heatsink – Experimental Engineering

Posted on May 1, 2018April 29, 2018 by The Engineer — 8 Comments

Sterling ProCharge Ultra PCU1210 Teardown & Repair

The Sterling charger we’ve had on board nb Tanya Louise since Feb 2014 has bitten the dust, with 31220 hours on it’s internal clock. Since we’re a liveaboard boat, this charger has had a lot of use while we’re on the mooring during winter, when the solar bank isn’t outputting it’s full rate. First, a bit of a teardown to explore the unit, then onto the repair:

There’s the usual mains input filtering on the left, with the bridge rectifier on it’s heatsink.
Underneath the centre massive heatsinks is the main transformer (not visible here) & active PFC circuit. The device peeking out from underneath is the huge inductor needed for PFC. It’s associated switching MOSFET is to the right.

On the other side of the PFC section is the main DC rail filter electrolytic, a 450v 150µF part. Here some evidence of long-term heating can be seen in the adhesive around the base, it’s nearly completely turned black! It’s not a decent brand either, a Chinese CapXon.
The PCB fuse just behind it is in the DC feed to the main switching supply, so the input fuse only protects the filter & Active PFC circuitry. Luckily this fuse didn’t blow during the failure, telling me the fault was earlier in the power chain.
The logic circuits are powered by an independent switching supply in the centre, providing a +5v rail to the microcontroller. The fan header & control components are not populated in this 10A model, but I may end up retrofitting a fan anyway as this unit has always run a little too warm. The entire board is heavily conformal coated on both sides, to help with water resistance associated with being in a marine environment. This has worked well, as there isn’t a single trace of moisture anywhere, only dust from years of use.
There is some thermal protection for the main SMPS switching MOSFETS with the Klixon thermal fuse clipped to the heatsink.

The DC output rectifiers are on the large heatsink in the centre, with a small bodge board fitted. Due to the heavy conformal coating on the board I can’t get the ID from this small 8-pin IC, but from the fact that the output rectifiers are in fact IRF1010E MOSFETS, rated at 84A a piece, this is an synchronous rectifier controller.
Oddly, the output filter electrolytics are a mix of Nichicon (nice), and CapXon (shite). A bit of penny pinching here, which if a little naff since these chargers are anything but cheap. (£244.80 at the time of writing).
Hiding just behind the electrolytics is a large choke, and a reverse-polarity protection diode, which is wired crowbar-style. Reversing the polarity here will blow the 15A DC bus fuse instantly, and may destroy this diode if it doesn’t blow quick enough.

Right on the output end are a pair of large Ixys DSSK38 TO220 Dual 20A dual Schottky diodes, isolating the two outputs from each other, a nice margin on these for a 10A charger, since the diodes are paralleled each channel is capable of 40A. This prevents one bank discharging into another & allows the charger logic to monitor the voltages individually. The only issue here is the 400mV drop of these diodes introduce a little bit of inefficiency. To increase current capacity of the PCB, the aluminium heatsink is being used as the main positive busbar. From the sizing of the power components here, I would think that the same PCB & component load is used for all the chargers up to 40A, since both the PFC inductor & main power transformer are massive for a 10A output. There are unpopulated output components on this low-end model, to reduce the cost since they aren’t needed.

A trio of headers connect all the control & sense signals to the front panel PCB, which contains all the control logic. This unit is sensing all output voltages, output current & PSU rail voltages.

The front panel is stuffed with LEDs & 7-segment displays to show the current mode, charging voltage & current. There’s 2 tactile switches for adjustments.

The reverse of the board has the main microcontroller – again identifying this is impossible due to the heavy conformal coat. The LEDs are being driven through a 74HC245D CMOS Octal Bus Transceiver.

Now on to the repair! I’m not particularly impressed with only getting 4 years from this unit, they are very expensive as already mentioned, so I would expect a longer lifespan. The input fuse had blown in this case, leaving me with a totally dead charger. A quick multimeter test on the input stage of the unit showed a dead short – the main AC input bridge rectifier has gone short circuit.

Here the defective bridge has been desoldered from the board. It’s a KBU1008 10A 800v part. Once this was removed I confirmed there was no longer an input short, on either the AC side or the DC output side to the PFC circuit.

Time to stick the desoldered bridge on the milliohm meter & see how badly it has failed.

I’d say 31mΩ would qualify as a short. It’s no wonder the 4A input fuse blew instantly. There is no sign of excessive heat around the rectifier, so I’m not sure why this would have failed, it’s certainly over-rated for the 10A charger.

Now the defective diode bridge has been removed from the circuit, it’s time to apply some controlled power to see if anything else has failed. For this I used a module from one of my previous teardowns – the inverter from a portable TV.

This neat little unit outputs 330v DC at a few dozen watts, plenty enough to power up the charger with a small load for testing purposes. The charger does pull the voltage of this converter down significantly, to about 100v, but it still provides just enough to get things going.

After applying some direct DC power to the input, it’s ALIVE! Certainly makes a change from the usual SMPS failures I come across, where a single component causes a chain reaction that writes off everything.

Unfortunately I couldn’t find the exact same rectifier to replace the shorted one, so I had to go for the KBU1010, which is rated for 1000v instead of 800v, but the Vf rating (Forward Voltage), is the same, so it won’t dissipate any more power.

Here’s the new rectifier soldered into place on the PCB & bolted to it’s heatsink, with some decent thermal compound in between.

Here is the factory fuse, a soldered in device. I’ll be replacing this with standard clips for 20x5mm fuses to make replacement in the future easier, the required hole pattern in the PCB is already present. Most of the mains input filtering is also on this little daughterboard.

Now the fuse has been replaced with a standard one, which is much more easily replaceable. This fuse shouldn’t blow however, unless another fault develops.

