EEVblog Electronics Community Forum
Products => Other Equipment & Products => Topic started by: mikeselectricstuff on January 02, 2012, 10:52:51 am
-
A switchmode mains supply with one transistor.... guess how well it performs...?
Crappy PSU analysis (https://www.youtube.com/watch?v=T88ej64aXUM#ws)
-
A killer PSU, ftw. :P
Alexander.
-
I was pleasantly surprised at how bad it actually was! :o
Dave.
-
My guess would be that, to save money, the manufacture reused a plastic case made for some other unit.
Either that or they ran out of the correct case so used whatever they had available
So it ended up with the wrong specs on it.
That would also explain why the pcb doesn't fit very well and doesn't use the proper guide rails in the case.
Sadly all they really care about is that it has enough current for the product they ship it to run.
-
Nice dissection.
Such Chinese power supplies regularly show up in the weekly reports of the EU's "rapid" alert system for non-food products (RAPEX). And you can be sure when even the EU bureaucrats manage to regularly pick up some that the problem is actually epidemic. The criminals really don't care if they kill people.
-
Years ago I designed some 2-3 W switching power supplies using 3 transistors to power panel meters and controllers, and they were light years better than this one. The industrial cost was about 20 cents higher (not accounting for filter components needed to pass emission tests).
I don't know how much the manufacturer ask for such an horror, but it's too much, in any case...
I have some Christmas tree LED lights here, that are built with the same mentality: Next week I will dissect one of them and post the results.
-
Sadly all they really care about is that it has enough current for the product they ship it to run.
In this case, not even that. It literally will do little more than light the LED on the hub.
-
Sadly all they really care about is that it has enough current for the product they ship it to run.
In this case, not even that. It literally will do little more than light the LED on the hub.
oh, my mistake, i miss-heard that bit as powering the hub leds normaly.
-
Mike,
Thanks for opening my eyes to the crap PSU's that come with some Cheap products. I just gutted a China sourced PSU that recently arrived with a DVB-T adapter and what I found didn't impress me either. The SMPSU uses more components and very slightly better design than your unit but the isolation provided by the transformer still looks suspect.
I will be going through my "cheap & cheerful" plugpack PSU's and giving my bin a good feed ! Unfortunately many small SMPSU plugpacks have welded cases so are very difficult to inspect. I will look at the cheap far East sourced units with great suspicion in the future !
Thanks again for a great video.
-
Thanks for opening my eyes to the crap PSU's that come with some Cheap products. I just gutted a China sourced PSU that recently arrived with a DVB-T adapter and what I found didn't impress me either. The SMPSU uses more components and very slightly better design than your unit but the isolation provided by the transformer still looks suspect.
At least the regulator of your board has a feedback node.
-
A switchmode mains supply with one transistor.... guess how well it performs...?
Wow what a P O S!
Thanks for sharing. I was thinking of maybe re-using 1 or 2 we got with USB hubs at work, but you've cured me of that idea.
-
The more interesting fact is that usually this cheap hubs will connect their external power supply parallel to the USB supply of your PC, frying your PC if the power supply is something like in this topis (or worse).
This product remebers me about car starting cables we bought some time ago, thick like thumb when looking the insulation outside and inside there is 2-3 strands of speaker wire.
-
Thanks for opening my eyes to the crap PSU's that come with some Cheap products. I just gutted a China sourced PSU that recently arrived with a DVB-T adapter and what I found didn't impress me either. The SMPSU uses more components and very slightly better design than your unit but the isolation provided by the transformer still looks suspect.
At least the regulator of your board has a feedback node.
And a fuse. Almost a high-end PSU by chinese standards.
-
Excellent dissection work on the transformer. ;D
We should have expected this given the "quality" of some of the Chinese bench equipment being posted here. At least things like the Atten hot air station and some of the bench PSUs are repairable to a point of being both useable and safe. This unit is not even worth saving for the parts.
-
Like I commented under the video, you get what you pay for.
I am not shocked about performance of the unit - because we all saw how little there is to actually do anything.
I'm shocked to see how much power is lost in the cable.
It doesn't look that thin, and I have similar looking cables going to motors in my little robots. Each engine draws good 0.5A when busy - and I don't see much of a drop on em.
What I am trying to say, is that the cable might be looking solid - but it must be made of pure quality materials too.
