LW DQGS power supply : review and teardown - Caution:SAFETY RISKS!!!

[ciccio]

This power supply was bought new for the very small price of 50 euros (including VAT) from a dealer in a electronics fair.
He had lots of them, and told me that he was buying them in 100 pieces lots from the manufacturer.

I had urgent need of using it to test some other equipment, so I put it to work immediately after opening the box. It operated without  any evident problem, the only  negative thing was noise from the internal cooling fan, always on.
After the initial hurry, I put it on one side of the instrument rack, and forgot about it.

Some days ago I decided to open it and install a thermal switch to control the fan, since it was mostly used at low current, and I though that the fan could be off most of the time, for a better environment it the lab.

The inside of the unit was full of surprises, some off them  really nasty, including some wiring faults that presented real safety hazards.

 The unit: it's a Chinese clone of another Chinese clone…. made by the HK Longwey company. It was really difficult to find the manufacturer's name,  because it is labeled LWDQGS… in any case I traced the company's site:
http://en.hklongwei.com/newEbiz1/EbizPortalFG/portal/html/index.html
They have a big range of simple laboratory instruments, each of them resembling a big-name model.

The unit can supply a nominal 30 Vdc at 5 A max.
A couple of rotary pots (coarse and fine)  sets output voltage, another couple sets current limit value. A pushbutton switch, not well documented in the too simple user's manual,  changes the current limit range from 0 to 5 A (HI) to 0 to 2.5 A (LO).  I see no real use for this…
Two LEDs show  operation mode (CV or CC).
Actual voltage and current values are displayed by two 3-digits large LED displays.
Output terminals are 5-way binding posts, that  do not (at least I believe they do not) respect actual  safety regulations.
Opening the case shows a very large and hefty transformer, that does not show signs of overheating even when heavily loaded, and a lot of cheaply-made, single sided circuit boards.
The dual meters, based on the classic 7107 ic,  are assemble on two boards interconnected by a flat ribbon cable, and there are wires everywhere….
I cannot understand the economics of this choice: a smaller dual sided board could handle both meters, and the final cost should be lower, even in Hong Kong or China.
The main board shows three 8-pins ICs and more than 10 transistors.
I could not  see the details because the wiring did not allow to reach near it, but two relays were obviously used to switch the transformer's  secondary windings, to reduce power dissipation.

The heath sink was a simple piece of sheet aluminum,  and the two power transistors were custom labeled units, coded:  LW H2388C  LONG WEI MEX.
This make me believe they have a great use of this parts…

The soldering to the power transistors pins was abysmal: see the photos.
One wire connecting the two collectors lugs was not well soldered, and it  detached  with a simple pull. If this happened  in operation, the transistor (or both) could be easily damaged.
Wires (including mains) from unit's rear to unit's front were  placed inside a slot in the heat sink, that had sharp corners that could damage the insulation, and that can become really hot.

The wire from the protective hearth pin of the IEC power inlet was not soldered (or at least the joint was detached by a simple pull), but  kept in place by heat-shrink tube that was kept in place  by a blob of hot melt glue (hot glue was everywhere in the unit's inside).
I must say that I had verified continuity from mains-in PE terminal to instrument case ground before first use of the unit, and it tested OK: maybe the heath shrinking was helping continuity.

All  mains-connected wires did not respect any safety regulation, in terms of colors and wiring/soldering method.  Live, neutral and PE were all white….

I ended up with a new ground wire (yellow-green) with adequate heat-shrinking, and some mechanical protection for the other wires.

After reassembling the unit (whit the thermostat-controlled fan and the modified wiring), I carefully tested it with loads from 1 to 3 A .
The meters read correctly (at least +- 1 LSD)
Ripple was about 1 mV p-p (see photo) at any voltage and current, but increased tenfold in constant current mode.

In searching the Net, I've found a report on a Swedish site:
http://www.elektronikforumet.com/wiki/index.php/LW_PS-305D

I translated it with Google, but not every thing was clear to me.
I understand that the tester measured a large overshoot when powering down.
I've  recorded many on/off  sequences with my DSO, and  they seem correct (no overshoot) at any output voltage and with/without load.
Maybe the unit in Sweden  had an unconnected wire….