Now everything is back together, a full load test charging a 200Ah 12v battery for a few hours will tell me if the fix is good. This charger won’t be going back into service onboard the boat, it’s being replaced anyway with a new 50A charger, to better suit the larger loads we have now. It won’t be a Sterling though, as they are far too expensive. I’ll report back if anything fails!

Posted on February 22, 2018 by The Engineer — 1 Comment

Brother PT-E300 Industrial Label Machine Teardown

Here a tape is installed in the printer. This unit can handle tape widths up to 18mm. The pinch rollers are operated by the white lever at the top of the image, which engages with the back cover.

This printer is supplied with a rechargeable battery pack, but AA cells can be used as well. Some of the AA battery terminals can be seen above the battery.

Pretty standard fare for a 2-cell lithium pack. The charging circuitry doesn’t appear to charge it to full voltage though, most likely to get the most life from the pack.

With the cartridge removed, the printer components can be seen. As these cartridges have in effect two rolls, one fro the ribbon & one for the actual label, there are two drive points.

The thermal print head is hidden on the other side of the steel heatsink, while the pinch rollers are on the top right. The plastic piece above the print head heatsink has a matrix of switches that engage with holes in the top of the label cartridge, this is how the machine knows what size of ribbon is fitted.

Most of the internal space is taken up by the main board, with the microprocessor & it’s program flash ROM top & centre.

The charger input is located on the keyboard PCB just under the mainboard, which is centre negative, as opposed to 99% of other devices using centre positive, the bastards.

The dot-matrix LCD is attached to the mainboard with a short flex cable, and from the few connections, this is probably SPI or I²C.

The printer itself is driven by a simple DC motor, speed is regulated by a pair of photo-interrupters forming an encoder on the second gear in the train.

The back case has the battery connections for both the lithium pack & the AA cells, the lithium pack has a 3rd connection, probably for temperature sensing.

Posted on March 22, 2017April 24, 2017 by The Engineer — Leave a comment

Philips LED PAR38 Lamp Teardown

These large LED Philips PAR38 lamps were recently on clearance sale in my local T.N. Robinsons electrical contractors for about £3, so I decided to grab one in the hopes I might be able to hack it into a low-voltage LED lamp. These are full-size PAR38 format, with most of the bulk being the large aluminium heatsink on the front. The back section with the power supply module is secured with silicone, so some unreasonable force was required to liberate the two pieces.

These lamps are rated at 18W in operation, and are surprisingly bright for this power level.

The front has the moulded multi-lens over the LEDs, to spread the light a bit further than the bare dies.

The LED array is two series strings of 4 LEDs, for ~24v forward voltage. Unusual for a high power LED array, this PCB isn’t aluminium cored, but 0.8mm FR4. Heat is transferred to the copper plane on the backside by the dozens of vias around the Luxeon Rebel LEDs. There is a thermal pad under the PCB for improved heat transfer to the machined surface of the heatsink.

The power supply & control PCB is pretty well made, it’s an isolated converter, so no nasty mains on the LED connections.

Posted on February 15, 2017March 27, 2017 by The Engineer — Leave a comment

16-Port SATA PCIe Card – Cooling Recap

It’s been 4 months since I did a rejig of my storage server, installing a new 16-port SATA HBA to support the disk drives. I mentioned the factory fan the card came with in my previous post, and I didn’t have many hopes of it surviving long.

The heatsink card has barely had enough time to accumulate any grime from the air & the fan has already failed!

There’s no temperature sensing or fan speed sensing on this card, so a failure here could go unnoticed, and under load without a fan the heatsink becomes hot enough to cause burns. (There are a total of 5 large ICs underneath it). This would probably cause the HBA to overheat & fail rather quickly, especially when under a high I/O load, with no warning. In my case, the bearings in the fan failed, so the familiar noise of a knackered sleeve bearing fan alerted me to problems.

A replacement 80mm Delta fan has been attached to the heatsink in place of the dead fan, and this is plugged into a motherboard fan header, allowing sensing of the fan speed. The much greater airflow over the heatsink has dramatically reduced running temperatures. The original fan probably had it’s bearings cooked by the heat from the card as it’s airflow capability was minimal.

Here’s the old fan removed from the heatsink. The back label, usally the place where I’d expect to find some specifications has nothing but a red circle. This really is the cheapest crap that the manufacturer could have fitted, and considering this HBA isn’t exactly cheap, I’d expect better.

Peeling off the back label reveals the back of the bearing housing, with the plastic retaining clip. There’s some sign of heat damage here, the oil has turned into gum, all the lighter fractions having evaporated off.

The shaft doesn’t show any significant damage, but since the phosphor bronze bearing is softer, there is some dirt in here which is probably a mix of degraded oil & bearing material.

There’s more gunge around the other end of the bearing & it’s been worn enough that side play can be felt with the shaft. In ~3000 hours running this fan is totally useless.

Posted on December 14, 2016 by The Engineer — Leave a comment

Mercury 30A Ham Radio SMPS

After having a couple of the cheap Chinese PSUs fail on me in a rather spectacular fashion, I decided to splash on a more expensive name-brand PSU, since constantly replacing PSUs at £15 a piece is going to get old pretty fast. This is the 30A model from Mercury, which seems to be pretty well built. It’s also significantly more expensive at £80. Power output is via the beefy binding posts on the front panel. There isn’t any metering on board, this is something I’ll probably change once I’ve ascertained it’s reliability. This is also a fixed voltage supply, at 13.8v.

Not much on the rear panel, just the fuse & cooling fan. This isn’t temperature controlled, but it’s not loud. No IEC power socket here, the mains cable is hard wired.

Removing some spanner-type security screws reveals the power supply board itself. Everything on here is enormous to handle the 30A output current at 13.8v. The main primary side switching transistors are on the large silver heatsink in the centre of the board, feeding the huge ferrite transformer on the right.

