Repair.IT – Ikea Striberg Led Lighting strip – Part 2

I mentioned a few days ago, that I received a shipment notification that my infrared emitters were in Australia and on the way to me. They did finally arrive on Friday. This fine Sunday morning, it was time to install the replacement infrared emitter. I did that, and was delighted that the Ikea Striberg Led lighting strip was now working very well in the automatic mode. I can put my hand about 10cm in front, and the light would go out – fantastic!

This meant that my diagnosis was correct and that the original infrared emitter had gone dim – I based this diagnosis on the measurement of the voltage drop on the original emitter. This voltage was higher than it should be, and indicated an internal high resistance fault. The replacement infrared emitter is a type OFL-3102 made by Multicomp Pro.

This is the link to the device –

The minimum purchase quantity was 5 units, so I bought 10 units, which meant that individually, the infrared emitter cost 20 cents each including gst. I now have 9 spares to go into my parts collection. If I bought these on their own, I would have had to pay $15 for delivery. However, I also bought some other things at the same time, so once I went over $50 for the order, I got free delivery. I often have other bits and pieces that I want to buy from Element14, so no big deal.

Anyway – that is the end of this story, or is it? While waiting for the infrared emitters to ship from the UK, where most of the stock was being held, I also decided to reverse engineer this lighting strip. I looked at each side of the board that the components are mounted on, and drew out the components as a schematic diagram. This is hand-drawn for the time being, as I wanted to understand how it works.

Most of the components are easily recognizable – a few transistors marked with a code that I needed to look up. Essentially, the 12V input connects directly to the leds via some limiting resistors. There are 3 leds in series with 3 parallel resistors – then 4 lots of these in parallel. 12 leds in total. The limiting resistors have markings 680, 680 and 750 – which translates to 68, 68 & 75 ohms each, then in parallel works out to be 23.4ohms in total. I measured the voltage drop across those resistors when the led lights were on, and got 2.5V, so each string of leds are passing about 106mA – so effectively 0.3W leds are being used.

When the switch is set to ON, then the negative or ground lead is then connected to the LED -ve, so all the leds light up. When the switch is set to Auto, then the ground lead goes to the ground that I have drawn, which means that the other electronics are enabled. The 12V input is converted to 5V through the use of the 78L05 voltage regulator – marked N1 at top right. The infrared emitter is driven from 5V through a 100ohm resistor, then switched with a N-channel MostFet marked A2SHB. Looking up A2SHB tells me that it is a GMS2302AL made by Dongguan Yushin Electronics. In the Auto mode, then the leds are switched on using another MosFet.

The rest of the circuit, shows the IR phototransistor being amplified by 3 transistors, then fed to the integrated circuit N2 on pin 7. Anyway, the integrated circuit N2 is marked 24. I don’t know what this is, but would guess that it functions as a proximity detector. Fortunately in my case, N2 is working fine – otherwise it might be a job to find a replacement. I noted what the pin functions for N2 are. The Reset pin 4 is determined because it has a capacitor which is then charged by a resistor. Pin 2 I guess would be an enable or some other purpose, but just has a pull-up resistor on it.

Essentially the function of N2 is – to pulse the IR Led, looks for a return pulse via IR Sensor, and if no return pulse, will activate LED Drv. In english this means if no pulse received, turn on the light – but keep sending pulses and checking for the return. Once pulses are being returned, turn off the light until it doesn’t detect return pulses.

That’s it for now – what’s for lunch!

Repair.IT – APC Back-UPS 650VA UPS Model BK650MI – Part 2

This is an update on a previous post –

Today, in the mail, I received a small packet – which felt like the relay I ordered for this UPS, seven weeks ago. Once I opened the packet, sure enough it was the relay, and I got very excited as it would mean that I could finish the repair of the UPS. Alas, it was not to be. Once I inspected the relay, my excitement dissipated and was left with a mild disappointment.

The relay that was supposed to be brand new (I checked the eBay listing) appeared to have been removed from existing equipment. The solder pins still had solder left on it, and the casing was discoloured and had some fine scratches on it.

In contrast, I then took another photo to compare it with the relay that I wanted to replace.

My faulty relay does look almost new – doesn’t it? Anyway, I did the usual “contact the seller”, mentioned that what I received wasn’t new and added a couple of photos showing the condition and that there was solder on the pins.

Since the supplied relay is not new, I cannot know what condition it is in, other than performing an extended test on it. For its purpose in switching the power to the outlets of the UPS, it is too critical to use a second-hand replacement unless there is no choice.

I did have a look at alternative relays and have found one that is a higher rated relay at 8 Amps instead of 6 Amps. The main problem is that the alternative is not a sealed relay – a factor that would be important, since the UPS will have a lead acid battery in it, which can release hydrogen during charging. A relay that is not sealed, means that hydrogen could potentially get inside the relay, and when the relay switches over, the slight spark could cause an explosion.