-
It probably would've been able to do its rated output with beefier components (like the small-signal glass body diodes(!) replaced with 1N4001s) and wire. It's just a simple oscillator circuit (like a Joule thief) but with a secondary step-down winding.
-
It probably would've been able to do its rated output with beefier components (like the small-signal glass body diodes(!) replaced with 1N4001s) and wire. It's just a simple oscillator circuit (like a Joule thief) but with a secondary step-down winding.
Maybe but wthout feedback, if it could do 1A at 5V, the n-load voltage would be excessively high
-
I'm a big advocate of using 5 volt cell phone chargers to power circuits. I can usually pick these up at the local thrift store for a buck a piece which is far cheaper than I can build anything and the name-brand ones seem to be well engineered.
I'll working on a project where I need to embed a 5-volt supply, so I looked at disassembling the wall wart and using its circuit board inside my enclosure. I took several of these chargers apart to evaluate the feasibility of doing this. What really surprised me was comparing a number of Samsung chargers. All in the same enclosure, all the same ratings, and no two boards alike! Here are some of the pictures I posted at Digital-DIY.
The First Victim
(http://digital-diy.com/forum/download/file.php?id=1719)
(http://digital-diy.com/forum/download/file.php?id=1720)
(http://digital-diy.com/forum/download/file.php?id=1721)
(http://digital-diy.com/forum/download/file.php?id=1722)
Looks like it's probably a reasonable circuit and it's a well-made board.
The Second Victim
From outward appearances, this should be the same charger.
(http://digital-diy.com/forum/download/file.php?id=1730)
(http://digital-diy.com/forum/download/file.php?id=1731)
(http://digital-diy.com/forum/download/file.php?id=1732)
The circuit board and circuit are completely different.
The Third Victim
Again, the same outward appearance.
(http://digital-diy.com/forum/download/file.php?id=1733)
(http://digital-diy.com/forum/download/file.php?id=1734)
The Fourth Victim
This is another Samsung charger in a slightly different package.
(http://digital-diy.com/forum/download/file.php?id=1735)
(http://digital-diy.com/forum/download/file.php?id=1736)
Different circuit boards and different implementations. They'll all different, but all seem to be nicely built.
As a comparison, here's a Nokia charger. I should note that the Samsung chargers are screwed together and can be opened without much trouble. The Nokia charger was glued together and required a brute-force approach to open it. After looking at the circuit board, this may be more than a coincidence!
(http://digital-diy.com/forum/download/file.php?id=1737)
(http://digital-diy.com/forum/download/file.php?id=1738)
-
I bet that last one was made by neither Nokia or Astec.
-
Of the first two, one was made in China and one was made in Korea. Therefore I am not surprised they are different inside. Each was most likely made by a different company to a different design.
-
The first 2 have different model numbers too. I can't make out the numbers on the rest.
-
i have this thing lying around, took apart, exact circuit as shown. mike said about fuse resistor, its actually a bleed resistor for the half bridge cap. yes its using half bridge rectifier. isolation is through the small (hi freq?) transformer, but not sure inside i havent watch mike's video to the end. fuse? i think the whole circuit is a fuse, or some random (weakest) component.
about the half bridge rectifier, the adapter only pulling power when the mains is on +ve side, zero load on -ve. my recent testing using Uni-T UT202 clamp Aac meter registered different current reading from the mains when using between full and half bridge rectifier, same switching volt-regulation circuit, same Adc output reading, same transformer, only different bridge. half bridge showing less efficient power from calculation/comparison, i'm not sure why though, my circuit (different bridge will give different result/output/efficiency?) or the clamp meter that cannot register half loaded mains correctly.
i dont see any problem with the china adapter circuit, in fact i like the design... simple, i would say this is a cheap PS, not crappy (what do you expect from a small and light transformer and cheap product?), or maybe i'm on the low standard side :P the only ridiculous thing is the spec written on the product, mine showing INPUT 220-240V 50Hz 350mA, OUTPUT: 5.1V-5.5V, 500mA±50mA, its like the adapter will draw min 5.1V which proved to be a bullshit from Mike's video. but mine is more humble on the current spec, at least it can do 700+mA on short circuit and not blowing to the face right away :D
just checked the noise, its around 50-200mV on open circuit and 1Vpp at short circuited iirc (i've pulled the output wire apart unintentionally :P cant do the retest again for more accurate figure, late nite GTG)
-
about the half bridge rectifier, the adapter only pulling power when the mains is on +ve side, zero load on -ve. my recent testing using Uni-T UT202 clamp Aac meter registered different current reading from the mains when using between full and half bridge rectifier, same switching volt-regulation circuit, same Adc output reading, same transformer, only different bridge. half bridge showing less efficient power from calculation/comparison, i'm not sure why though, my circuit (different bridge will give different result/output/efficiency?) or the clamp meter that cannot register half loaded mains correctly.
