All the site migration hassle now out of the way, the blog should have returned to it’s usual speed & responsiveness by now 🙂
Here’s a couple of photos from my current project, a full write up to come as things go together:
Since the 4×18650 battery pack supplied with my Cree head torch is pretty shit, even by China’s standards, I figured something I could put my own cells into would be a better option. An eBay search turned up these battery boxes, not only with a direct battery output for my torch, but also a USB port for charging other devices when I’m low on charge.
The output to the lamp connector is directly connected to the battery, through the usual Lithium Ion protection, but the USB output is controlled from a single power button. Battery charge condition is displayed on 3 LEDs. Not sure why they used blue silicone for the seal & then used green LEDs… But it does work, even if a little dim.
Essential information. Does claim to be protected, and from the already existing electronics for the USB this would be expected in all but the cheapest crap.
An IP rating of IPX4 is claimed, yet just above that rating is a notice not to be used in water. Eh?
This is sealed with an O-Ring around the edge of the top cap & silicone seals around the cable & retaining screw. I did test by immersion in about 6″ of water, and it survived this test perfectly fine, no water ingress at all.
The casing holds a PCB at the bottom end with the cell straps.
Someone wasn’t that careful at getting the brass screw insert properly centred in the injection mould when they did this one. It’s mushed off centre, but i’s solidly embedded & doesn’t present any problems to usability.
The top cover holds the cell springs & the electronics.
Removing the pair of screws allows the top cap to open up. The cable, button & LEDs are robustly sealed off with this silicone moulding.
Here’s the PCB, not much on the top, other than the power button & battery indicator LEDs.
Desoldering the cell springs allows the PCB to pop out of the plastic moulding. There’s more than I expected here!
Bottom left is a DC-DC converter, generating the +5v rail for the USB port, this is driven with an XL1583 3A buck converter IC.
Bottom right is the protection IC & MOSFETs for the Lithium Ion cells. I wasn’t able to find a datasheet for the tiny VA7022 IC, but I did manage to make certain it was a 7.4v Li-Ion protection IC.
Top right is a completely unmarked IC, and a 3.3v SOT-23 voltage regulator. I’m assuming that the unmarked IC is a microcontroller of some sort, as it’s handling more than just the battery level LEDs.
A pretty decent 4-core cable finishes the job off. For once there’s actually some copper in this cable, not the usual Chineseuim thin-as-hair crap.
Here’s another Dyson teardown, in my efforts to understand how marketing have got hold of relatively simple technology & managed to charge extortionate amounts of money for it.
This is the DC35, the model after the introduction of the brushless digital motor.
On this version the mouldings have been changed, and the back cover comes off, after removing the battery retaining screw. It’s attached with some fairly vicious clips, so some force is required. Once the cap is removed, all the electronics are visible. On the left is the motor itself, with it’s control & drive PCB. There’s another PCB on the trigger, with even more electronics. The battery connector is on the right.
Here’s the trigger PCB, which appears to deal with DC-DC conversion for powering the brush attachments. The QFN IC with yellow paint on it is an Atmel ATTiny461 8-bit microcontroller. This is probably controlling the DC-DC & might also be doing some battery authentication.
Here’s the motor & it’s board. The windings on the stator are extremely heavy, which makes sense considering it’s rated at 200W. The main control IC is a PIC16F690 from Microchip. Instead of using an off the shelf controller, this no doubt contains software for generating the waveforms that drive the brushless motor. It also appears to communicate with the other PCBs for battery authentication.
Desoldering the board allows it to be removed from the motor itself. The pair of windings are connected in anti-phase, to create alternating North-South poles depending on polarity. Since the existing controller is unusable due to software authentication with the other parts, I might have a go at building my own driver circuit for this with an Arduino or similar.
The blower assembly is simple plastic mouldings, pressed together then solvent welded at the seam.
The impeller is just a centrifugal compressor wheel, identical to what’s used in engine turbochargers.
The inside face of the control PCB holds the 4 very large MOSFETs, IRFH7932PbF from International Rectifier. These are rated at 30v 20A a piece, and are probably wired in a H-Bridge. There’s a bipolar Hall switch to sense rotor position & rotation speed, and an enormous pair of capacitors on the main power bus.
Not much on the other side of the PCB other than the microcontroller and associated gate drive stuff for the FETs.