But then I saw a YouTube video of a repair of an APC Smart-Ups UPS and it appeared to contain a combination of sealed and non-sealed relays, so maybe I am over-thinking too much about the risks. I do try to replace with identical or higher rated parts if I possibly can.

In the meantime, all I can do at this stage for the UPS is to wait for the seller to respond. I got an email this afternoon that my order of the infrared emitters has been shipped. I should get that tomorrow hopefully and work on the led lighting strip.

Replace.IT – Coin battery in a Hantek DSO5062B Digital Storage Oscilloscope

I bought this Hantek DSO5062B Digital Storage Oscilloscope back in July 2013 as a replacement for my aging Hung Chang DS-635 35MHz Oscilloscope. The DSO5062B is a 60MHz oscilloscope and has a colour LCD display.

My DS-635 was bought in the early 1980’s and had served faithfully but had started to have intermittent faults. I meant to have a look at fixing it, but at the time, there was an offer on these Hantek’s, so it was a good opportunity to upgrade.

The digital storage has been very handy to look at waveforms and digital protocols, like serial output from unknown equipment. Just recently I got it out to help with the repair of the Ikea Striberg Led Lighting Strip. Once triggered, it would sample the waveform and then I can look and inspect various aspects of the waveform at my leisure. The old analog oscilloscopes could not do that – we would have to set and change trigger settings, delayed trigger etc – to view the waveforms for the detail that we wanted.

While using it for the Striberg, I noticed that it was not keeping the time. It is supposed to contain a real-time clock, but the manual didn’t mention anything about it. Eventually I found that the second page of the Utilities Menu allowed me to adjust the date and time.

It was then a matter of highlighting each field, press V0 then rotate to change the number, press V0 to set it, then rotate to the next field etc. After doing this, I left it for a short time, then turned off the DSO.

After turning on the DSO, the clock was reset again – looks like it works while powered on, but then forgets. The manual was checked again, and there was no mention of a clock battery. Sometimes these are powered by a NiCad battery which has probably gone flat, or maybe there is a coin battery – but the manual was of no help at all. A check of Google only showed replacing the lithium battery for portable Hantek oscilloscopes.

I then decided to open the case – two screws near the mounting feet, and two long screws under the carry handle. Once the case back is removed, this is what I see.

The power supply is the main part visible, then the main circuit board protected by a metal cover. If the oscilloscope had been powered up recently, the big capacitor on the right can still hold a lethal charge – even after a couple of minutes, so leave it for a while, if you are going to work on the power supply.

The metal cover has five screws, and once removed, I spied a coin battery in the corner. If I was to choose the most inconvenient place to put a removable coin battery, that would be the spot.

After some effort, I managed to pop the battery out – it was a CR2025 and measured 0V on my multimeter. I didn’t have any new CR2025’s in my spare battery box, so opted to borrow (steal) one from a camera remote control – one that I knew I wouldn’t need for a long time.

Once installed, I put the metal shield back on, then the case. Turning on, I set the date and time then let it sit for a while. 30 minutes later, I turned it on, and saw that the time and date were now correct. Hooray! Another repair done, by replacing an undocumented CR2025 lithium coin battery in the digital storage oscilloscope. Now to update my register of battery requirements. That’s it for now.

[P.S. Some of you might have seen the little sticker with red writing – that says “Remove seal after washing”. This covers the piezo speaker so that the solvent (used to wash the printed circuit board to remove flux and other contaminants after wave soldering) doesn’t get inside it and stop it from working. Actually I think it is better this way, as the beeps can get too loud, otherwise.]

Repair.IT – Ikea Striberg Led Lighting strip – Part 1

I mentioned while fixing the Ikea PAX wardrobe that my son had installed these Striberg Led Lighting strips that would turn on automatically once the doors were opened. After installation, a few years later – he noticed that the left lighting strip would stay on after the door was closed. As a workaround, he changed the switch to manual instead of automatic so he would need to turn it on whenever he wanted the light.

The Striberg is no longer available, but looks almost identical to the Norrfly which can be controlled by a wireless controller. This is what the Norrfly looks like – the picture is from Ikea’s website. The light is very handy especially at night and is supposed to turn on when the sliding door is opened. The left strip distance to the back of the door is about 5cm. The right strip distance to the back of the door is about 8cm – and that one works fine.

The faulty strip seems to detect around 2cm which is not good enough, so the lighting strips were removed from the wardrobe for examination. I removed the sensor cover and the light separator from the lighting strip in order to see what is happening.