The transformer’s low voltage output tap comes straight out instead of being on pins, due to the size of the winding cores. Four massive diodes are mounted on the black heatsinks for output rectification.

The supply is controlled via the jelly bean TL494 PWM controller IC. The multi-turn potentiometer doesn’t adjust the output voltage, more likely it adjusts the current limit.

Power to initially start the supply is provided by a small SMPS circuit, with a VIPer22A Low Power Primary Switcher & small transformer on the lower right. The transformer upper left is the base drive transformer for the main high power supply.

Posted on November 14, 2016March 27, 2017 by The Engineer — Leave a comment

Project Volantis – Storage Server Rebuild

For some time now I’ve been running a large disk array to store all the essential data for my network. The current setup has 10x 4TB disks in a RAID6 array under Linux MD.

Up until now the disks have been running in external Orico 9558U3 USB3 drive bays, through a PCIe x1 USB3 controller. However in this configuration there have been a few issues:

Congestion over the USB3 link. RAID rebuild speeds were severely limited to ~20MB/s in the event of a failure. General data transfer was equally as slow.
Drive dock general reliability. The drive bays are running a USB3 – SATA controller with a port expander, a single drive failure would cause the controller to reset all disks on it’s bus. Instead of losing a single disk in the array, 5 would disappear at the same time.
Cooling. The factory fitted fans in these bays are total crap – and very difficult to get at to change. A fan failure quickly allows the disks to heat up to temperatures that would cause failure.
Upgrade options difficult. These bays are pretty expensive for what they are, and adding more disks to the USB3 bus would likely strangle the bandwidth even further.
Disk failure difficult to locate. The USB3 interface doesn’t pass on the disk serial number to the host OS, so working out which disk has actually failed is difficult.

To remedy these issues, a proper SATA controller solution was required. Proper hardware RAID controllers are incredibly expensive, so they’re out of the question, and since I’m already using Linux MD RAID, I didn’t need a hardware controller anyway.

A quick search for suitable HBA cards showed me the IOCrest 16-port SATAIII controller, which is pretty low cost at £140. This card breaks out the SATA ports into standard SFF-8086 connectors, with 4 ports on each. Importantly the cables to convert from these server-grade connectors to standard SATA are supplied, as they’re pretty expensive on their own (£25 each).
This card gives me the option to expand the array to 16 disks eventually, although the active array will probably be kept at 14 disks with 2 hot spares, this will give a total capacity of 48TB.

Here’s the card installed in the host machine, with the array running. One thing I didn’t expect was the card to be crusted with activity LEDs. There appears to be one LED for each pair of disks, plus a couple others which I would expect are activity on the backhaul link to PCIe. (I can’t be certain, as there isn’t any proper documentation anywhere for this card. It certainly didn’t come with any ;)).
I’m not too impressed with the fan that’s on the card – it’s a crap sleeve bearing type, so I’ll be keeping a close eye on this for failure & will replace with a high quality ball-bearing fan when it finally croaks. The heatsink is definitely oversized for the job, with nothing installed above the card barely gets warm, which is definitely a good thing for life expectancy.

Update 10/02/17 – The stock fan is now dead as a doornail after only 4 months of continuous operation. Replaced with a high quality ball-bearing 80mm Delta fan to keep things running cool. As there is no speed sense line on the stock fan, the only way to tell it was failing was by the horrendous screeching noise of the failing bearings.

Above is the final HBA installed in the PCIe x1 slot above – a parallel SCSI U320 card that handles the tape backup drives. This card is very close to the cooling fan of the SATA card, and does make it run warmer, but not excessively warm. Unfortunately the card is too long for the other PCIe socket – it fouls on the DIMM slots.

The tape drives are LTO2 300/600GB for large file backup & DDS4 20/40GB DAT for smaller stuff. These were had cheap on eBay, with a load of tapes. Newer LTO drives aren’t an option due to cost.

The main disk array is currently built as 9 disks in service with a single hot spare, in case of disk failure, this gives a total size after parity of 28TB:

/dev/md0:
        Version : 1.2
  Creation Time : Wed Mar 11 16:01:01 2015
     Raid Level : raid6
     Array Size : 27348211520 (26081.29 GiB 28004.57 GB)
  Used Dev Size : 3906887360 (3725.90 GiB 4000.65 GB)
   Raid Devices : 9
  Total Devices : 10
    Persistence : Superblock is persistent

  Intent Bitmap : Internal

    Update Time : Mon Nov 14 14:28:59 2016
          State : active 
 Active Devices : 9
Working Devices : 10
 Failed Devices : 0
  Spare Devices : 1

         Layout : left-symmetric
     Chunk Size : 64K

           Name : Main-PC:0
           UUID : 266632b8:2a8a3dd3:33ce0366:0b35fad9
         Events : 773938

    Number   Major   Minor   RaidDevice State
       0       8       48        0      active sync   /dev/sdd
       1       8       32        1      active sync   /dev/sdc
       9       8       96        2      active sync   /dev/sdg
      10       8      112        3      active sync   /dev/sdh
      11       8       16        4      active sync   /dev/sdb
       5       8      176        5      active sync   /dev/sdl
       6       8      144        6      active sync   /dev/sdj
       7       8      160        7      active sync   /dev/sdk
       8       8      128        8      active sync   /dev/sdi

      12       8        0        -      spare   /dev/sda

/dev/md0:

Version : 1.2

Creation Time : Wed Mar 11 16:01:01 2015

Raid Level : raid6

Array Size : 27348211520 (26081.29 GiB 28004.57 GB)

Used Dev Size : 3906887360 (3725.90 GiB 4000.65 GB)