The battery pack is similar to the DC16 in it’s construction, a heavily clipped together plastic casing holding 6 lithium cells. In this one though there’s a full battery management system. The IC on the top of the board above is a quad Op-Amp, probably for measuring cell voltages.
The other side of the BMS board is packed with components. I wasn’t able to identify the QFN IC here, as it’s got a custom part number, but it’s most definitely communicating with the main motor MCU via I²C over the two small terminals on the battery connector.
The Dyson DC16 is one of the older handheld vacuums, before the introduction of the “Digital Motor”. (Marketing obviously didn’t think “Switched Reluctance Motor” sounded quite as good).
These vacuums have a very large DC brush motor driving the suction turbine instead, the same as would be found in a cordless power tool.
Popping the front cap off with the ID label, reveals the brains of the vacuum. The two large terminals at the right are for charging, which is only done at 550mA (0.5C). There are two PIC microcontrollers in here, along with a large choke, DC-DC converter for supplying the logic most likely. The larger of the MCUs, a PIC16HV785, is probably doing the soft-start PWM on the main motor, the smaller of the two, a PIC16F684 I’m sure is doing battery charging & power management. The motor has a PCB on it’s tail end, with a very large MOSFET, a pair of heavy leads connect directly from the battery connector to the motor.
Just out of sight on the bottom left edge of the board is a Hall Effect Sensor, this detects the presence of the filter by means of a small magnet, the vacuum will not start without a filter fitted.
The battery pack is a large custom job, obviously. 4 terminals mean there’s slightly more in here than just the cells.
Luckily, instead of ultrasonic or solvent welding the case, these Dyson batteries are just snapped together. Some mild attack with a pair of screwdrivers allows the end cap to be removed with minimal damage.
The cells were lightly hot-glued into the shell, but that can easily be solved with a drop of Isopropanol to dissolve the glue bond. The pack itself is made up of 6 Sony US18650VT High-Drain 18650 Li-Ion cells in series for 21.6v nominal. These are rated at a max of 20A discharge current, 10A charge current, and 1.3Ah capacity nominal.
There’s no intelligence in this battery pack, the extra pair of terminals are for a thermistor, so the PIC in the main body knows what temperature the pack is at – it certainly gets warm while in use due to the high current draw.
Hidden in the back side of the main body is the motor. Unfortunately I wasn’t able to get this out without doing some damage, as the wiring isn’t long enough to free the unit without some surgery.
The suction is generated by a smaller version of the centrifugal high-speed blowers used in full size vacuums. Not much to see here.
Since I got this without a charger, I had to improvise. The factory power supply is just a 28v power brick, all the charging logic is in the vacuum itself, so I didn’t have to worry about such nasties as over-charging. I have since fitted the battery pack with a standard Li-Po balance cable, so it can be used with my ProCell charger, which will charge the pack in 35 minutes, instead of the 3 hours of the original charger.
I recently came across these on eBay, so I thought I’d grab one to see how they function, with all the metrics they display, there’s potential here for them to be very useful indeed.
One of the best parts is that no wiring is required between the sensor board & the LCD head unit – everything is transmitted over a 2.4GHz data link using NRF24L01 modules.
Above is the display unit, with it’s colour LCD display. Many features are available on this, & they appear to be designed for battery powered systems.
Another PCB handles the current & voltage sensing, so this one can be mounted as close to the high current wiring as possible.
The transmitter PCB is controlled with an STM8S003F3 microcontroller from ST Microelectronics. This is a Flash based STM with 8KB of ROM, 1KB of RAM & 10-bit ADC. The NRF24L01 transceiver module is just to the left.
There’s only a single button on this board, for pairing both ends of the link.
The high current end of the board has the 0.0025Ω current shunt & the output switch MOSFET, a STP75NF75 75v 75A FET, also from ST Microelectronics. A separate power source can be provided for the logic via the blue terminal block instead of powering from the source being measured.
Here’s the display unit, only a pair of power terminals are provided, 5-24v wide-range input is catered for.
Unclipping the back of the board reveals the PCB, with another 2.4GHz NRF24L01 module, and a STM8S005K6 microcontroller in this case. The switching power supply that handles the wide input voltage is along the top edge of the board.
Unfortunately I didn’t get any instruction manual with this, so some guesswork & translation of the finest Chinglish was required to get my head round the way everything works. To make life a little easier for others that might have this issue, here’s a list of functions & how to make them work.
On the right edge of the board is the function list, a quick press of the OK button turns a function ON/OFF, while holding it allows the threshold to be set.