The light separator is a piece of molded plastic that the led emitter and light sensor goes into, and forces it to view in front – and not light from the side. When operating, I cannot see any light from the led emitter, and expect that it would be an infrared led and the light sensor is also an infrared sensor.

Here is a closeup of the sensor arrangement, with the clear led emitter on the left and the red infrared detector on the right. I would guess that it works by turning on the led emitter, then detecting the light pulse through the detector. If the light pulse is seen within a certain time, then the door is closed. Now it seems that either the detector is not working properly or the emitter is not working properly.

I connected both lighting strips one on top of the early, with the working one at the top. Then using a digital camera and an infrared pass filter, I took a photo of the led emitters. The infrared pass filter will block all visible light and only allow infrared light within a range of frequencies to pass through to the camera. Most digital cameras have some sensitivity to infrared light.

Clearly, I can see that the working strip has a much brighter emitter – so that explains it!

How, I can hear you ask? Let me explain – the infrared reflectivity varies depending on the surface. Some materials reflect more infrared than others. The back of the sliding wardrobe doors have different surfaces. The left door is a mirror door and has a mottled plastic appearance. The door on the right is a wooden panel door, so has a wooden laminate surface on the back.

Since the non-working strip has a dim led emitter, it needs a closer reflector in order to sense the reflected light. To illustrate this, I tested it with an aluminium anti-static bag which has a matt highly reflective surface. The non-working strip was able to turn off at a distance of 8.5cm with this aluminium reflective surface, whereas a yellow post-it note would need to be 3cm away for the sensor to detect the returning light.

The general assumption in this case would be that if the sensor doesn’t see any light return, that the door is open – or in this case, the emitter is not sending enough light.

Light travels at about 299,705 km/s in air, so if I want to calculate the time it takes to say move 30cm (15cm to reflector and then return), it would be approximate 1nanosecond (1ns). I would guess that the sensor arrangement would be that it will send an infrared light pulse, and if it detects the returning pulse within 1ns, that the door is closed and turn off the led lighting strip.

Ok, so now I see (with the camera) that the non-working lighting strip is not sending much light, what would cause that? One thing might be a bad led, i.e. it has gone dim. Another possibility is that the driving voltage to the led emitter has dropped – how might I measure this? After some pondering, I can compare the working one with the non-working one.

I decided that I should get out my digital storage oscilloscope and measure the driving waveform across the led emitter – for both strips. One thing though is that my oscilloscope probe has a ground and if I ground one side of the led emitter, and the led power supply has a fault in it, I might trip my circuit breakers. The best way would be to run the lighting strip from a battery inverter, that way – it is isolated from the mains circuits. So that is what I did, connected the lighting strip power pack to my battery inverter.

Then connecting my DSO, I captured some waveforms – firstly the working strip.

I can see a pulse every 67ms or so – going to about 2.8V.

Reducing the horizontal timebase to 80us per division, I can see that the pulse is 118us in length. So this is now my reference. Let’s see what the non-working strip shows me.

The pulses are still about 67ms each time, but the pulse is about 5V.

The pulse length in this case is 123us – still similar enough, so why the different in voltage? Ok, maybe the led drive circuit is using a constant current drive – yes, that is more likely, because led’s like to be driven with a constant current.

So what does this tell me? I don’t know what current is passing through the led emitter at this point, but generally 5V across the led – doesn’t look good, since the forward voltage of an infrared led is usually around 1.6-2V or so. It may increase with high current, but usually high current means higher brightness and in this case, its brightness is low. Possibly all that I can conclude at this time is that the led might be faulty or has a high resistance internal fault.

If the led is faulty, then maybe I can either replace it or swap the led from the working strip. I checked my stock of parts, but unfortunately didn’t have any 3mm infrared leds. I know that just down the road, Jaycar in Rydalmere, has some in stock, but with the current CoVid stay at home restrictions in place – that isn’t essential travel.

As I had some other parts I wanted to get, I decided to order two types of infrared emitters from Element14 – which needs to be shipped from the UK, should take a week, so will update this in Part 2.

[P.S. It might be that all of those spikes might be related to my battery inverter which produces a modified sinewave output. Maybe when the infrared emitter parts come in, I will power the lighting strip from battery, then make some more measurements.]

Repair.IT – Ikea PAX wardrobe with the back coming apart

As mentioned in the previous post, not everything is about technology. This post is about a wardrobe, specifically, an Ikea PAX wardrobe that had been used by my younger son for many years. When he moved out a while ago, his bedroom sat unused and we had a look at the wardrobe with the view of making use of it, since it was now half empty.

The wardrobe comprised of two self standing sections that were connected together and with two sliding doors. The left wardrobe has the usual thin back that was joined in the middle with a glossy tape. With time and of course clothing pushing on it, the tape had decided to come apart, generally as the adhesive lost its grip.