Raid Devices : 9

Total Devices : 10

Persistence : Superblock is persistent

Intent Bitmap : Internal

Update Time : Mon Nov 14 14:28:59 2016

State : active

Active Devices : 9

Working Devices : 10

Failed Devices : 0

Spare Devices : 1

Layout : left-symmetric

Chunk Size : 64K

Name : Main-PC:0

UUID : 266632b8:2a8a3dd3:33ce0366:0b35fad9

Events : 773938

Number Major Minor RaidDevice State

0 8 48 0 active sync /dev/sdd

1 8 32 1 active sync /dev/sdc

9 8 96 2 active sync /dev/sdg

10 8 112 3 active sync /dev/sdh

11 8 16 4 active sync /dev/sdb

5 8 176 5 active sync /dev/sdl

6 8 144 6 active sync /dev/sdj

7 8 160 7 active sync /dev/sdk

8 8 128 8 active sync /dev/sdi

12 8 0 - spare /dev/sda

The disks used are Seagate ST4000DM000 Desktop HDDs, which at this point have ~15K hours on them, and show no signs of impending failure.

Here’s a screenshot with the disk array fully loaded running over USB3. The aggregate speed on the md0 device is only 21795KB/s. Extremely slow indeed.

This card is structured similarly to the external USB3 bays – a PCI Express bridge glues 4 Marvell 9215 4-port SATA controllers into a single x8 card. Bus contention may become an issue with all 16 ports used, but as far with 9 active devices, the performance increase is impressive. Adding another disk to the active array would certainly give everything a workout, as rebuilding with an extra disk will hammer both read from the existing disks & will write to the new.

With all disks on the new controller, I’m sustaining read speeds of 180MB/s. (Pulling data off over the network). Write speeds are always going to be pretty pathetic with RAID6, as parity calculations have to be done. With Linux MD, this is done by the host CPU, which is currently a Core2Duo E7500 at 2.96GHz, with this setup, I get 40-60MB/s writes to the array with large files.

Since I don’t have a suitable case with built in drive bays, (again, they’re expensive), I’ve had to improvise with some steel strip to hold the disks in a stack. 3 DC-DC converters provides the regulated 12v & 5v for the disks from the main unregulated 12v system supply. Both the host system & the disks run from my central battery-backed 12v system, which acts like a large UPS for this.

The SATA power splitters were custom made, the connectors are Molex 67926-0001 IDC SATA power connectors, with 18AWG cable to provide the power to 4 disks in a string.

These require the use of a special tool if you value your sanity, which is a bit on the expensive side at £25+VAT, but doing it without is very difficult. You get a very well made tool for the price though, the handle is anodised aluminium & the tool head itself is a 300 series stainless steel.

Posted on October 27, 2016 by The Engineer — Leave a comment

LG Flatron 22EA53VQ-P Power Issue

I was recently given a pretty nice LED backlit 1080p LG monitor, with the instruction that it wouldn’t power on correctly. The monitor would power on as far as the standby light, but when fully powered on, would flash the backlight momentarily then shut down. A power supply issue was immediately suspected.

I popped the covers off the monitor itself first, thinking that it was an electrolytic gone bad in the backlight DC-DC converter. Not to mention the fact that cracking into a wall-wart type of PSU is only occasionally possible without the use of anger & large hammers. (Cracking the glue with the handle of a screwdriver doesn’t work so well when the factory went a bit nuts with the glue/ultrasonic welder). As can be seen in the photo, there’s not much inside these monitors, the logic is a single-chip solution, the rest of the PCB is dedicated to supplying the power rails for the various circuits. On the left is the power input & the DC-DC converter for the backlight, along with the DC-DC converter supplying the logic circuits. None of the capacitors here are damaged, everything looks good.
I then measured the output of the PSU, which under no load was the correct 19v DC. However applying any load caused the output voltage to drop like a proverbial brick. Applying a full load of 1.3A saw the output voltage drop so severely that the PSU tripped on it’s UVLO.

At 200mA of load the factory PSU is already dropping to 18v, with a 5.3kHz switching frequency appearing.

At higher load the frequency increases to 11.5kHz & the output voltage has dropped to 11.86v!

750mA was as high as I could make the supply go without it tripping itself out – the UVLO circuit trips at 9v. 12.6kHz is now riding on the severely low DC at this point.

The power supply is supposed to be rated at 1.3A at 19v, however with this fault it’s getting nowhere near that. The LG brand is on this PSU but it’s contracted out to Shenzen Honor Electric Co. Ltd.

Here’s the problem with this PSU. The output electrolytic has ballooned. I don’t have an ESR tester, but this cap has gone way past it’s sell-by date. It’s position right next to the heatsink with the output rectifier diodes has probably cooked it. The PSU isn’t that badly built for a Chinese one – there’s plenty of creepage distance on the PCB & even a couple of isolation slots.

Posted on January 3, 2016 by The Engineer — Leave a comment

Totally Wicked Forza / JoyeTech eVIC 60W Teardown

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.

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.

Posted on January 3, 2016January 7, 2016 by The Engineer — Leave a comment

Duratool ZD-915 12v Conversion

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.

Posted on October 31, 2015 by The Engineer — 2 Comments

Cisco EA6700 / Linksys AC1750 Router

Since the boat was still running it’s internal network on 10/100M speeds, an upgrade was decided on, the internal WiFi signal strength was also pretty poor further than a few feet from the NOC.

The new router is a Cisco/Linksys AC1750 model, with gigabit networking, and full 802.11ac 2.4/5GHz Wireless. This router also has a built in media server, print server, USB3 & USB2.