When the output is disabled by one of the protection functions, turning that function OFF will immediately enable the output again.
The UP/DOWN buttons obviously function to select the desired function with the cursor just to the left of the labels. Less obviously though, pressing the UP button while the very top function is selected will change the Amp-Hours display to a battery capacity icon, while pressing DOWN while the very bottom function is selected will change the Watts display to Hours.
The round circle to the right displays the status of a function. Green for OK/ON Grey for FAULT/OFF.
I’ve not yet had a proper play with all the protection functions, but a quick mess with the OVP setting proved it was very over-sensitive. Setting the protection voltage to 15v triggered the protection with the measured voltage between 12.5v-13.8v. More experimentation is required here I think, but as I plan to just use these for power monitoring, I’ll most likely leave all the advanced functions disabled.
This camera has now been retired after many years of heavy use. Exposure to a 3-year old has caused severe damage to the lens mechanism, which no longer functions correctly.
Pretty much standard interface for a digital camera, with a nice large LCD for it’s time.
With the front cover removed, the lens assembly & battery compartment is exposed.
Removing the rear cover exposes the LCD module & the main PCB, the interface tactile switches are on the right under a protective layer of Kapton tape.
Flipping the LCD out of it’s mounting bracket reveals the main camera chipset. The CPU is a NovaTek NT96432BG, no doubt a SoC of some kind, but I couldn’t find any information. Firmware & inbuilt storage is on a Hynix HY27US08561A 256MBit NAND Flash, with a Hynix HY5DU561622FTP-D43 256Mbit DRAM for system memory.
I couldn’t find any info on the other two chips on this side of the board, but one is probably a motor driver for the lens, while the other must be the front end for the CCD sensor input to the SoC.
The other side of the PCB handles the SD card slot & power management. All the required DC rails are provided for by a RT9917 7-Channel DC-DC converter from RichTek, an IC designed specifically for digital camera applications.
Top left above the SD card slot is the trigger circuitry for the Xenon flash tube & the RTC backup battery.
Once the main PCB is out of the frame, the back of the lens module with the CCD is accessible. Just to the left is the high-voltage photoflash capacitor, 110µF 330v. These can give quite the kick when charged! Luckily this camera has been off long enough for the charge to bleed off.
Finally, here’s the 7-Megapixel CCD sensor removed from the lens assembly, with it’s built in IR cut filter over the top. I couldn’t find any make or model numbers on this part, as the Aluminium mounting bracket behind is bonded to the back of the sensor with epoxy, blocking access to any part information.
Die images of the chipset to come once I get round to decapping them!
I’ve been working on something new for the blog, as I have a massive collection of scrap ICs – that is to decap the silicon & get them under a microscope!
Here’s the first image, a Texas Instruments part. This has been stitched together from 140 separate images to create the final version.
As the CO meter I bought on eBay didn’t register anything whatsoever, I decided I’d hack the sensor itself apart to make sure it wasn’t just an empty steel can. It turns out that it’s not just an empty can, but there are some reasons why the thing doesn’t work 😉
The cell was crimped together under the yellow shrinkwrap, but that’s nothing my aviation snips couldn’t take care of. The photo above shows the components from inside.
The endcap is just a steel pressing, nothing special here.
Also pretty standard is the inlet filter over the tiny hole in the next plate, even though it’s a lot more porous that I’ve seen before in other sensors.
Next up is the working electrode assembly, this also forms the seal on the can when it’s crimped, along with insulating it from the counter electrode & external can. The small disc third from left is supposed to be the electrode, which in these cells should be loaded with Platinum. Considering where else they’ve skimped in this unit, I’ll be very surprised if it’s anything except graphite.
Next up is the counter electrode, which is identical to the first, working electrode. Again I doubt there’s any precious metals in here.
Another steel backplate finishes off the cell itself, and keeps most of the liquid out, just making sure everything stays moist.
Finally, the rear of the cell holds the reservoir of liquid electrolyte. This is supposed to be Sulphuric Acid, but yet again they’ve skimped on the cost, and it’s just WATER.
It’s now not surprising that it wouldn’t give me any readings, this cell never would have worked correctly, if at all, without the correct electrolyte. These cheap alarms are dangerous, as people will trust it to alert them to high CO levels, when in fact it’s nothing more than a fancy flashing LED with an LCD display.