The two back panels then had a gap where air could come into the wardrobe through the back – not an ideal situation as dust can also enter in that fashion. The wardrobe was 2.1m high, and together it measured 1.5m wide. It was very heavy, so was only able to be shifted after emptying and removing the drawers, shelved and doors – essentially anything that could be removed.

Once we were able to move it out, away from the wall – we could better assess the situation.

The tape had lifted from the backing board. I thought maybe I can stick it down, but as it had a waxed appearance – my duct tape wouldn’t stick to it. I remembered that I had some thin wooden slats that were left over from a few old blinds. These slats were used as a weight for the bottom of the blind and for the blind cord to attach to. I put these away since it might come in handy sometime.

It turns out that this would be quite suitable to brace the back of the wardrobe. So on Sunday, after fixing the wooden barrel planters, it was time to work on the wardrobe. I cut two pieces from those slats, each 1.5m in length – then pre-drilled some small holes for the attachment screws. I decided to attach the first brace, at 1.2m height on the back of the wardrobe, screwing into the side panels, then the next one at 1.8m height.

That seems better! Now to attach the backing board. I remembered that I had some T-nails that could be used with my staple gun. I should nail the backing board to the brace. As it turned out, that was not an easy task – the T-nails would bounce unless I had someone pushing on the brace – at the same time as I was pushing and then firing with the staple gun. My older son and I worked out a method whereby he would hold a large piece of wood and push onto the brace at the position that I wanted to staple.

Actually for safety reasons, it was better to do it this way – in case I missed the slat and the T-nail went straight through the backing boards. Of course I had measured the height and marked a line on the inside board so that I knew where to staple. In that way, I was able to staple T-nails along both braces. I left the other section alone since that didn’t have this problem. The next job was closing the gap. To do this, I had to cut the old tape off, and then put black duct tape down the gap, then add a few strips of duct tape sideways.

At least now, the gap was closed at the back. For the front, it might be best to use a white-coloured duct tape, so it wouldn’t be noticeable – but due to Covid lockdown restrictions, that will need to be left for another day. The wardrobe was moved back into place, then the drawers fitted – and then the doors refitted.

Oh, one more thing. The wardrobe had been fitted with Ikea Striberg led light strips, one for each section. The Striberg has an automatic mode whereby it detects that the door is closed and will turn off the led lights – except that one of them wasn’t working properly. The Striberg light strips were removed while the repair was in progress – so that will likely be the subject to another future blog post.

Repair.IT – Vintage wooden barrel planter

On a cold Winter’s day, during the Covid lockdown in Sydney – what can we do to pass the time, considering that during a lockdown, only essential travel away from home is permitted. This post is not about electronics, or computers, but is about fixing things to get some more use out of them.

Some time ago, pre-lockdown, we had visited a family friend. My wife and I had gone to help out with some pruning of trees in her back yard. She had a couple of wooden barrel planters lying around, left to the elements and thought, that perhaps, we might get some better use out of these, so had given them to us.

These barrel planters were of unknown make – appeared to have been hand made to some extent, and was also quite old. There were steel straps around the barrel planter, and the screws would need some tightening.

The planters were of a good size, with an internal diameter of near 50cm. Three castor wheels had been installed on the bottom of each planter to allow it to roll on the ground – but I found that these were loose, and had made a mental note to do something about it.

On Sunday, it seemed like just the thing for a cold morning. I had a piece of wood that I decided would be suitable and cut several pieces of it to help support the castor wheels.

The wooden blocks were first fastened from the sides with two screws each. I removed each castor, and sprayed some silicone lubricant into the bearings which were a bit rusted – the lubricant should protect the bearings. Then the castors were fastened with four screws instead of the original two screws.

After that was done, I put it aside as I didn’t have enough wood to do the same with the other planter. It already had five holes drilled in the base to allow water to leak out, so would be a good planter to use. While I was at it, I tightened the screws along each steel strap. Now on to the next job!

Recharge.IT – AEG L1825R 2.5Ah 45 Wh Li-Ion Battery Pack

This is a battery pack from my AEG Brushless Combo Kit.

The Combo kit came with a hammer drill/driver, an impact driver and two battery packs with a charger.  I used this AEG kit when I was building my timber deck back in 2016.  Since then, I had been using the drill/driver sporadically.  Recently when I went to use it, found that the drill would not run so put the battery pack to charge while I used the other battery pack.  After it was fully charged, I found that pressing the test button shows that the battery pack is fully charged, but it still didn’t work on the drill.  I checked the AEG website for my warranty, but found that the battery packs were now out of warranty.

