I would include a switch to bypass the filter but I am CW-only, and found I didn’t use the very wide passband at all. (Also, if I want to bypass the filer, I can plug a speaker or earphones into the earphone jack on the front panel.) I made the filter connection using the speaker jack on the PCB.

I also added a zero-beat indicator using the basic design outlined in an earlier blog entry here. Since the audio filter is pretty narrow, this wouldn’t be essential, but I have a poor ear for sidetone tone frequency.

With no AGC and the peaking nature of the audio filter, a sudden loud signal can be pretty irritating. I considered a number of AGC options but settled on the audio limiting or AVC approach described in my earlier blog entry. This has been a huge improvement. Note that audio connections for this can found on the long connector on the uBitx PCB under the encoder.

I was unsatisfied with the supplied speaker so replaced it with one from the junk box – an Elecraft K2 speaker, I think. Note, though, that I did some work on a local ham’s uBitx and his speaker seemed fine, so it may just be that mine was defective. I also included a jack on the rear panel so that I could use an external speaker; I’ve found that the inexpensive MFJ-281 gives better sound.

I added an LED RF sensor so that I could be sure that I was really getting RF out.

My latest addition is a 10-segment LED audio-based S-meter. I used a kit board from Russia I got on eBay and after I was sure it was more-or-less working, cut out the front panel to mount the LED. The board itself is in the back right corner.

I first took the signal from the top of the volume control (using the multiple-pin connector under the encoder) to that it would be unaffected by the volume control setting. However, the passband of the uBitx is so crazy wide that I couldn't tell what signal I was reading. I am now taking the signal from the audio output and have marked the volume control where it has to be set to give an accurate reading. Also, the AVC has to be turned down on loud signals. It's not an ideal arrangement but it does give me a general idea of signal strength.

A problem that I cannot fix is that the T/R changeover tends to clip off the first dot or dash. I had not noticed this using a 250 millisecond delay and 18-24 wpm, but it showed up as a problem in a 15 wpm QSO where there was a longer delay between words. The delay is adjustable and it is best to choose a delay time that is long enough that there is no changeover during a transmission. This is definitely not a QSK rig.

One note: many uBitx users are replacing the display with a Nextion one and using KD8CEC firmware (see, for example, uBitx.net). I have heard reports about RFI from that display, however, and even though that would add some additional features, I decided not to go that direction.

It's been a fun winter project but it may be time to find another project to carry me through the rest of the winter. With these mods, though, it is a rather fun rig to use.