Teardown time! Here’s the router with the cover removed. Most of the fun stuff is hidden under the shields, but these aren’t fully soldered down & the covers are removable. The 6 antennas can be seen spaced around the edge of the housing, the main CPU is under the large heatsink upper centre. The radio power amplifier stages are underneath the shields, while the main RF transceivers are just outside the shields.

Wireless N is provided by a Broadcom BCM4331, this provides full dual-band 3×3 802.11n support. Being 3×3 it is actually 3 separate transceivers in a single package, to get much higher throughput rates of 600Mbit/s.

Wireless AC is provided by it’s sister IC, the BCM4360, with throughput speeds of 1.3Gbit/s. Both of these transceiver ICs connect back to the main CPU via PCI Express.

To get increased range, there are a trio of Skyworks SE5003L +23dBm 5GHz power amplifier ICs under the shield, along with the TX/RX switching & antenna matching networks. Heatsinking for these is provided by a sink screwed to the bottom side of the PCB. The outputs to the antennas can be seen at the top of the image.

The 2.4GHz section is fitted with a trio of Skyworks SE2605L +23dBm 2.4GHz power amplifiers, with a similar heatsink arrangement under the board. Unlike the 5GHz section, the 2.4GHz antenna feeds are soldered to the PCB here instead of using connectors.

The main CPU is a BCM4708 Communications Processor from Broadcom, as for the other Broadcom chips in this router, very little information is available unless under NDA, but I do know it’s a dual core ARM Cortex A9 running at 1GHz, with built in 5-port gigabit ethernet switch.

Working RAM for the processor is a Hynix H5TQ2G63DFA 256MB part.

More to come on the installation of the new networking, with it’s associated 4G mobile gateway connection system.

73s for now!

Posted on October 22, 2015 by The Engineer — Leave a comment

Netgear GS308 Gigabit Switch

Here’s a new addition to the network, mainly to replace the ancient Cisco Catalyst 3500 XL 100MB switch I’ve been using for many years, until I can find a decently priced second hand commercial gigabit switch.

Here’s the switch with some network connections on test. So far it’s very stable & draws minimum power. I’ve not yet attempted to run my core links (NAS) through yet, as I’ve not yet seen a consumer grade switch that can stand up to constant full load without crashing.

Here’s the switch with it’s lid popped. The magnetics can be seen at the back, next to the RJ-45 ports, the large IC in the centre is the main switching IC, with a heatsink bonded to the top. Very minimal design, with only a couple of switching regulators for power supply & not much else.

Here’s a closeup of some of the support components. There’s a 25MHz crystal providing a clock signal for the switch IC, just to the right of that is an EEPROM. I imagine this is storing the switch configuration & MAC address. Further right is one of the switching DC-DC converter ICs for power.

As a quick test, here’s 500GB of data being shifted through the switch, at quite an impressive rate. I’m clearly maxing out the bandwidth of the link here. Soon I will upgrade to a 10G Ethernet link between the NAS & main PC to get some more performance.

Posted on October 20, 2015October 18, 2015 by The Engineer — Leave a comment

DIY SMPS Fan Controller

Now the controllers have arrived, I can rejig the supplies to have proper thermal control on their cooling.

Here’s the top off the PSU. The board has been added to the back panel, getting it’s 12v supply from the cable that originally fed the fan directly. Luckily there was just enough length on the temperature probe to fit it to the output rectifier heatsink without modification.

To connect to the standard 4-pin headers on the controller, I’ve spliced on a PC fan extension cable, as these fans spent their previous lives in servers, with odd custom connectors.

Here’s the controller itself, the temperature probe is inserted between the main transformer & the rectifier heatsink.
I’ve set the controller to start accelerating the fan at 50°C, with full speed at 70°C.

Under a full load test for 1 hour, the fan didn’t even speed up past about 40% of full power. The very high airflow from these fans is doing an excellent job of keeping the supply cool. Previously the entire case was very hot to the touch, now everything is cool & just a hint of warm air exits the vents. As the fan never runs at full speed, the noise isn’t too deafening, and immediately spools back down to minimum power when the load is removed.

Posted on September 27, 2015 by The Engineer — Leave a comment

DIY SMPS Cooling

The power supplies I have recently built from surplus Cisco switch boards have started displaying a rather irritating problem – continual load of over 9A causes the supplies to shut down on overheat.

This was partially expected, as the original switches that these supplies came from are cooled by a monster of a centrifugal blower that could give a Dyson a run for it’s money. The problem with these fans is that they’re very loud, draw a lot of power (3-4A) and aren’t small enough to fit into the case I’ve used for the project.

The solution of course, is a bigger fan – I’ve got some Delta AFB0612EHE server fans, these are very powerful axial units, shifting 60CFM at 11,000RPM, with a power draw of 1.12A.
They’re 60mm diameter, so only just fit into the back of the case – although they stick out of the back by 40mm.

Here’s the fan, not the beefiest I have, but the beefiest that will fit into the available space.
These will easily take fingers off if they get too close at full speed, so guards will definitely be required.

To reduce the noise (they sound like jet engines at full pelt), I have ordered some PWM controllers that have a temperature sensor onboard, so I can have the fan run at a speed proportional to the PSU temperature. I will probably attach the sensor to the output rectifier heatsink, since that’s got the highest thermal load for it’s size.

More to come when parts arrive!

73s for now 🙂

Posted on September 2, 2015 by The Engineer — Leave a comment

USB1100 Digital Message Unit

This is basically an industrial, rugged MP3 player, in an extruded aluminium case.
They are used in commercial settings for generating telephone hold music or continual playback of background music in shops.

It’s quite a compact unit, in a nice aluminium case, designed for mounting into a comms setup. This unit will play any MP3 file, up to a maximum size of 11MB.