I searched the local Bunnings website for replacements.  These AEG batteries are not cheap, and I could not find the 2.5Ah version.  I did find the 2.0Ah 18V battery pack for $89 or the 3.0Ah 18V battery pack for $129.  I put it aside until I had a chance to do something with it since I still have one working battery pack.

Whilst I was working on the Segway battery pack, I thought that I should check the batteries inside the battery pack.

The battery pack is held together with eight security Torx screws.  Fortunately I have the proper T-10H screwdriver bit so could remove the screws with ease.  Four screws held the top cover.

Another four screws held the sides – these sides need to be removed to expose the tabs on the battery cells inside the battery pack.

To my surprise, both sets of screws were the same length – this is not often the case, so means that I don’t need to keep each set of screws separated.

I used my multimeter to check each cell.  The first one I measured was 2.56V – which wasn’t right.  The remainder all measured 4.1V each, so it appears that this one cell is the problem.  As my Swallow Advance charger was still on the bench, I can use it to charge the bad cell.

I don’t need to kick start the cell, since there is some charge in the cell.

After it completed charging, there wasn’t much capacity delivered. I ran another charge cycle and it topped up a little, but the cell voltage drops from 4.1V down to 3.9V – so it appears that the cell is not retaining its charge capacity.

Anyway, I reassembled the battery pack and put it on my drill and was happy to confirm that the drill now worked with this battery, when it failed to do so the other day. Then I put the battery pack into the AEG charger to let it charge. Again, time will tell if this particular lithium-ion cell will remain usable.

Oh, one more thing – I found on eBay that there are compatible 4Ah battery packs available – a lot cheaper at $36.99 – and appear to be for Ridgid branded drills, and works with AEG. That is always an alternative since replacing a single cell would require dismantling of the battery pack and spot welding the contacts again – I don’t have a battery spot welder, maybe I should get one.

Repair.IT – Lithium Battery Pack from a Segway Ninebot One S2 Unicyle

My friend Thomas owns this Segway Ninebot One S2 Unicycle.  The Ninebot One S2 is powered by two battery packs, and he got a warning from his Segway phone app that the batteries were “severely disparate”. Further troubleshooting found that one battery pack seemed to be low in charge and just wouldn’t charge.  Replacements for these battery packs were not available in Australia, so he was looking for other solutions. He then remembered that I do all sorts of repairs and contacted me about it. I said that I would be very happy to have a look at it.

He brought it to me a while ago, and I had a quick look at it. When the ND1501-B battery pack is removed from the unicycle, there will usually be a small green led that is lit to show that the battery pack is working. The left side battery pack seemed to be ok, but the right side battery pack was bad – when viewed with the power button facing forwards.

I couldn’t make a diagnosis until the battery pack could be opened up.  There should be an electronic circuit board inside the battery pack that would monitor the status of the batteries and if anything goes wrong, it would disconnect the output to protect the battery pack from further damage.  The battery pack comprises of two halves that clip together, and unfortunately – it was very difficult to open without causing some damage to the casing.

Even a Google search did not come up with information on how to open the battery casing – I did find a YouTube video of one that had been opened – but the only instructions were to pry it apart. After some thought, I chose to put the battery pack into my milling machine, and carefully mill off a section of the case.

Getting ready to mill a hole near one of the tabs

This would be where one of the clips would be located.  That can be my entry point to determine how the clips work and how to open the casing.  It sounded easy, and I could see how the clip engaged, but just could not push the casing apart at that point. I tried a number of times, and would put it aside, then come back to look at it again. I really didn’t want to damage the casing any more than I already had.

Anyway, I eventually worked out that I had to lever the area where the clip was located outwards, and then the case started to separate. Then I used another tool to pry the sides apart bit by bit as the clips began to disengage.  There appeared to be some sort of glue used to join the sides together, so prying the case halves, caused the glue seal to break and then the sides eventually came apart.

Now that the battery pack was open, I could look at the battery layout and how it works.

The battery pack consists of 15 lithium cells.  The part number is LGEBMG11865.  When I look that up, I find that it is a Lithium-Ion cell, 3.62V and 2850mAh – this will be important for when I do the cell charging.  The 15 cells are joined in a series arrangement with 10 cells in one group, and 5 cells in the group above.

Coincidentally, it seems that those 5 cells were all showing little or no voltage, whereas the other cells of the group of 10 were on average 4.1V each which indicate a fully charged state.  From this, I would surmise that perhaps that group of 5 cells at some point, had less capacity than the rest, so when the battery pack was being used to its capacity, those cells went completely flat before the protection circuitry could prevent this.  Once that had happened, the battery pack would no longer charge.

I have a Swallow Advance Digital charger which could handle 1-5 cells, but this will only work if the cells still had some charge. The only way to find out if the lithium cells are still workable is to charge each cell individually.  To do this, I need to kick start the cell first using another charger before I can use the Swallow charger.