If the clamp meter does not report the true RMS value, you should expect it to be wrong for non-sinusoidal currents like those drawn by rectifier/smoothing cap combinations, switch-mode power supplies or half-wave rectification.
-
I'm a big advocate of using 5 volt cell phone chargers to power circuits. I can usually pick these up at the local thrift store for a buck a piece which is far cheaper than I can build anything and the name-brand ones seem to be well engineered.
I'll working on a project where I need to embed a 5-volt supply, so I looked at disassembling the wall wart and using its circuit board inside my enclosure. I took several of these chargers apart to evaluate the feasibility of doing this. What really surprised me was comparing a number of Samsung chargers. All in the same enclosure, all the same ratings, and no two boards alike! Here are some of the pictures I posted at Digital-DIY.
Is the nokia also a one transistor power supply? ??? Or does it have surface mount components on the back?
-
Is the nokia also a one transistor power supply? ??? Or does it have surface mount components on the back?
What you see is the entire circuit.
-
Is the nokia also a one transistor power supply? ??? Or does it have surface mount components on the back?
What you see is the entire circuit.
It's really strange: there is no trace of mains filtering components.
The Samsung chargers in the other photos have at least a bridging capacitor from secondary to primary (the blue one), which in my experience is mandatory to pass CE testing with ungrounded power supplies.
I cannot believe that ASTEC, a reputable power supply manufacturer, could made this unit...
Are you sure it is not a fake one?
Best regards
-
I stand corrected on the Nokia supply. There are SMT components on the back of the board include a couple transistors of some type and maybe 15 resistors and capacitors. Sorry I did not notice these before.
A curious feature of the board is a bunch of solder dots where there are not component leads to be soldered. Look at the bottom center and upper right corner in the picture below.
-
A curious feature of the board is a bunch of solder dots where there are not component leads to be soldered. Look at the bottom center and upper right corner in the picture below.
Presumably intended to reduce trace resistance by adding solder. The skinny serpentine tracks on the secondary side are unusual and presumably act as fuses.
To me the most curious feature is the resist free triangles which look like the are intended to encourage flash over across the transformer isolation. Some one have a justification for those?
-
I stand corrected on the Nokia supply. There are SMT components on the back of the board include a couple transistors of some type and maybe 15 resistors and capacitors. Sorry I did not notice these before.
A curious feature of the board is a bunch of solder dots where there are not component leads to be soldered. Look at the bottom center and upper right corner in the picture below.
Hmm... Still crappy anyway, I don't see any feedback loop.
-
The Samsung chargers in the other photos have at least a bridging capacitor from secondary to primary (the blue one), which in my experience is mandatory to pass CE testing with ungrounded power supplies.
It is not mandatory, and there is no CE testing. The CE mark is a self-certification that you claim the product is in line with all mandatory regulations. Of course, the fun starts with figuring out the relevant regulations first.
When it comes to emissions, there is also no regulation how you keep your device below the required levels, just that you have to. A Y cap is just the easy way out. But you find high quality SMPS' without a Y cap where they manage to keep emissions within the required levels.
-
i dont see any problem with the china adapter circuit, in fact i like the design... simple
simple design will pay. you should be aware of power factor correction which is poorer with this "simple" design. this will effect the line in the viccinity and interfere with other devices. with no optimization, this smps like design will probably have higher switching loss, which is not very good... source=wiki
-
I've got a docking station for my phone that has the power supply connected in parallel too. What's the proper way to do it? Can I just disconnect the USB V+ pin and power the docking station exclusively from the external power supply?