Ironically enough, when I connected a real electrochemical CO detector cell to the circuit from the alarm, it started working, detecting CO given off from a burning Butane lighter. It wouldn’t be calibrated, but it proves everything electronic is there & operational. It’s not surprising that the corner cut in this instance is on the sensor cell, as they contain precious metals & require careful manufacturing it’s where the cost lies with these alarms.
I did a little more digging into the PSU circuitry of the small coin counting machine, and it’s even more strange than I thought!
The part I originally thought was a transformer on the PSU board is in fact a DC-DC converter module!
Here’s the device after desoldering it from the PCB. It turns out that instead of a transformer, it’s an inductor.
Underneath is the controller electronics, with an COB controller & the switching transistors are under a protective covering of silicone.
Driving this whole lot of PSU randomness is the mains transformer, with a secondary voltage of 35v.
The only reason I can think of that the manufacturer went to this much expense with the power supply is stability – a coin counting machine that miscounts due to power supply surges, sags & spikes wouldn’t be very much use. It’s not likely I’ll see anything similar again, unless I manage to get hold of something like medical grade equipment.
Here’s some teardown photos of an old De La Rue coin counter, used in businesses for rapid counting of change into large bags.
An overview of the whole mechanical system of the counter. Coins are loaded into the drum at the rear of the machine, which sorts them into a row for the rubber belt to pick up & run through the counter. The coin type to be sorted is selected by turning the control knobs on the right.
The control knobs adjust the width & height of the coin channel so only the correct sized coins will be counted.
The counter is driven by a basic AC induction motor, the motor power relay & reversing relay is on this PCB, along with the 5v switching supply for the main CPU board.
The SMPS on this board looks like a standard mains unit, but it’s got one big difference. Under the frame next to the main motor is a relatively large transformer, with a 35v output. This AC is fed into the SMPS section of the PSU board to be converted to 5v DC for the logic.
I’m not sure why it’s been done this way, and have never seen anything similar before.
The edge of the coin channel can be seen here, the black star wheel rotates when a coin passes & registers the count.
Here’s the main controller PCB, IC date codes put the unit to about 1995. The main CPU is a NEC UPD8049HC 8-bit micro, no flash or EEPROM on this old CPU, simply mask ROM. Coin readout is done on the 4 7-segment LED displays. Not much to this counter, it’s both electronically & mechanically simple.
Coin counting is done by the star wheel mentioned above, which drives the interrupter disc on this photo-gate. The solenoid locks the counter shaft to prevent over or under counting when a set number of coins is to be counted.
Under the frame, here on the left is the small induction motor, only 6W, 4-pole. The run cap for the motor is in the centre, and the 35v transformer is just visible behind it.
Main drive to the coin sorting mech is through rubber belts, and bevel gears drive the coin drum.
Here’s an old HDSPA 3G USB modem stick that I got with a mobile phone contact many years ago. As it’s now very old tech, and I have a faster modem, not to mention that I’m no longer with Orange (Robbing <expletive>), here’s a teardown of the device!
The top shell is just clipped into place, while a pair of very small screws hold down the orange piece at left to hold the PCB stack in the casing. Not much to see here, but it’s clear that there’s a lot crammed into a very small space.
Here’s the PCB stack removed from the outer casing. The main antenna is on the right, attached with another small screw. Every IC on the boards is covered with an RF can. No problems there, pliers to the rescue!
Here’s the top PCB, all the shields have been removed. On the left is a Qualcomm PM6658 Power Management IC with integrated USB transceiver. This is surrounded by many of the power management circuits.
The integrated SD Card slot is on the right side. with what looks to be a local switching regulator for supply voltage. This might also provide the SIM card with it’s power supply.
The other side of the top board reveals more power management, with another switching regulator, and a truly massive capacitor at the top edge. I’m guessing this is a solid Tantalum.
The other PCB holds the main chipset & RF circuits. On the left here is a Samsung MCP K5D1G13ACH IC. This one is a multiple chip package, having 1Gbit of NAND Flash & 512Mbit of mobile SDRAM.
To it’s right is a Qualcomm RTR6285 RF Transceiver. This IC supports multiband GSM/EDGE/UMTS frequencies & also has a GPS receive amplifier included.
At bottom right is an Avago ACPM7371 Wide-Band 4×4 CDMA Power Amplifier. The external antenna connector is top right.