I have a TP4056 charger that I had previously used to fix a completely flat Samsung phone battery.  This TP4056 charger is designed to charge a single cell with a specific current until it reaches 4.1V. I connected this TP4056 charger to my multimeter so that I can monitor the voltage, then used the multimeter probes to contact the first cell.  I could see the cell voltage slowly rise from zero to 1V – this was promising.  After disconnecting the TP4056, I could see that the cell was still holding some voltage.

Ok – great.  I then connected the TP4056 again, and charged the cell to about 1.5V. This kick start process takes short period of time – usually around 15-30 seconds, so I had to hold the probes on the lithium cell until I got to that voltage. I know if the Swallow charger can detect some voltage on the lithium cell above 1V that it would be able to charge the cell.

At this stage, I then connected the Swallow and it was able to commence charging.  The way the Swallow charger works is that it first checks the cell voltage and if it is below a certain threshold, it stops and gives a low voltage error.  This time however, it could measure some voltage and began to charge at a constant current of 0.1C.

Remember earlier, I noted that the cell capacity was 2850mAh – this is what I set in the Swallow charger configuration.  0.1C is one-tenth of the capacity and so it charges initially at 280mA or so, until the cell voltage rises to near 3.6V – at which time, it switches to CCCV mode, i.e. constant current at 1C and constant voltage.

I could see that once the charger was running in CCCV mode, it increased the maximum charge current to 2.85A, which then slowly decreases while approaching a target voltage of 4.1 Volts.  The target voltage will differ depending on the cell type – this one is Lithium-Ion, so it will have a nominal voltage of 3.6V and end up at 4.1V.  Lithium-Ion Polymer would be different – with nominal at 3.7V and full charge at 4.2V.

It is important to use the correct charging method for the battery type – and since the Swallow Advance is an automatic charger, I just need to input that the battery type is Lithium-Ion, capacity is 2850mAh and nominal voltage is 3.6V or 1 cell since I am charging a single cell at a time.  When the charge cycle has completed, the charger will give a short series of beeps and then display a status screen.

This shows that it took 221 minutes and 17 seconds, and delivered approximately 2353mAh capacity to the battery.  Ok, on to the next cell – kick start the charging first, then connect the Swallow charger, and then the next, etc until all five are done. For the remaining three cells, I decided to connect the cells to the Swallow charger, then using the TP4056 would kick start the cell by connecting the probes to the Swallow charger output. Then when the cell voltage hit 1.5V, remove the probes and quickly press and hold the Enter button to start the Swallow charger.

My Swallow charger could have charged those 5 cells together at once, why didn’t I do this, you might ask?  Yes, I could have kick started each cell individually, then used the Swallow charger to charger all five cells as once. But this might overcharge some cells and undercharge others as the charger doesn’t have a charge balancer. Since I didn’t know whether each cell was working, I might have caused damage to any good cells.  By charging the cells individually, I can at least determine how much capacity I delivered to each cell using the Swallow charger.

The table shows each battery, how long it took in minutes and seconds, then the delivered charge capacity. As each cell was fairly consistent at around 2350mAh – I thought this was ok for cells that were a year or two old, as it is normal for cells to lose a bit of capacity over time. And yes, the individual cell charging was taking 3-4 hours each – I didn’t have to sit and watch it of course – once the charge is completed, the charger will give a series of beeps and then sits monitoring the cell voltage until I disconnect the cell.

I checked the overall voltage of the cells and got 60.2V – which was satisfactory. On the output though, I could measure only 5.5V – but I think that was because the battery pack is in safe mode. After I had checked that everything was intact, no stray wires sitting on the battery pack, charger disconnected completely – I pressed the S1 switch.

The switch press caused the green led to start blinking – which is a very good sign. Just think of the S1 switch as being the start button for the microcontroller (the black square chip with lots of leads) to turn on. Once it is on, the battery pack will continue running until the battery pack either fails or goes completely flat. Ideally, if the Ninebot is going to be unused for a while – then it should be placed on charge once a month or so.

As to why it failed in the first place – since it was a group of five cells, I suspect that maybe those five cells had a lower capacity in the first place – or may not have been fully charged when the battery pack was manufactured. Then over time, those cells eventually went completely flat – causing the Segway app to generate a warning about the batteries. It isn’t uncommon for cells to go completely flat, so whatever the reason – these cells seem to be ok – were able to charge, so time will tell if this repair is going to last.

If it happens again, I can always charge the cells again – now that I know how to open the battery pack. Other actions that I could take – would be to run a number of discharge/charge cycles. This is usually done to determine the actual capacity of the cells, and can sometimes help to rejuvenate failing cells. These cells are most likely failing, but at least they are working, so until it fails again – best to get some use out of it. Using the Ninebot One will help to discharge the cells, then putting it back on charge, will help to exercise the cells anyway.

