An Incomplete Repair: BladeRF x40 Software-Defined Radio

My last trip to Singapore, Japan and Hong Kong kept me away from home for 80 days, in which time I thoroughly enjoyed doing things I couldn’t otherwise do from home. Despite arriving back home almost a month ago, it has taken a while for me to finally overcome the holiday hangover and complete the process of “unpacking” my life once again. While I haven’t had any time to report on any of my travel in-depth (yet), the first trip went smoothly and (mostly) trouble-free; the latter proved somewhat cursed.

On the one hand, I returned home to find my luggage had not made it with me, leaving me with an anxious two-day wait to find out if it would be coming back. On the other hand, a number of things failed during this holiday – including:

  • various items of clothing (socks, t-shirts, undergarments, wallet, shoes)
  • my Mi Band 2’s silicone strap
  • the eye-cup on my Nikon D3300’s viewfinder
  • the frame on a set of sunglasses snapped
  • the auto-focus motor on my Tamron lens (noisy, but itself a replacement for the lens that had a loose front element on the last trip)
  • an external drive (bad sectors, but no data lost due to backup and drive retested OK)
  • a radio antenna (physical wear)
  • the USB port on the BladeRF x40 SDR (very sad)
  • the charging port on a Fenix PD40R torch (also sad)
  • my Audio-Technica ATH-ANC9 noise cancelling headphones (multiple failures)

Luckily for me, I did have enough equipment to keep me going for the most part. As an engineer, I attempted a repair with the limited resources I had, and if not, managed to purchase a substitute or replacement. I guess my luck lay in the fact the failures happened at times where it didn’t critically impact my enjoyment of my holiday. The trail of broken things even led home, leaving me with things awaiting a repair – the most urgent was the door-lock to the house! Fixing everything once I got back was a high priority, to the point that I didn’t even have the time to document my fixes or blog about them (and probably never will).

Thankfully, the luggage did find its way back to me in one piece. It was complete with the requisite transfer tags, coded for automatic forwarding, complete with tags telling the handlers to “rush” the bag through Singapore transit as the passenger was already on board, but even with 2 hours and 50 minutes at Changi Airport to make the transfer, I did and the bag didn’t. At least I didn’t have to live without my gear for the duration of my holiday.

My Sorrow

My happiness with being reunited with my luggage quickly subsided into sorrow when I pulled out my radio gear container to be reminded about what happened during one of my last days in Hong Kong.

The BladeRF x40 and XB-200 transverter have been my space-and-weight-efficient travelling SDR radio companion since 2014. It’s not perfect by any means – I’ve had some issues with conducted power-noise on the unit and reception images due to slight I/Q imbalances, but it is a mighty combination that easily provided 20Mhz+ of real usable spectrum to make scanning and recording the airwaves both fast and efficient. Add a simple SMA extendable whip for regular VHF/UHF transmissions, combine a Sony AN-LP1 for shortwave, a few adapters and a chunky laptop to boot and you basically have my travel-kit.

In 2016, I finally got around to posting about the unit, and I had remarked:

As this is an earlier BladeRF, it has a microUSB-B connector on the input which feels a little fragile …

Little did I know that those words would foreshadow exactly what happened a week before the end of my holiday. Knowing that many places may look upon radio-enthusiasts with an ire of suspicion, I packed my radio gear out of sight every single day before leaving the hotel room. However, this day I was a little pressed for time and placed the gear into the luggage case without disconnecting the USB lead. When it came to unpacking the radio, the cable had taken the connector off the board, along with all but one of the traces.

At this time, I was kicking myself for not ensuring I had removed the cable before packing it, and for not making sure it wouldn’t be somehow “butted up” against something which might break the connector. I was also somewhat miffed that I didn’t end up waiting a little longer since the more modern variants of the board have a full size USB-B connector that anchors itself solidly to the board with through-hole mounts rather than surface-mount pads.

With this SDR out of action, I only had the RTL-SDR to fall back upon. At about 1/10th the bandwidth with a much more compromised dynamic range, I quickly decided just to pack it all up and deal with it once I got home. It was a sizeable investment to obtain the board, and I really didn’t want to have to fork out again just to get another – especially when the exchange rate is not as favourable as it was when I first bought it. However, losing the capability was also something I didn’t want to happen. I enjoyed the convenience of a wideband SDR too much to abandon it now.

An Idea to Try?

A closer look at the damage shows just how bad it was. The pads are practically all torn off, with the exception of the USB 2.0 ground pad which is somewhat lifted. The anchoring pads at the front have delaminated partially as well.

The connector shows exactly what was attached to it at the point it tore off the board. At this stage, the connector is practically useless.

Not willing to give up on the board, I decided to try and rescue it. How? The first step was to try and restore some solderable area to the traces. As the traces don’t fan out much from the pads, I decided to sand away the solder resist with some fine-grained wet-and-dry sandpaper (ignoring any ESD precautions).

To stop copper oxidation, I tinned the exposed copper almost immediately. In theory, if I set the connector back about half the width of a connector, I should be able to make the connection. Right?