On the other side of the main PCB is a Qualcomm MSM6246 Baseband processor. Not sure about this one as I can’t find anything resembling a datasheet. Another micro-coax connector is in the centre, probably for factory test purposes, as it’s not accessible from the outside.
Just above the coax connector is a Qorvo RF1450 SP4T (single-pole 4-throw) High Power (34.5dBm) GSM RF Switch.
Upper right is an Avago FEM-7780 UMTS2100 4×7 Front End Module.
Under that is an RFMD RF3163 Quad-Band RF Power Amplifier Module.
These units are used to broadcast local audio, such as from a public address system or local microphone. They accomplish this by producing a modulated magnetic field that a hearing aid is capable of picking up.
Not many controls on this bit of equipment. A bi-colour LED for status indications, a microphone, external audio input, charging input & a power switch.
Popping the cover off reveals a small lead-acid battery, 2.1Ah at 12v. This is used when the loop is unplugged.
Here’s the main PCB, which takes care of the audio & battery charging. The inductive loop itself is just visible as the tape-covered wire bundle around the edge of the casing.
Here’s the input section of the main PCB. The microphone input is handled by a SSM2166 front-end preamplifier from Analog Devices.
This audio is then fed into a TDA2003 10W Mono Power Amplifier IC, which directly drives the induction coil as if it were a speaker. Any suitable receiving coil & amplifier can then receive the signal & change it back into audio.
I have yet another receipt printer, this one appears to be brand new. It’s possibly the smallest thermal 80mm printer I have at the moment, and has both USB & Serial interfaces.
There’s not much to these printers at all. Removing a single screw allows the case halves to separate, showing the guts. The controller is based around a Texas Instruments TMS320VC5509AFixed-Point DSP. It’s associated Flash ROM & RAM are to the right.
Power supply is dealt with in the top right of the PCB, with the interface ports further left.
Here’s the thermal mechanism itself, with the large print head. The stepper motor to drive the paper through the printer is just peeking out at top right. The paper present sensor is just under the left hand side of the print head.
A lot of the electronics I use & projects I construct use batteries, mainly of the lithium variety. As charging this chemistry can be a little explosive if not done correctly, I decided a proper charger was required. This charger is capable of handling packs up to 6 cells for Lithium, and up to 20v for lead-acids.
The usual DC input barrel jack on the left, with an external temp sensor for fast charging NiCd/NiMH chemistry batteries. The µUSB port registers under Linux as USB HID, probably so drivers aren’t required. Unfortunately the software is Windows only, but it doesn’t provide anything handy like charging graphs or stats. Just a way to alter settings & control charging from a PC. On other versions of this charger there’s a setting to change the temp sensor port into a TTL serial output, which would be much handier.
The other side of the charger has the main DC output jacks & the pack balancing connections.
Here’s the top cover removed from the charger, showing most of the internals. A standard HD44780 LCD provides the user interface, the CPU & it’s associated logic is hidden under there somewhere.
The PCB has nice heavy tracks to handle the 6A of current this charger is capable of.
The output side of the board. Here the resistive pack balancing network can be seen behind the vertical daughter board holding the connectors, along with the output current shunt between the DC output banana jacks & the last tactile button.
Unfortunately the LCD is soldered directly to the board, and my desoldering tool couldn’t quite get all the solder out, so time to get a bit violent. I’ve gently bent the header so I could see the brains of the charger. The main CPU is a Megwin MA84G564AD48, which is an Intel 8081 clone with USB support. Unfortunately I was unable to find a datasheet for this part, and the page on Megwin’s site is Chinese only.
I was hoping it was an ATMega328, as I have seen in other versions of this charger, as there are custom firmwares available to increase the feature set of the charger, but no dice on this one. I do think the µUSB port is unique to this version though, so avoiding models with that port probably would get a hackable version.
There’s some glue logic for controlling the resistor taps on the balancing network, and a few op-amps for voltage & current readings.
All the power diodes & switching FETs for the DC-DC converter are mounted on the bottom of the PCB, and clamped against the aluminium casing when the PCB is screwed down. Not the best way to ensure great contact, but Chinese tech, so m’eh.
It’s well known that there are two versions of the 701 type controller available for Eberspacher heaters, the version with the blue logo is the official un-restricted model, while the version with the white logo is a version built for BT that restricts the heater to 1 hour runtime & has no diagnostics built in.
As these devices are microcontroller driven, I assumed that the hardware would be the same, only the code running in the micro being the bit that Eberspacher changed. This option would certainly have been the lowest cost.