I reassembled the battery pack, put duct tape over the hole that I had made and installed it back into the Ninebot One S2. Then the rubber seals installed and the covers fastened into place. I found and installed the Segway-Ninebot app on my phone – then powered on the Ninebot One, and connected to it via Bluetooth. Once I registered my app, I was able to see that the battery packs were at 100% and 99% – which I thought was fantastic.

All of this work had been done over several days – not consecutively, and summarises the main points of this particular repair. Today is a Sunday as I write this article which was started a couple of days ago – maybe after my badminton this afternoon, I could try going for a little ride! Or at least attempt to – since this would be my very first experience on an electric unicycle.

[P.S. The white connector in that picture above connects to each battery individually, so when charging batteries 12 onwards, I inserted small jumper wires to contact the pins, i.e. white and black wire, then black and blue, blue and yellow and finally yellow and red. The jumper wires then go to my Swallow charger – for battery 12, negative would be the white, and positive would be the black. Then the next battery, negative would be black, and positive would be blue – and so on. For the low charge currents that I am using, those wires can easily carry the charging load.]

[P.P.S – This is the third lithium battery article – you remember, all good things come in three’s! Actually, there is another lithium battery article coming up next – so maybe it is four’s!]

Remove.IT – Very swollen lithium battery from a Sound Blaster E5

It seems that once I get started on lithium batteries, it doesn’t seem to end. This post is about a Sound Blaster E5.

The Sound Blaster E5 is a 24-bit/192kHz high resolution USB DAC and portable headphone amplifier and contains a lithium battery. The E5 belongs to my son, who had been using it for many years. My son didn’t use it with its portable capability, but was permanently plugged into a USB port on his computer.

A while ago, he noticed that the metal name plate didn’t seem to be attached properly. Little did he know that this was a sign of a swollen battery – and from the Google searches on this topic, it seems likely that almost every E5 will suffer this fate sooner or later. Creative had tried to fix the problem by offering a firmware patch, but if the battery is very swollen, that patch will not help. My son had taken it apart by following instructions found online. The battery was very swollen, and there were wires everywhere, so due to the potential fire hazard, it was put aside.

The battery in this case is a lithium-ion polymer battery. Lithium-ion batteries usually contain electrode sheets that are wrapped around like a Swiss roll. When the battery is overheated, or overcharged or even left discharged for long periods of time, it can swell or pillow. Gases build up inside, and if the protective envelope is punctured, it can cause an explosion or fire.

It was at this stage that I was involved. After some careful examination, I could see that the battery was a bit loose on one side, so it was most likely held in by adhesive tapes. If I could separate the battery from the adhesive, I might have a chance to detach it, and bring it outside of the case. Nothing of course is as easy as it sounds, and due to the very swollen nature of the battery pack, it was essential not to damage it by using sharp utensils. Ideally, I would be wearing a full face mask to protect myself from fire, and wearing fireproof gloves – but I don’t have those handy.

Yesterday, it was time to look at this delayed job. It had been sitting there for a month or so, and as long as I was careful – I should be able to remove the battery. I made a couple of soft plastic spatulas from an old store loyalty card. I could have used an old credit card but those are usually cut up and disposed of. The edges of the spatulas were rounded so they won’t have a sharp corner, and although it helped, the spatula would stick to the adhesive. I then got a worn wooden clothes peg, the type that was held together by a spring. By separating the peg, I was able to use the pieces to prod the tape until the battery was loose.

The clothes peg pieces were much stiffer, so that made it easier to do the job. Once the battery was out of the case, I could then cut the wires off the battery, as I would need to use the wires eventually – if I want to restore this device back into a working condition.

For the moment though, I was concerned with removing the fire hazard. Once the wires were cut, I can then have a closer look.

A lot of “Do Not’s”

From the battery model number, I can see that it is a 535058 1S2P. This means that originally, the battery is 5.3mm thick, with a width of 50mm and length of 58mm. It is also 2 battery cells in parallel, which you can see in the next photo.

It has pillowed out to be just over 20mm, which caused the case to balloon out and hence why the metal nameplate was loose. I will store the battery in a tin box for the time being. If this Sound Blaster E5 had been sitting under a desk, out of sight, it could eventually have exploded or started a fire, so we dodged this bullet this time around. If we have battery operated devices, we should check them from time to time. If they use lithium batteries or even NiCd and NiMH, it makes sense to charge them, from time to time.

