Dec 202020

So, some years back, my father purchased a Hema HX-1 GPS navigator to replace an older Hema unit… and for the past few years, the unit itself has performed just fine.

Then, about a month ago, he goes out to the car, and sees that the GPS has attempted a partial deconstruction. The LiPo cell inside had expanded, partially popping the case open. The unit was unresponsive to external power input, or really, any input.

Since the job of opening the thing was half-done, the least I could do is finish the job and open it up completely to survey the damage.

I didn’t take any photos of the GPS at this point. Apparently, this failure mode is very common with this model of GPS. First thought for me was getting that battery out! I started by disconnecting it from the PCB.

When disconnecting the battery, cut ONE WIRE AT A TIME! A LiPo cell, even a damaged one like this one, can easily generate enough current to spot-weld a set of side-cutters shut. You can also trigger overheating of the already damaged cell leading to possible explosion of the cell.

I cut the wires close to the cell so I’d have some wire to work with should I need it. I didn’t like my chances of accidentally shorting something in the process, so I figured by leaving the leads out of the dud cell as short as possible, I’m less likely to have them short each-other. The battery is a fairly standard 3.7V 5Ah single-cell LiPo battery with built-in thermister.

Next step was to remove the battery. This is tricky because the battery is stuck-down with double-sided tape, and the case makes accessing the underside difficult. You do NOT want to bend the cell, as that will cause fireworks too. Also, do NOT use metal implements to shift the battery, one puncture and there’ll be fireworks.

I found a semi-flexible plastic divider from a storage box worked: I just wedged it in under the cell, then moved it side-to-side to slowly “cut” the double-sided tape. You want something that is fairly flat, stiff, but with some “give” in it. Also, you’ll want some patience, this will be a slow process.

The old battery, removed from the GPS.

The code 2016ABAD didn’t escape my attention, maybe this cell’s credentials were dodgy from the start? Anyway. You can see how close I cut the output leads: I don’t want anyone ever using this cell again, and this is the easiest way to discourage it.

Next, I needed to see if the GPS was still alive. So, I stripped the ends of the old battery leads, spliced on a bit of extra cable, and hooked it to my bench supply set at 3.6V. A little current, and the GPS eventually woke up, moaning bitterly about a “low” battery, but the thing is, it worked, so it was worth proceeding.

For the replacement battery, I measured up the old one and found this 6Ah 3.7V cell to be a good match: it was almost the exact same size, featured a thermister the same as the old one, and was about the same capacity. If your cell lacks a thermister, you might have to bodge a 10k NTC in somehow. The pin-out of this particular cell though, pretty much exactly matched the pads on the PCB:

Existing battery connections.

Now, I could just take any suitable cell, cut the connector off (again, one wire at a time!) and solder it to the PCB, but I didn’t like my chances of accidentally bridging connections.

The cell I was buying had a JST-style connector, so I figured I’d ask about where I could get a mating socket. Turns out, Core Electronics have those too. The idea was I was going to mount this to the pads, then simply plug the cell in. I found I could simply bend the pins on this connector so that instead of exiting at a right-angle, they continued straight, making it very easy to simply “tack”-solder the connector to the old pads.

I just needed to ensure I got the connector around the right way!

Installing a battery connector.

Here’s a close-up without the annotations.

Close up of installed connector.

The install looked clean, so I tried firing up. No dice, it didn’t respond to the power button. Okay, maybe the battery is a little low, let’s try applying power. The GPS went into a boot-loop, seemingly unable to complete a full boot sequence. I left it that way for a few hours, but it didn’t improve.

I measured the cell voltage at about 3.8V… I figured maybe the charging circuitry just needed a helping hand. I had bought a LiPo charger a few weeks before, a clearance item, but not used it until now. I made up a pig-tail to mate with the cell’s connector.

LiPo charger with pigtail

I then unplugged the battery from the GPS, and plugged it into the charger. The POWER LED illuminated the moment the battery was connected, and once I fished around for a 5V 1A PSU (had a 2A one in the junk box), plugged that in, the STATUS LED illuminated. Reading the data sheet, that meant it was charging.

I left it run for a few hours then checked again, the STATUS LED had gone dark, indicating charging was complete. So I powered off the charger, unplugged the battery and put it back in the GPS. Held down the power button, and… SUCCESS! It booted straight up, and reported about 80% battery capacity (okay, whatever… at least it turned on).

Now, all was not completely rosy… in the opening and shutting of the GPS, I managed to tear off the coax from an antenna. I wasn’t sure whether it was for GPS reception or WiFi at this point, the antenna is a stick-on passive type, which is unusual for GPS, but a likely candidate for WiFi.

At this point, I had no GPS fix, so I did have concerns about that. I left the GPS for a bit, and over time, it eventually acquired a satellite fix, which might just be down to the GPS being powered down for over a month.

The old antenna and its broken coax feed.

Either way, I should try fixing that before I close it up. At first I was going to try soldering to the antenna, but that proved impossible, so I peeled off the old one, and set about replacing it outright. GPS reception didn’t seem too bad, but I did notice WiFi was down to about 1 bar of reception.

The coax is some of the tiniest cable I’ve seen, so I decided I’d just replace that too. The case design does not make accessing the solder pads easy, and I’m probably in need of a new tip on my soldering iron. Removing the old coax was easy enough… soldering to those pads was a pain.

My thought was to make a simple “dipole” antenna with some RG-195 coax, stripping off its jacket to reduce the outer diameter, putting a small length of heat-shrink to insulate part of the braid, then folding the braid back over the heat-shrink-insulated section to form a “sleeve” balun. I had thought I’d just tack the coax to the PCB, but this proved a nightmare, so I wound up using copper enamel wire to bridge from the pads to my antenna.

New antenna, connected to PCB using copper enamel wire

This is not ideal, as the wire will be inductive, but thankfully it’s short, so I’ll get away with it. You’ll notice there’s a small plastic “poker” that presses the power button on the PCB: I had to cut away a little bit of plastic there to allow the coax to pass underneath. I then stuck the other end of the antenna down with electrical tape. VSWR at 2.4GHz is probably horrible, but it seems to work anyway:

WiFi, seemingly working okay with the homebrew antenna

At this point, the battery was in, the antenna was fixed, and it was time to button it all up. I used some electrical tape to help keep the battery in place, figuring that with the tight tolerances in there it likely wouldn’t move far anyway.

GPS took a little while to “wake up”, having been disconnected from power for nearly a month, but eventually it picked up the satellites and had a reasonably accurate fix. given I was indoors I’m happy with that.

GPS status screen, showing a ~5m accuracy on the fix from 7 satellites
OziExplorer (Android) showing my current position.