Here’s the user connections on the end of the unit. The device takes a standard 12v DC input, and has a single button for setup, user feedback is given through the multi-colour LED next to the power jack.
Both 8Ω & 600Ω audio outputs are provided for maximum compatibility. Volume & tone controls are also here.
On the other end of the unit is a single USB port for loading the audio files from a USB drive, and a reset button.

Here’s the single PCB removed from the casing. Unfortunately the main CPU has had it’s part number sanded off, and I can’t be bothered to try & find out what kind of processor it is at this point. To the right of the CPU are some flash ROM & SDRAM, along with the single USB port at bottom right.
The left side of the board is dedicated to audio output & voltage regulation, there are a fair few linear regulators in this unit.

Here’s the audio output side of the board, the transformer on the left is to provide the 600Ω output, the audio amplifier IC (BA5416) is just behind it. To the right are some of the main voltage regulators, a 5v one on the heatsink & a LM317.

The audio codec is a CS4271 from Cirrus Logic, a really high quality part, 24-bit resolution, 192kHz Stereo codec. Considering this is for telephone & PA systems that aren’t that high fidelity, it’s well built!

Posted on August 12, 2015 by The Engineer — Leave a comment

Chinese 12v 10A Power Brick Analysis

I recently ordered a PSU to run one of the TVs I converted to 12v operation, and being an older TV, it’s a fairly heavy load at 6.5A. eBay to the rescue again, with a cheap 10A rated supply.

Like all similar supplies these days, it’s a SMPS unit, and feels suspiciously light for it’s power rating.

Luckily this one is easy to get into, no ultrasonic welding on the case, just clips. Here’s the top cover removed, big alloy plate between the heatsinks.

The top heatsink plate was glued to the top of the transformer with silicone, some gentle prying released it. From the top, things don’t look too bad. There’s some filtering on the mains input & it’s even fused!

Here’s a closeup of the primary side of the PSU, the main DC bus capacitor is a Nichicon one, but it’s clearly been recovered from another device, look at the different glue on the end!
it’s also flapping about in the breeze, the squirt of silicone they’ve put on does nothing to stop movement.
Also here is the mains input fuse, filter capacitor & common mode choke. At least there is some filtering!

The main control IC is a UC3843B High Performance Current Mode PWM Controller, operating at a switching frequency of 250kHz.
The main switching transistor is visible at the bottom left corner, attached to the heatsink.

Here’s the secondary side of the supply. The transformer itself is OK, nice heavy windings on the output to suit the high current.
It’s using proper opto-isolated feedback for voltage regulation, with a TL431 reference IC.
The output diodes are attached to the heatsink at the top of the photo, I couldn’t read any numbers on those parts.

The output filter capacitors are low quality, only time will tell if they survive. I’ll put the supply under full load & see what the temperature rise is inside the casing.

On the bottom of the PCB things get a little more dire. There isn’t really much of an isolation gap between the primary & secondary sides, and there’s a track joining the output negative with mains earth, which gets to within 2mm of the live mains input!

As with all these cheapo supplies, there’s good points & bad points, I will update when I’ve had a chance to put the supply under full load for a while & see if it explodes!

Posted on July 14, 2015July 19, 2015 by The Engineer — 1 Comment

Cisco PSU Hack & Switched Mode PSU Background

Recently I decommissioned some networking equipment, and discovered the power supplies in some switches were single rail 12v types, with a rather high power rating. I figured these would be very good for powering my Ham radio gear.

They’re high quality Delta Electronics DPSN-150BP units, rated at a maximum power output of 156W.

These supplies have an adjustment pot for the output voltage regulation, but unfortunately it just didn’t have quite enough range to get from 12.0v to 13.8v. The highest they would go was ~13.04v.

After taking a look at the regulator circuit, I discovered I could further adjust the output voltage by changing a single resistor to a slightly lower value.

Firstly though, a little background on how switched mode power supplies operate & regulate their output voltage.

Here’s the supply. It’s mostly heatsink, to cool the large power switching transistors.

The first thing a SMPS does, is to rectify the incoming mains AC with a bridge rectifier. This is then smoothed by a large electrolytic capacitor, to provide a main DC rail of +340v DC (when on a 240v AC supply).

Above is the mains input section of the PSU, with a large common-mode choke on the left, bridge rectifier in the centre, and the large filter capacitor on the right. These can store a lot of energy when disconnected from the mains, and while they should have a discharge resistor fitted to safely drain the stored energy, they aren’t to be relied on for safety!

Once the supply has it’s main high voltage DC rail, this is switched into the main transformer by a pair of very large transistors – these are hidden from view on the large silver heatsinks at the bottom of the image. These transistors are themselves driven with a control IC, in the case of this supply, it’s a UC3844B. This IC is hidden under the large heatsink, but is just visible in the below photo. (IC5).

Here’s the main switching transformer, these can be much smaller than a conventional transformer due to the high frequencies used. This supply operates at 500kHz.
After the main transformer, the output is rectified by a pair of Schottky diodes, which are attached to the smaller heatsink visible below the transformer, before being fed through a large toroidal inductor & the output filter capacitors.
All this filtering on both the input & the output is required to stop these supplies from radiating their operating frequency as RF – a lot of cheap Chinese switching supplies forego this filtering & as a result are extremely noisy.

After all this filtering the DC appears at the output as usable power.

Getting back to regulation, these supplies read the voltage with a resistor divider & feed it back to the mains side control IC, through an opto-isolator. (Below).

The opto isolators are the black devices at the front with 4 pins.

For a more in-depth look at the inner workings of SMPS units, there’s a good article over on Hardware Secrets.

My modification is simple. Replacing R306 (just below the white potentiometer in the photo), with a slightly smaller resistor value, of 2.2KΩ down from 2.37KΩ, allows the voltage to be pulled lower on the regulator. This fools the unit into applying more drive to the main transformer, and the output voltage rises.