I just received another UPS battery delivery today, since my main UPS started beeping recently. My eBay purchase history shows that the batteries I just ordered, had also been bought back in 2018 – so a life of 3 years in a UPS is quite reasonable, but I would have preferred it to be longer. I ordered the two 12V 18Ah lead acid batteries yesterday and they got delivered today from Melbourne. I will have to schedule the battery replacement, since the main UPS powered up my internet and internal network – maybe a Sunday morning would be good.

Anyway, that’s it for now – another lithium battery article. Good things come in three’s – next article to come will be about the lithium battery pack from a Segway Ninebot One S2 Unicycle.

Replace.IT – Swollen battery for Xiaomi Redmi Note 4

This story begins about two weeks ago. Do you know that often – one thing leads to another? This story is a bit like that. I noticed that my phone, the Xiaomi Mi A3 had some very fine scratches on the screen. When I bought that phone, I had also ordered a glass screen protector which arrived in many pieces, so that got refunded and I didn’t get around to getting another screen protector. My wife’s phone was the Xiaomi Mi A1 and it had a plastic screen protector that was very worn, so had patches on it where the top coating was missing.

I ordered glass screen protectors for both phones from eBay. Eventually they arrived a few days later, so last week – I was putting the glass screen protectors on the phones. After doing this, I then had a look at my old phone, which was a Xiaomi Redmi Note 4 – for which I had a spare plastic screen protector. I thought while I had the cleaning materials out, I should swap the screen protector for this phone at the same time.

After removing the Note 4 from its case, I noticed that the screen was bulging out. Then I noticed that the back was bulging too! I remember that for a while now, this Note 4 was getting quite hot when being used, and now I realize that it must be due to the failing battery. This battery had been replaced in October 2018 – just about three and a half years ago. I still had all the tools to remove the back cover, so decided to do just that.

There are two tiny screws on the bottom where the micro usb (charging) port is located – it needed a small pentalobe screwdriver to remove the screws. Then it was a matter of using the plastic prying piece, which looks a bit like a guitar pick. I finally remembered that it needs to be pried where the metal back cover is contacting the plastic around the screen. The best part to do this was where the charging port is located, so eventually removed it. I saw that the battery was swollen – puffed up a bit like a balloon, but not that big, of course.

The battery is hermetically sealed, or at least – it should be. When it begins to fail, it can generate more gases than can be recombined, so these excess gases lead to the bulging of the battery. If it wasn’t airtight, exposure to water vapour in the air due to humidity can cause catastrophic failure, i.e. you hear stories of phones burning – yeap. This can happen also if the gases continue to increase until the plastic around the battery breaks (pops) – then yeah, burning can happen, since lithium is a highly reactive metal.

A quick check on eBay shows that the BN41 battery is still available, so was able to order it a couple of days ago. It should arrive sometime next week. In the meantime – what do I do with this swollen battery, and come to think of it, the batteries that had been swapped from the UPS’s, and all the other batteries. The lithium battery should be stored in a fireproof bag, but I don’t have one handy, so might be best to be placed in a tin box.

Remember those old alkaline batteries, AA, AAA and 9V ones! These are all meant to be recycled, and not thrown into the garbage. I had been putting the old batteries into plastic bags and stored in the garage. Also a lot of rechargeable batteries, NiCd, NiMH – were similarly stored. Old battery packs from broken cordless tools were also stored in the garage. I did mention that one thing leads to another – I started by replacing screen protectors, then needed to get a replacement battery for a phone, and now need to dispose of all of the old batteries. Come to think of it, I still have the old car battery that came out of the Honda Civic recently when it failed and needed a replacement.

I looked on the local government website for battery recycling and eventually found a place nearby – Battery World at Northmead North Parramatta. I sent them an email asking about whether I can drop off these old batteries and they replied back that I could do this any time. So on Wednesday, I will most likely bundle all of these batteries, including any others that I find and drop them off to be recycled.

I realize that not everyone sends batteries to be recycled, and often the alkaline batteries are tossed into the garbage along with everything else that is discarded. In Australia, about 98% of lead-acid batteries are recycled. Most of these are of course, car and truck batteries – usually when being replaced, the servicing guy will take the old batteries away to be recycled. If you buy certain brands of UPS batteries, they used to come with a freight label to send the old battery back – I don’t know if they still do this.

Lithium-ion and associated batteries though – only about 2% of them get recycled here in Australia. The CSIRO says that Australia produces 3,300 tonnes of lithium-ion battery waste a year, and this increases by 20% each year, and may hit 100,000 tonnes by year 2036. Lithium-ion batteries are used almost everywhere now – laptops, phones, portable electronics, electric vehicles and now even some UPS’s are using them instead of lead-acid ones.

So let’s do our bit for the environment – if you have old batteries lying around, check with your local council or local government on what to do with them. Now I just need to wait for the phone battery to arrive, and I can complete my repair.