The next step was to order a new connector from element14 – 2443129 made by Global Connector Technology, as it was the only compatible part with local stock. I then decided to tin the connector pins and then try to reflow them for a more solid connection.

I had to add some extra solder to positive connection to bridge the gap. At this stage, it visually looked somewhat poor, but due to the fine pitch, it was hard to do it by hand unassisted. Part of the problem was the solder flux was getting in the way of the joints being properly wetted, and the thin traces made it hard to confirm whether the connection is made.

As a result, I plugged it in … but nothing happened. Nothing.

As it turned out, the connections weren’t being properly made due to the flux that was getting sticky and messy. In an act of desperation, I got out my reflow hot air gun and cooked the connector. Now it wouldn’t even accept the plug. Damn it!

Take Two: We Learned Something Right?

Knowing that repairs don’t always go right the first time, I ordered two connectors as that was all they had in local stock. The first connector was then desoldered from the board, the board cleaned-up, and then an attempt to mount the fresh connector was made.

This time, the connector did manage to make connection with all pins and the SDR was detected at USB 3.0 rate. Great, or so I thought …

I knew the connector would be fragile in its present state, so I mixed up some high-strength slow-cure Araldite two-part epoxy to run over the connector. That would provide mechanical support to the connector. What I didn’t realize was that it was very runny, so it ended up gluing the cable connector into place as well.

But once the epoxy had set, it was found not to be working. Visual inspection didn’t find anything amiss, but it seems likely that during the course of putting the epoxy on, the physical stress of rotating the board around with the cable connected to the board was enough to crack the joint.

Last Ditch Effort

Now, I had a board with a blob of epoxy set rock hard, and a cable attached that doesn’t work. I felt absolutely defeated. So I decided it was probably lost and didn’t take any photos.

But once I realized that Araldite loses strength at high temperatures, I waved a hot air gun at the epoxy blob enough to burn away the epoxy enough to remove the connector and give me one last shot. But this time, it took away even more of the remaining traces.

My last effort centred around a two stage connector-less process. The first step would be to attach the USB 2.0 “high-speed” wires (red-white-green-black), so that if that succeeded, I would have a partially working unit with a bandwidth restriction. Once that was solidified, I would attempt to attach the USB 3.0 differential pairs (five connections including the drain wire) to see if I could enable high speed. In between each stage, I would secure the wires with more Araldite epoxy.

The USB 2.0 wires were a small challenge, but not one I haven’t attempted before. In fact, I succeeded, thus quickly applied epoxy. It was quite runny, so started to impinge on unrelated components and pads in short order. After three days curing time, I attempted to attach the USB 3.0 wires, but after four attempts and an hour of fiddling, I could not get USB 3.0 connectivity at all. Sadly, despite all my desire to get it to work, I couldn’t.

I even tried to solder to the legs of the filter chip, to no avail. The wires wouldn’t stay put, and when they did, they would often short-out to adjacent pins.

With that, I officially gave up on making further attempts. The traces were all raggedy. Besides, a half-working unit is better than a dead unit, right? As a result, I epoxied the stripped wire ends of the USB 3.0 cable to the board so as to provide mechanical support, so hopefully the repair allows the unit to do some limited travelling.

Conclusion

Most products of the modern era are not really designed for repair. Notwithstanding this, when something is damaged, the likelihood of repair is never guaranteed. The BladeRF x40 was my favourite SDR for travel, and its capabilities are far beyond those of my existing arsenal. Losing it was unacceptable.

As a result, I hatched a plan to try to repair it, but the first attempt was not successful. A second attempt was partially successful, but proved exactly how fragile soldering to thin traces are. It just won’t work with the mechanical stress of the plug/unplug cycle.

A change of plan to directly wire to the board proved to be the partial success that reclaimed the board. The USB 2.0 connections were made reliably, but the USB 3.0 proved too much of a challenge. At least, having something is better than nothing …

Now, the SDR is capable of about 10-11MSPS with no buffer drops depending on which USB 2.0 controller is in use. Previously, I would regularly operate at 28 to 32MSPS, so I suppose it’s a bit of a loss in that regard. In the end, I could have done a lot worse. After all, the success rate for an attempted repair probably sits at 60/40, where there is a 40% chance of success. Many things are a total loss, sometimes because of a lack of skill/bad luck/mistakes being made, but not attempting a repair would have guaranteed a total loss. A repair costs time, but it can help to save the environment and teach you a lot about what powers the products you use. Of course, most of the complete losses aren’t worth posting about – so you don’t see them.

The newer BladeRF boards feature full size USB B connectors which are more robust and probably would not have failed under the same strain. Buying a new BladeRF x40 is an option, but it’s a pricey one that as a hobbyist, I can’t easily justify.

About lui_gough

I’m a bit of a nut for electronics, computing, photography, radio, satellite and other technical hobbies. Click for more about me!

This entry was posted in Computing, Electronics, Radio and tagged , , , , , , , . Bookmark the permalink.

Error: Comment is Missing!