It’s important to note that making too drastic a change to these supplies is likely to result in the output filter capacitors turning into grenades due to overvoltage. The very small change in value only allows the voltage to rise to 13.95v max on the adjuster. This is well within the rating of 16v on the output caps.

Now the voltage has been sucessfully modified, a new case is on the way to shield fingers from the mains. With the addition of a couple of panel meters & output terminals, these supplies will make great additions to my shack.

More to come on the final build soon!

Posted on September 30, 2014October 6, 2014 by The Engineer — Leave a comment

HPI Nitrostar F4.6 Ignition Conversion

As there was no other online example of someone converting a glow/nitro car engine onto CDI ignition, I thought I would document the highlights here.
The engine is currently still running on glow fuel, but when the required fuel lines arrive I will be attempting the switch over to 2-Stroke petrol mix. This should definitely save on fuel costs.

The engine in this case is a HPI NitroStar F4.6 nitro engine, from a HPI Savage X monster truck.

Above is the converted engine with it’s timing sensor. As The installation of this was pretty much standard, a complete strip down of the engine was required to allow the drilling & tapping of the two M3x0.5 holes to mount the sensor bracket to. The front crankshaft bearing has to be drifted out of the crankcase for this to be possible.

Detail of the ignition hall sensor. The bracket has to be modified to allow the sensor to face the magnet in the flywheel. Unlike on an Aero engine, where the magnet would be on the outside edge of the prop driver hub, in this case the hole was drilled in the face of the flywheel near the edge & the magnet pressed in. The Hall sensor is glued to the modified bracket with the leads bent to position the smaller face towards the back of the flywheel.
The clearance from the magnet to sensor is approx. 4mm.

Detail of the magnet pressed into the flywheel. A 3.9mm hole was drilled from the back face, approx 2mm from the edge, & the magnet pressed into place with gentle taps from a mallet & drift, as I had no vice to hand.
Initial timing was a little fiddly due to the flywheel only being held on with a nut & tapered sleeve, so a timing mark can be made inside the rear of the crankcase, across the crank throw & case to mark the 28 degree BTDC point, the flywheel is then adjusted to make the ignition fire at this point, before carefully tightening the flywheel retaining nut to ensure no relative movement occurs.
The slots in the sensor bracket allow several degrees of movement to fine adjust the timing point once this rough location has been achieved.

Definitely the tiniest spark plug I’ve ever seen, about an inch long. Some trouble may be encountered with this on some engines – the electrodes stick out about 2mm further into the combustion chamber than a standard glow plug does. This causes the ground electrode to hit the top of the piston crown. (This happens on the HPI NitroStar 3.5 engine). The addition of another copper washer under the plug before tightening should cure this problem.

Ignition module. Due to the depth of the plug in the heatsink head on these engines, I will have to modify the plug cap to straighten it out, as it will not fit in this configuration.
However, ignition modules are available from HobbyKing with straight plug caps, this makes modification unnecessary

The ignition & components used on this system were obtained from JustEngines.

Posted on December 3, 2013December 4, 2013 by The Engineer — Leave a comment

555 Flyback Driver

Here is a simple 555 timer based flyback transformer driver, with the PCB designed by myself for some HV experiments. Above is the Eagle CAD board layout.

The 555 timer is in astable mode, generating a frequency from about 22kHz to 55kHz, depending on the position of the potentiometer. The variable frequency is to allow the circuit to be tuned to the resonant frequency of the flyback transformer in use.

This is switched through a pair of buffer transistors into a large STW45NM60 MOSFET, rated at 650v 45A.

Input power is 15-30v DC, as the oscillator circuit is fed from an independent LM7812 linear supply.

Provision is also made on the PCB for attaching a 12v fan to cool the MOSFET & linear regulator.

Board initially built, with the heatsink on the linear regulator fitted. I used a panel mount potentiometer in this case as I had no multiturn 47K pots in stock.

Bottom of the PCB. The main current carrying traces have been bulked up with copper wire to help carry the potentially high currents on the MOSFET while driving a large transformer.
This board was etched using the no-peel toner transfer method, using parchement paper as the transfer medium.

Main MOSFET now fitted with a surplus heatsink from an old switchmode power supply. A Fan could be fitted to the top of this sink to cope with higher power levels.

This is the gate drive waveform while a transformer is connected, the primary is causing some ringing on the oscillator. The waveform without an attached load is a much cleaner square wave.

I obtained a waveform of the flyback secondary output by capacitively coupling the oscilloscope probe through the insulation of the HT wire. The pulses of HV can be seen with the decaying ringing of the transformer between cycles.

Corona & arc discharges at 12v input voltage.

Download the Eagle schematic files here: [download id=”5561″]

Posted on April 22, 2013April 22, 2013 by The Engineer — Leave a comment

Wearable Raspberry Pi – Some Adjustments

As the first USB hub I was using was certainly not stable – it would not enumerate between boots & to get it working again would require waiting around 12 hours before applying power, it has been replaced. This is a cheapie eBay USB hub, of the type shown below.

These hubs are fantastic for hobbyists, as the connections for power & data are broken out on the internal PCB into a very convenient row of pads, perfect for integration into many projects.

I now have two internal spare USB ports, for the inbuilt keyboard/mouse receiver & the GPS receiver I plan to integrate into the build.

These hubs are also made in 7-port versions, however I am not sure if these have the same kind of breakout board internally. As they have the same cable layout, I would assume so.

Here is a closeup of the back of the connectors, showing a couple of additions.

I have added a pair of 470µF capacitors across the power rails, to further smooth out the ripple in the switching power supply, as I was having noise issues on the display.

Also, there is a new reset button added between the main interface connectors, which will be wired into the pair of pads that the Raspberry Pi has to reset the CPU.
This can be used as a power switch in the event the Pi is powered down when not in use & also to reset the unit if it becomes unresponsive.