-
Here's a vid showing a more acceptable PSU
Inside a reasonable quality Chinese PSU (https://www.youtube.com/watch?v=1F3XlFI1JBo#ws)
-
You forgot to test it under load and do all the measurements :P
It looks more solid, but does it perform any better ?
-
You inspired me to pick up a few from the local charity shop. I looked for screw held cases and came away with a 5v/1amp switcher (Audiovox branded) that holds to 4.5v at 800ma and a couple of linear supplies, one 9V at 500ma which consists of a rectifier, cap and xformer and a fairly hefty (1kg!) wall wart which outputs 15v @ 1650ma with a regulated output. All that for $2.50... ;D
-
Prior to this thread, I was browsing my usual ebay equipment suppliers store and noticed he was selling 140 (yes 140) 240V to USB 5V 1A adapters with a start price of GBP9.99 + free postage. they look a decent type from the outside. The catch..... the 3 pin UK plug part that attaches to them is missing. As a project part the lack of the plug is of no consequence so I placed a low bid.
I won 140 USB power supplies for GBP9.99 :o
What the heck I am going to do with 140 of the little blighters, I have no idea ! I will take one apart to see how well it is made. If they are decent they may have been a bargain.... if not...oh well you don't win them all ;D My wife will (not!) be pleased when a large box of 140 little boxes arrives ;)
-
I won 140 USB power supplies for GBP9.99 :o
I have 20 or so small project boxes sitting on the shelf. They came as a lot of 25. It was cheaper to buy the lot than a single one for a particular project :)
Regarding USB power supplies. I had an opportunity to look inside an Atten AT1501D power supply for mobile phone repair http://www.atten.com.cn/ShowProduct.aspx?G=2&productId=739 (http://www.atten.com.cn/ShowProduct.aspx?G=2&productId=739)
Well, not too pretty.
A steel case, but no PE/Ground connection, just a two wire L and N mains cable, permanently attached to the unit. And nothing inside the unit that looked remotely like double isolation. A tiny transformer 220V (not 240V) primary and 20V and 10V secondary windings.
A funny mixture of analogue display (current) and digital display (volts), although the analogue current display might help for some repair jobs.
An LM723 for the adjustable output, with an 2N3055 in TO3 bolted to the case without an additional heatsink. The USB 5V might come from an 7805 inside the case. A heatsink and a cap, supposed to be well cooked by the 7805 due to its proximity to the 7805, prevented me to see the exact type. A Holtek IC (OTP MCU?) in a DIP socket (luxury ;) ) to drive the voltmeter display and to drive (luxury, luxury ;) ) a beeper. No idea if/when the beeper beeps.
The pot for the voltage adjustment was a single turn pot. The output range can be switched between 0-5V and from 0-15V. I.e. there isn't much granularity in the 0-15V range. I think I spotted a protection diode at the output.
USB A socket on the front to power the mobile phone from the power supply, and a USB B socket on the back is supposed to go to a PC. The USB data lines are supposed to be connected between the back and the front, while the power to the front is supposed to come from the PSU. Although I can't rule out that the PSU power was just connected in parallel with the USB power from the PC.
A strange little power supply, and the missing PE/Ground connection without double isolation is IMHO a show-stopper.
-
http://www.pcbheaven.com/userpages/how_cheap_is_a_power_supply/ (http://www.pcbheaven.com/userpages/how_cheap_is_a_power_supply/)
via HackADay
Makes you think a bit :P Just because a power supply "works" doesn't mean much.
-
Don't get confused with the phone chargers.
Those phones chargers are designed to do what they have to do, no matter the number of transistors or ICs on them.
Some of them are tested under temperature chambers and more functional test before they go to mass production, even if third companies are doing the manufacturing.
Of course the cell phone companies are always looking for ways to reduce cost and still get the same results, that is why some designs have only one transistor. Most of the beauty happens inside the phone anyways.
-
As a comparison, here's a Nokia charger. I should note that the Samsung chargers are screwed together and can be opened without much trouble. The Nokia charger was glued together and required a brute-force approach to open it. After looking at the circuit board, this may be more than a coincidence!
(http://digital-diy.com/forum/download/file.php?id=1737)
(http://digital-diy.com/forum/download/file.php?id=1738)
That adapter may look like a fake, but several genuine Nokia adapters I have taken apart look very similar. Unlike most adapters, they are basically current sources with voltage limit, which is very useful for charging batteries or powering LED arrays.