Here’s the PCB removed from the plastic housing. There are definitely some differences that I can tell. As the un-restricted version has an extra wire for the diagnostic serial interface, and this board has no unpopulated parts, the PCB is definitely a different version.
In the centre is a Microchip PIC16C622 microcontroller, the OTP version in this case for cost reductions. (I may try reading the binary from this chip in the future, chances are it’s code protected though).
Below the micro is an NXP PCF8577C 32-segment LCD controller, this has an I²C interface to the PIC.
The temperature control function on these heaters is done via applying a resistance to one of the control lines, between 1750Ω-2180Ω, ±80Ω. (Very odd values these, not to mention no standard components can create this range easily, bloody engineers >_<). This is accomplished in hardware with a BU2092F I²C shift register from Rohm, which is connected to a bank of resistors. The microcontroller will switch combinations of these into the circuit to get the range of resistances required.
The rest of the circuit is local power regulation & filtering.
There’s not much on the other side of the PCB, just the LCD itself & the contacts for the buttons.
I got one of these to test since I’ve been in need of some small DC pumps for fluid transfer use. At £2 I can definitely afford to experiment.
On the eBay listing, these pumps are rated at 3-12v DC, (I thought that was a bit wide of an operating range), I looked up the motor, an RS-360SH on Mabuchi’s website, they only have models in this range rated at 7.2v & 24v. Judging by the size of the windings on the armature & the fact that after a few minutes operation on 12v it gets rather hot, I’m going to say this is the 7.2v motor.
Removing the screws releases the end cover & the pair of gears inside. This operates like any other hydraulic gear motor, albeit with much wider tolerances. It has no capability to hold pressure when the power is removed, and can be blown through easily.
Flow & pressure under power are quite good for the pump’s size, even though it’s noisy as hell.
I’ve been a vaper now for many years, after giving up the evil weed that is tobacco. Here’s my latest acquisition in the vaping world, the JoyeTech eVIC 60W. This one is branded by Totally Wicked as the Forza VT60.
Powered by a single 18650 Li-Ion cell, this one is a Sony VTC4 series, 2100mAh.
Under the battery a pair of screws hold the electronics in the main cast alloy casing.
After removing the screws, the entire internal assembly comes out of the case, here’s the top of the PCB with the large OLED display in the centre.
On the right side of the board is the USB jack for charging & firmware updates. The adjustment buttons are also at this end.
On the left side of the board is the main output connector & the fire button. Unlike many eCigs I’ve torn down before, the wiring in this one is very beefy – it has to be to handle the high currents used with some atomizers – up to 10A.
Removing the board from the battery holder shows the main power circuitry & MCU. The aluminium heatsink is thermally bonded to the switching MOSFETs, a pair under each end. The switching inductor is under the gap in the centre of the heatsink.
A close up of the heatsink shows the very slim inductor under the heatsink.
The main MCU in this unit has a very strange part number, which I’ve been unable to find information on, but it’s probably 8081 based.
Inkeeping with everything else in my shack being low voltage operated, I had planned from the outset to convert the desoldering station to 12v operation. It turns out this has been the easiest tool to convert in my shack so far.
The factory SMPS is a fairly straightforward 18v 12A unit, with only a single small oddity: the desoldering gun’s heating element is controlled from inside the supply.
Next to the output rectifier on the heatsink is a large MOSFET, in this case a STP60NF06 from ST Micro. This is a fairly beefy FET at 60v & 60A capacity, RDS On of <0.016Ω.
This is driven via an opto-isolator from the main logic board. I’ve not yet looked at the waveform on the scope, but I suspect this is also being PWM’d to control temperature better when close to the set point.
Rather than fire up the soldering iron & build a new element controller circuit (Lazy Mode™), I opted to take a saw to the original power supply. I cut the DC output section of the PCB off the rest of the supply & attached this piece back to the frame of the base unit. I also added a small heatsink to the MOSFET to make sure it stays cool.
Since the fan & vacuum pump are both already 12v rated, those are connected directly to the DC input socket, that I’ve installed in place of the original IEC mains socket. The 18v for the heating element is generated by a 10A DC-DC converter, again from eBay.
Oddly, the iron itself is rated at 24v 80W, but the factory supply is only rated to 18v. I’m not sure why they’ve derated the system, but as the station already draws up to 10A from a 13.8v supply, increasing the voltage any further would start giving my DC supplies a problem, so it can stay at 18v for now.