Posted on April 16, 2013April 22, 2013 by The Engineer — Leave a comment

Wearable Raspberry Pi Part 2.5 – Battery Pack PCM

The final part for the battery pack has finally arrived, the PCM boards. These modules protect the cells by cutting off the power at overcharge, undercharge & overcurrent. Each cell is connected individually on the right, 12v power appears on the left connections. These modules also ensure that all the cells in the pack are balanced.

Posted on April 6, 2013April 22, 2013 by The Engineer — Leave a comment

Wearable Raspberry Pi SMPS Modifications

A few modifications were required to the SMPS modules to make the power rails stable enough to run the Pi & it’s monitor. Without these the rails were so noisy that instability was being caused.

I have replaced the 100µF output capacitors & replaced them with 35v 4700µF caps. This provides a much lower output ripple.

There are also heatsinks attached to the converter ICs to help spread the heat.

Posted on January 25, 2013 by The Engineer — Leave a comment

1.5W 445nm Lab Laser

Here is a followup from the 1.5W laser module post.

The module has been fitted into a housing, with a 2.2Ah Li-Poly battery pack. Charging is accomplished with an external 12.6v DC power supply.

Above can be seen the pair of switches on the top, the keyswitch must be enabled for the laser to fire.

When armed, the ring around the push button illuminates blue, as a warning that the unit is armed.

Inside the unit. The Li-Poly battery pack is at the bottom, with it’s protection & charging circuitry on the top. The switches are wired in series, with the LED connected to illuminate when the keyswitch is turned to the ON position.

The push button applies power to the laser driver module, which regulates the input power to safely drive the semiconductor laser in the aluminium heatsink housing.

Posted on July 25, 2012July 25, 2012 by The Engineer — Leave a comment

Raspberry Pi GPIO Breakout

After seeing these on eBay for £8.99 I thought it might be a good deal – interfacing with the RasPi’s GPIO & it has built in power supplies.

As a kit, it was very easy to assemble, the PCB quality is high, and is a fairly good design. It worked first time, the regulators hold the rails at the right voltages.
However there are some issues with this board that bug me.

The documentation for the kit is *AWFUL*. No mention of the regulators on the parts list & which goes where – I had to carefully examine the schematics to find out those details.
The 4x 1N1007 diodes required weren’t even included in the kit! Luckily I had some 1N4148 high speed diodes lying around & even though they’re rated for 200mA continuous rather than the specified part’s 1A rating, the lack of heatsinking on the regulators wouldn’t allow use anywhere near 1A, so this isn’t much of a problem.

Component numbering on the silkscreen isn’t consistent – it jumps from R3 straight to R6! These issues could be slightly confusing for the novice builder, and considering the demographic of the RasPi, could be seen as big issues.

On the far left of the board are the 5v & 3.3v regulators, well placed on the edge of the board in case a heatsink may be required in the future. However the LM317 adjustable regulator is stuck right in the middle of the PCB – no chance of being able to fit a heatsink, & the device itself seems incredibly cheap – the heatsink tab on the back of the TO-220 is the thinnest I have ever seen. Not the usual 2-3mm thick copper of the 5v & 3.3v parts – but barely more than a mm thick, so it’s not going to be able to cope with much power dissipation without overheating quickly.

As the adjustable rail can go between ~2.5v – 10v, at the low end of the range the power dissipation is going to shoot through the roof.

The GPIO connector – this could have been done the other way, at the moment the ribbon cable has to be twisted to get both the Pi & the GPIO board the same way up. Just a slight fail there. See the image below

The power rails are not isolated out of the box – there is no connection between the 5v & 3.3v rails & the Pi’s GPIO, but the GND connections are linked together on the board.

Getting the ribbon cable through the hole in the ModMyPi case was a bit of a faff – the connector is too big! I had to squeeze the connector through at a 45° angle. The case is also remarkably tight around the connector once it’s fitted to the board – clearly the designers of the case didn’t test the an IDC connector in the case before making them!
Everything does fit though, after a little modification.

Here is the unit all built up with the case. The top cover just about fits with the IDC connector on the GPIO header.

More to come once I get some time to do some interfacing!

Posted on October 4, 2011 by The Engineer — Leave a comment

LED Lighting Part 1

Here I will document progress in replacing standard halogen MR10 lights with LEDs.

These units are from TruOpto, available through Rapid Electronics in the UK. They are 3W total, from 3x 1W emitters on an aluminium back plate.

Here is the LED attached to a heatsink for testing purposes – these units dissipate nearly 2W in heat at full output.

As the lights are to be run from a 12v battery bank, for simplicity a master regulator is required to provide a stable 11.4v rail for LED supply.

I have used a Texas Instruments part – PTN78020WAH. This is a 6A capable adjustable regulator module.

The LED lights are to be fully dimmable – the low voltage PWM dimmers are in progress of being built.

Posted on May 23, 2011 by The Engineer — Leave a comment

PSi 150 Power Inverter

This is a small 120W power inverter, intended for small loads such as lights, fans, small TVs & laptop computers.

End cover of the unit, 12v DC input cord at the top, power switch & indicator LEDs at the bottom.

Opposite end of the unit, with the standard 240v AC 50Hz Mains output socket.

Cover removed from the top of the unit. Main power transformer is visible in the centre here, MOSFET bank is under the steel clamp on the left, the aluminium case forms the heatsink.

On the right is a KA3525 switchmode PWM controller & on the left is a LM324N quad Op-Amp IC. The buzzer on the far left is for the low battery warning.

PCB removed from the casing, with the MOSFET bank on the right hand side. Two potentiometers in the centre of the board tweak the frequency of the switcher & the output voltage.