-
My box of 140 USB power supplies arrived today :)
Wife wasn't too chuffed to see them :(
These power supplies are the type that have a common power block and optional connectors to suit various countries, so making the unit a universal travel charger. I paid GBP9.99 inc postage for a box of 140 as they are missing the connectors. Thats 7p each :o I wasn't expecting much but they have turned out to be quite interesting. The unit produces 5V @ 700mA according to the label but I have not tested one yet.
The units look well put together and helpfully have tri-wing screws holding the case together instead of the welded construction so common these days. Upon opening the unit I was horrified to see a single transistor sitting amongst several other components.......but I jumped to an incorrect conclusion. Upon turning the PCB over I found a cluster of SMT components including a 'TH2267.1' 8 pin chip. I believe this chip to be the switching controller for the transistor. The switching transistor drives a decent looking little ferrite transformer. The transformer appears to have some sort of feedback winding that goes to the TH2267.1 chip (for output regulation?).
At the secondary of the transformer a single diode is used to rectify the ac and this is smoothed with a 470uF capacitor.
I did a quick search for the TH2267.1 and had quite a surprise when I found that many people wanted the same information but the chip is custom and Melexis Will not release any information at all, including its function ! See below from the Melexis FAQ site:
FAQs by Topic
TH2267.1
Q: Do you have anybody datasheet or same information about IC "TH2267.1" 8pins, SOP-8.
A: The requested datasheet can not be provided to you because the TH2267 is a custom specific product and therefore not open for everybody like our ASSP's listed on our web site.
Q: Why won't you guys make a datasheet available for this TH2267.1 chip. It's used in lots of Nokia cellphone chargers and a bit of info would help a lot of people who want to repair these things!
A: For customer specific parts like this one we have a contract with our customer in which is stated that we are not allowed to disclose any information without their permission.
Q: Is it possible to inquire about the approximate price of your product TH2267.1?
A:This is also confidential information which we cannot disclose.
Q:Can anybody please tell me what this IC TH2267.1 is doing at all?
A:We are not allowed to even disclose the function of the IC.
Finally the design looks decent enough, but the PCB track layout looks a little wild for my liking. The SMPSU case has several safety markings and indicates that the unit was manufactured in Germany. I was hoping to modify some of the 140 units for other usefull voltages for use in little projects. It would appear the TH2267.1 may have burst that balloon :( I will try to work out its functionality via teh circuit around it. I have a few that I can risk in order to experiment on the design ;)
All good fun ;D
Apologies for the overexposed pictures, they were taken in a hurry.
Aurora
-
At the secondary of the transformer a single diode is used to rectify the ac and this is smoothed with a 470uF capacitor.
What's the purpose of the four diodes then? They're arranged as you'd expect in a bridge rectifier.
-
The bridge is on the primary side and converts the 100 to 240 VAC to high voltage DC.
The DC is chopped by the switching transistor, and sent to the primary of the transformer.
The single diode is on the secondary side of the transformer.
The single switching transistor and single diode on the secondary strongly suggest a flyback topology. There is no need for more diodes - would not improve anything.
Two switching transistors and a half or full bride on the secondary would suggest a forward converter topology.
-
I did a quick search for the TH2267.1 and had quite a surprise when I found that many people wanted the same information but the chip is custom and Melexis Will not release any information at all, including its function ! See below from the Melexis FAQ site:
Sounds like some reverse engineering is in order! It's fair enough though, if a customer has a part custom designed and made for them then it is proprietary. The only company that could possibly disclose information about the part is Nokia.
-
I'd love to see that chip decapped and put under a microscope. It shouldn't be too complicated so this and a bit of probing around with a scope could tell really a lot about its function.
-
Data sheet is the first link on Gooooogle.
th2267.1 (http://www.scribd.com/doc/10271023/th22671#)
-
DOH !
How did I miss that ! :-[
Many thanks
Aurora
-
oPossum,
I have just taken a good look at the data sheet you referred me to.
Thanks to you I now have all the information that I need to make use of these little SMPSU's.