For a long time I’ve needed a decent vacuum desoldering tool, as I do much stripping of old PCBs for random parts.
Solder wick works well for most things, but it’s expensive & can be fiddly. It also doesn’t keep very long as the copper braid oxidises & after that point it never seems to work particularly well, even when soaked in fresh flux.
As usual eBay to the rescue! I managed to pick this one up for £80.
Removing the lid reveals the internals. Front & centre is the vacuum pump, with the mains supply behind it. There’s also a very noisy cooling fan at the back. Not sure why since the unit never gets warm enough to actually warrant a fan.
On the other side is the PSU. This is an 18v 12A rated SMPS, with a bit of custom electronics for controlling the iron element. Mounted to the back case is a small black box, more to come on this bit.
Cracking the case of the PSU reveals a pretty bog-standard SMPS, with a surprising amount of mains filtering for a Chinese supply. The DC outputs are on the right.
From the rail markings, this is clearly designed to output some more voltage rails – possibly for other models of unit. In this case though, a single 18v rail is present. The iron’s element connects directly to the supply, controlled via an opto-isolated MOSFET.
As both the fan & the vacuum pump motor are 12v devices, some provision had to be made to reduce the 18v from the power supply to a more reasonable value. Inside the black plastic box are a pair of 1Ω 5W power resistors, connected in series. The output from this connects to the fan & vacuum pump. Because cheap, obviously.
Finally, here’s the controller PCB, the main MCU is an 8081 derivative, with a Holtek HT1621B LCD controller for the front panel temperature readout. Iron temperature is achieved by a thermocouple embedded in the heater, I imagine the potentiometer on the left side of the PCB is for calibration.
Here’s another retired piece of tech that we used to route media from the NAS to the main TV. It was retired since it’s inability to support XBMC/Kodi & having some crashing issues.
After attacking the case with the screwdriver (Torx in this case), the main board comes out. The CPU in this looks *very* familiar, being a PoP device. There are unpopulated places for an ethernet interface & USB port here.
After a little digging is turns out the CPU in this device is a BCM2835, with 256MB of RAM stacked on top. It’s a Raspberry Pi! Even the unpopulated part for Ethernet is the same SMSC LAN9512!
There’s 32MB of Flash for the software below the CPU.
On the far right of the board is a Broadcom BCM59002IML Mobile Power Management IC.
On the bottom of the PCB is the WiFi chipset, a Broadcom BCM4336, this most likely communicates with the CPU via SDIO. There’s also a section below for a Bluetooth chipset.
For a while I’ve wondered how these pancake type (AKA “Shaftless”) vibration motors operate, so I figured I’d mutilate one to find out.
These vibrators are found in all kinds of mobile devices as a haptic feedback device, unlike older versions, which were just micro-sized DC motors with an offset weight attached to the shaft, these don’t have any visible moving parts.
These devices are crimped together, so some gentle attack with a pair of snips was required to get the top cover off.
It turns out these are still a standard rotary DC motor, in this case specifically designed for the purpose. The rotor itself is the offset weight, just visible under the steel half-moon shaped section are the armature coils.
The armature lifts off the centre shaft, the coils can clearly be seen peeking out from under the counterweight.
The underside of the armature reveals the commutator, which in this device is just etched onto the PCB substrate, the connections to the pair of coils can be seen either side of the commutator segments.
The base of the motor holds the brushes in the centre, the outer ring is the stationary permanent magnet. These brushes are absolutely tiny, the whole motor is no more than 6mm in diameter.
Everyone at some stage must have seen these EAS security tags in shops, usually attached to clothing with a steel pin. As some of this year’s presents had been left with the tags attached, I had to forcibly remove them before wrapping could commence.
These are just a plastic disc about 50mm in diameter, with an internal locking mechanism & RF tag inside.
After some careful attack with a saw around the glue seam, the tag comes apart into it’s halves. The RF coil & it’s ceramic capacitor can be seen wrapped around the outside of the tag. The capacitor in this case isn’t even epoxy dipped to save that extra 0.0001p on the manufacturing price. In the top centre is the pin locking mechanism, enclosed in a small plastic pill.
Popping off the back cap of the lock shows it’s internals.
The lock itself is very simple. The centre section, held in place by a spring, carries 3 small ball bearings. The outer metal frame of the lock is conical in shape.
