QRP Labs QMX SSB beta firmware relased | Hacker News Reader

QRP Labs QMX SSB beta firmware relased

qrp-labs.com

72 pointsDrAwdeOccarim24 comments

These new little software based radios are amazing! I'm traveling in the UK now now with a truSDX -- I've run it from all sorts of places with a wire hanging out hotel windows.

SSB in these types of radios--including the QMX-- is generated with the "phase method" https://electronicspost.com/draw-and-explain-the-block-diagr...

You can change the phase of the Si5351 chip fast enough over its seral interface to accomplish this.'

I've been an Extra since 1977, and I've never seen a better time to be a radio hobbiest! So many things are possible today. I still run mostly CW, copying in my head, but I enjoy the digital modes when doing "hotel room" operations with makeshift antennas.

I was wondering how Hans got it to work. All the groups.io threads and I couldn't find a simple explanation.

I have a partially assembled QMX+ sitting on my bench. I need to get it finished.

I'll let somebody else talk about why this firmware update is somewhat earth shattering for the ham radio community and why the QMX itself is a huge deal. We're talking a $1K SDR for $150, multiband up to 6M, fully digital.

I can’t find this in the spec: does it tune continuously from 160 to 50m or only in the ham bands?

If it's anything like the PE1NNZ uSDX/derivatives it tunes only in the (rough) ham bands - obviously if you modify the firmware to remove the end limits of the ham band you can tune to other frequencies but fundamentally, the thing has a bunch of relays and coils for the different bands.

Only the ham bands, and there are three different models that hit different harmonics of the bands: 80/60/40/30/20m; 60/40/30/20/17/15m; and 20/17/15/12/11/10m.

Tl;dr: yes, but it is really only intended for amateur use.

QMX(+) can technically tune anywhere, you can go into the configuration and adjust the band RF ranges so you have full control over what frequencies the QMX(+) will attempt to tune. These configurations are there so a law abiding ham can keep themselves from making the silly (and illegal) mistake of transmitting out of band.

But it has band pass filters around the ham bands, so you’ll lose sensitivity the further you get from a ham band in general. There would be some places where you wouldn’t lose sensitivity- the 10m and 12m band use the same filter, so you would get full sensitivity between 25-28 MHz. As far as I’m aware you must select a band pass filter and cannot bypass without a hardware mod (I honestly wouldn’t be surprised if Hans put in a bypass for diagnostic purposes but I don’t know). There is almost certainly a way for you to make one of the band pass filters very wide, but this would be a substantial modification, requiring you to figure out the inductance values yourself.

On the Tx side the issue is low pass filters. You can configure which filter bank gets used, so you could configure it to transmit on any frequency. But the filters are designed specifically to keep the the transmissions clean and legal- there are certainly frequencies where you would have to select a low pass filter that does not suppress a harmonic to legal levels.

Footnote: the 160m band has issues on the QMX+ but I’m pretty sure that will get fixed in another firmware release. But for now the useable range is ~3.5-54 MHz and not all the way down to 1.5 MHz.

  has band pass filters around the ham bands, so you’ll lose sensitivity the further you get from a ham band in general.

Thank you, was looking for this.

> is somewhat earth shattering

why would it be earth shattering?

(tr)usdx has ssb support for a similar price and the internal microphone actually works. It also has a digital interface (a bit unstable at times though) for digimodes without neding soundcards and aditional cables.

Maybe earth shattering isn’t the right word, but this is a big deal for fans of QRP labs. The (tr)usdx is a bit cheaper hardware, which limits the performance. The QMX has a more powerful processor and will outperform the (tr)usdx by several dB in terms of distortion.

Hans (the designer) is an amazing developer and very thorough, and his kits are very high quality. He includes several diagnostic tools in his firmware, so you don’t just build a kit, but you can diagnose and tune the performance with the kit itself. A beta release from Hans is practically a release candidate in everyone else’s eyes because he’s so thorough.

In terms of functionality, this SSB mode means you can do practically any modulation now, where previously you could only do single frequency fixed amplitude modes. So now you can do PSK, winlink, SSB voice, and whatever modes become popular in the future.

Hans took a while developing this because he wanted to get the max performance from the tiny ~5W transmit power, so he implemented a technique called controlled envelope SSB which was clearly challenging to get right, and I’m not sure the (tr)usdx even has the processing power to implement it.

I’ll probably buy another kit because of this just because I’m so excited (I built a QMX+ last year). This makes the QMX a really good radio. It has always been high quality, it was just held back by only being able to do CW and a certain digital modes. Being able to do voice and ALL digital modes at this price range and quality is unbeatable.

There's a lot of unusual and flaky design in the boring plumbing parts of these things that makes them prone to self destruction. For example, using the MCU to make discrete buck regulators that power the MCU is a o.O choice that I think has no reasonable justification. There are perfectly wonderful multi output integrated chips with external synchronization that cost almost no money, take up less space, are more performant, and make it impossible for software to immolate the hardware upon which it runs.

Gifted embedded software, but the electronics is simply not that robust. The company could really use an electrical engineer to complement its software engineer.

The QMX works well enough that I'm not going to critique the choice of voltage regulation or other plumbing bits. The boring parts work, even if they weren't designed by a professional EE. It uses low power on receive and is reliable, which are my primary concerns. I don't expect to get the same design as a team of Yaesu engineers would use on a $5k transceiver. If it's not hurting RF performance, then it is fairly low on the wish list.

Overall, the kit is still a killer value and buildable by someone who can follow directions and is decent at soldering. This is in stark contrast to many other kits which require a lot of sleuthing to figure out issues because the instructions aren't clear, or they changed some components and didn't update their documentation. And these other kits often don't have any better design for their power management sides either.

There have been a few board revisions to address some of the bigger issues that have come up with early board spins, and I like that mods have been made public for most of the early builders to keep their kits working.

The most common problem seem to be blown transistors from transmitting into a mismatched load, and that is mostly the operator's fault and common in many radios. I like that Hans put some SWR protection in the QMX to keep the finals from blowing as often as they might.

So many toroids.

Do you know has it been adapted to work on the 4 and 5 metre bands?

I can't say for sure, but I think you're going to start to struggle with performance and I wouldn't want to spend a lot of effort on it.

It is already a challenging to get good 6m performance from the QMX+. You would have to redesign the filters for 4/5m and you might still get poor performance due to being close to the limit of the si5351 chip (the frequency source), as well as the fact that parasitic capacitance on the PCB is going to start to matter. You might also have to modify the firmware, so it might end up being a lot of effort for the modification.

I built a QDX - great little radio.

You might be interested in the QubeDX: https://rz01.org/when-two-become-one/

Very nice to see this on the front page!

Would it be capable of Winlink now? If not what would that take?

Maybe. When using ARDOP, my laptop generates an audio stream that is then transmitted over SSB, so theoretically it might work.

We have SSB, audio in/out, and cat control, so all the pieces are there.

I'll be testing this weekend, but theoretically yes- multi-tone modes like winlink are specifically mentioned and this release was held until it could pass a two-tone test with decent performance.

did it work?

it could always do winlink at the slower VARA speeds

I don't know anything about the product, but just reading the notes, as soon as I read:

With ALL features enabled (CESSB, compression, mic AGC, Transmit equalization, noise gate, phase and amplitude pre-distortion, you name it), the CPU utilization is approximately 93%. So in other words there is no problem enabling everything.

... I knew I had found some embedded developers. :) Cheers from another STM32 developer. :)

Reply on news.ycombinator.com