Unlike the SMPSU detailed at the start of this thread by the OP, these German SMPSU's use a sophisticated ASIC with all manner of control and safety features....one extreme to the other I think :) The transformer has a feedback winding and a separate winding to power the TH2267.1 after the start up sequence. Quite a decent smpsu after all :D
Thanks again for making me realise I had missed the important datasheet. With the FRIWO part no. I was able to find the original source of the datasheet. Obviously FRIWO were willing to release the information when the manufacturer was not able to do so due to an NDA.
-
No need to get excited about an obscure custom part; switcher ICs are very common. One company "Power Integrations" has dozens of different product lines for flyback, forward, non-isolated, low parts count, etc. all available through Digi-Key with very good datasheets. I'm currently weighing my options for a project, it's between design my own around an IC (fun, expensive, dangerous, hooray) or buy a box of plug packs on ebay like Aurora did :-)
EDIT: two of the pictures on Jon Chandler's post on the second page have PI chips, for example
-
... I'm currently weighing my options for a project, it's between design my own around an IC (fun, expensive, dangerous, hooray)...
I considered building a supply for a project, but most of the chips depend on a specialized transformer. I found some designed for specific ICs, but they're about $15 each in small quantities. Makes recycling $1 supplies from the thrift store look pretty good.
-
I wouldn't be surprised if the TH2267.1 was based on some other IC; there aren't that many ways to make a switcher IC and put it in an 8-pin package. Look around enough and you might find another with the same pinout and functionality.
-
A switchmode mains supply with one transistor.... guess how well it performs...?
Apart from only being able to produce a fraction of the required power and shoddy construction I don't think there's anything wrong with using a cheap unregulated blocking oscillator fly-back converter as an SMPS. This is obviously the wrong application for it, a stable 5V supply is required for USB devices.
Did you actually count the number of turns on different windings?
You also mention about the temperature rise of the transistor and the power consumption. If the SMPS really used as much power as your meter says, the transistor would've melted long ago. I think it's just measuring the apparent power.
I thing the design is worth copying and would be a good way to power LEDs from the mains. Obviously the issue of isolation needs to be addressed, unless the LEDs are going to be in a fully insulated enclosure.
As others have mentioned another possible application is battery charging where there's another regulator. I've seen an unregulated mobile phone charger SMPs before which does have proper isolation between the mains and DC power out.
I've toyed with the idea of building a similar PSU but I wouldn't know what kind of tape to use for separating the primary and secondary. Is there a standard for this? I've dismantled apparently well-made SMPSes before and noticed they all seem to use a similar kind of yellow sticky tape between the primary and secondary.
-
Most transformers use Mylar tape for insulation.
-
Sorry to bump an old thread but I didn't know where else to post this.
On the subject unregulated SMPSes I've found an example on the TC33167 datasheet.
It looks like parts from an electronic halogen transformer could easily be used to make this circuit which could be used as an alternative to an unregulated supply using a bulky transformer.
http://www.icwic.com/icwic/data/pdf/cd/cd057/Switching,%20DC-DC%20Regulator,%20Controller/a/85248.pdf (http://www.icwic.com/icwic/data/pdf/cd/cd057/Switching,%20DC-DC%20Regulator,%20Controller/a/85248.pdf)
-
I bet that last one was made by neither Nokia or Astec.
I have a genuine Nokia charger I don't need anymore, do you guys want me to take it apart?
-
also have on my hands a few old chargers of various sources and brands that i could probably give a teardown of,
-
No need to get excited about an obscure custom part; switcher ICs are very common. One company "Power Integrations" has dozens of different product lines for flyback, forward, non-isolated, low parts count, etc. all available through Digi-Key with very good datasheets. I'm currently weighing my options for a project, it's between design my own around an IC (fun, expensive, dangerous, hooray) or buy a box of plug packs on ebay like Aurora did :-)
EDIT: two of the pictures on Jon Chandler's post on the second page have PI chips, for example
Why don't you use a Micro-USB connector for charging your project? You can supply an adapter easily or have the customer buy their own adapter like many manufacturers are doing.
-
for $7, i just got the "piss poor" Mains to USB adapter/charger and an electronic cigarette. dont turn it on, take it aparght! :P
edit: i believe this kind of thing is produced in humongous and distributed throughout the worldwide. for a single tiny of this, PF and harmonics noise may not be an issue, but imagine tens of thousands of households using it in harmony at the same time, i cant imagine what will happen to the power line... maybe the time has come where smps will take over the world! :P