When the pin is pushed into the tag, the conical shape of the lock chamber causes the ball bearings to grab onto it, helped by the action of the spring that pushes the ball bearing carrier further into the cone.
This also means that any attempt to force the mechanism causes it to lock tighter onto the pin.
In normal operation, removal is achieved by a strong magnet that pulls the ball bearing carrier back slightly against it’s spring, allowing the pin to disengage & be pulled out.
This design is incredibly simple & cheap to make, and gains it’s locking strength from friction alone.
I would consider the RF coil being around the outer edge of the device a bit of a security risk – a quick chop with a sharp pair of wire cutters would disable the tag’s alarm functionality instantly. Making the coil slightly smaller & keeping it out of reach of the edge of the tag would help in this regard.
With a recent order from a Chinese seller on eBay, this little gadget was included in the package as a freebie:
I’ve not smoked for a long time, so I’m not too sure what use I’m going to find for this device, but it’s an electronic lighter!
Pushing the slider forward reveals a red-hot heater, mounted in the plastic (!) frame.
Pushing the other way reveals a USB port to charge the internal battery.
A couple of screws releases the end cap from the cover & the entire core unit slides out. Like all Chinese toys it’s made of the cheapest plastic imaginable, not such a good thing when heat is involved.
The element itself is a simple coil of Nichrome wire, crimped to a pair of brass terminals. The base the heater & it’s terminals are mounted to is actually ceramic – the surround though that this ceramic pill clips into is just the same cheap plastic. Luckily, the element only remains on for a few seconds on each button push, there’s no way to keep it on & start an in-pocket fire, as far as I can see.
The main PCB clips out of the back of the core frame, the large pair of tinned pads on the left connect to the heater, the control IC has no numbering of any kind, but considering the behaviour of the device it’s most likely a standard eCig control IC.
The other side of the board has the USB port on the right, the Lithium Polymer cell in the centre, and the power button on the left. The cell itself also has no marking, but I’m guessing a couple hundred mAh from the physical size.
I needed a decent WiFi adaptor for my latest Pi LCD project, so after trawling eBay for cheapy USB adaptors, I found this one.
Unlike most USB WiFi radios these days, it actually has a proper RP-SMA antenna connector, not the low-gain built in jobbies that never seem to work too well.
There are a few versions of this adaptor, all of which seem to use the same casing, there’s a button push cut into the plastic for a WPS button that doesn’t exist on this model. This is fine, as I don’t enable WPS on any of my network equipment anyway. (It’s insecure, and can be cracked in minutes).
Here’s the rest of the essential details, the model is BL-LW08-AR, rated at 300Mbit/s.
Here’s the PCB removed from the casing, there are a pair of PCB antennas on here, but they’re not connected to the RF circuitry in this model, the links are missing.
The chipset used is a Realtek RTL8191SU, there isn’t much more in this device, as it’s all built into the silicon.
This detector has now been retired from service since it’s a fair bit out of date. So here’s the teardown!
Unlike older detectors, this unit has a built in battery that never needs replacing during the life of the sensor, so once the unit reaches it’s expiry date it’s just trashed as a whole.
4 screws hold the cover on, here’s the internals of the detector. There’s a 3v CR123A LiMnO² cell at the right for power, rated at 1500mAh. A 7 year life is quite remarkable on a single cell!
The sensor is just to the left of the lithium cell, and is of quite unusual construction. Previous CO sensor cells I’ve seen have been small cylinders with a pair of brass pins. This one appears to use a conductive plastic as the connections. These sensors contain H²SO⁴ so they’re a bit hazardous to open.
There are no manufacturer markings on the sensor & I’ve not been able to find any similarly shaped devices, so I’m unsure of it’s specifications.
The alarm sounder is on the left, the usual Piezo disc with a resonator to increase the loudness.
The brains of the device are provided by a Microchip PIC16F914 microcontroller. This is a fairly advanced device, with many onboard features, and NanoWatt™ technology, standby power consumption is <100nA according to Microchip’s Datasheet. This would explain the incredible battery life.
The choke just at the right edge of the photo is actually a transformer to drive the Piezo sounder at high voltage.
Here’s the PCB with the LCD frame removed. Not much to see on the this side, the silence/test button top right & the front end for the sensor.
Here’s a closer look at the front end for the CO sensor cell itself. I haven’t been able to decode the SMT markings on the SOT packages, but I’m guessing that there’s a pair of OpAmps & a voltage reference.