What should I write as the last comment?
This unit is CHEAP, but crap-made.
There are safety risks in actual use.
If you have one,  open it up and check the wiring before further use.
If you don't have one,  DO NOT buy it.

The problem that bothers me is: this manufacturer can supply OEM units with other logos on it, and you can end up with a more expensive, crappy unit, with big name on it…

[RCMR]

Yes, I have one very similar and it is almost as bad.

I could not figure out why the voltage and current meters on mine would start acting crazy when the supply was delivering any reasonable load so I threw the scope on the output and saw that there was *no* smoothing.

A quick inspection of the board showed that the main filter cap only had one wire soldered to the PCB (someone asleep on the job that day).  I soldered the other leg and that fixed the problem.

However, the mains wiring was very dangerous.  Thin wires very poorly soldered to the mains switch with no strain relief and not even any heatshrink to cover the active wires.  One of the bolts holding the power transistor to the heatshrink had been cross-threaded so was jammed about 2mm out of the hole with the transistor not being held against the sink itself.

I tidied it up and it works okay now but I pity anyone who doesn't properly inspect and fix up these cheap Chinese supplies.  I suspect it will not be long before we read about the first electrocution as a result of some neophyte using one.

[Bored@Work]

I suspect it will not be long before we read about the first electrocution as a result of some neophyte using one.

Unfortunately media doesn't report about such things.

[saturation]

Thanks for the details teardown and review.  Hopefully others will look into their units, or even the look a likes.

This looks just like the Mastech, but luckily the Mastech brand are better made, as its been around for a while 

I think as a rule, you should always tear down a Chinese brand item, there is always something to remediate.  I hope the parts used are quality enough not to wear prematurely, but I would burn in your unit anyway to be sure, you never know.

http://www.mastechpowersupply.com/dc-power-supply/linear-power-supply/mastech-regulated-variable-dc-power-supply-gps-3030d-30v-3a/prod_62.html

Your unit vs Mastech:

[ciccio]

From what I see my unit and the Mastech one are the same: I've seen other similar, but they did not have the current range pushbutton.
The plastic panel and the metalwork are the same.
I paid a lower price for  mine....

[saturation]

They are very similar, maybe copies of someone else, but as you see in this photo Mastech has Mastech PCBs [ this is just an example].  Although build quality is better than the one you pictured, I'd still be wary and check inside, as quality control can vary.  Most of the time the Mastech's are ok.  Its about time I revisited my PSU and took photos of its build since its nearly identical to your unit.  Also notable is they are not so crazy about hot glue.

From what I see my unit and the Mastech one are the same: I've seen other similar, but they did not have the current range pushbutton.
The plastic panel and the metalwork are the same.
I paid a lower price for  mine....

[saturation]

Here's a quick look inside my Mastech.  The PCBs don't say Mastech; PCBs all say 'designed by Lys'.  Anyone know what that means?

Generally the build quality is better that the previous photo I showed,  at least they were decent enough to align components, there is no hot glue mess at all, the layout is neater.  I have 2 minor reservations about the layout and made fixes.  What the photos show is the original layout as it comes, out of the box.

The design is c1980s through hole, easy to service if needed.   Note clean layout and alignment of components.  My only concern here are the 2 white wires on the side of the transformer, those are from the hot line of the 120VAC heading to the switch.  I rerouted them and removed the pull tension on them.  I also added more insulation to an exposed lead on the full wave bridge rectifier you see bolted to the floor of the PSU, just behind the transformer.

The very large heat sink is the large aluminum wall you see to the right of the fan; the design really keeps it cool but the fan is noisy.





This view is of the PCB for the LED readouts. You can see the hand soldering and cleanout of the flux.  Hand solder job on this is quite good.



Details of the power controller PCB and close up of the designer's label and soldering quality.



Close up of the power controller PCB and the alignment of parts, compare to the sloppy placement in my previous post with a photo from another Mastech PSU.

[Kiriakos-GR]

the transistor (or both) could be easily damaged.
Wires (including mains) from unit's rear to unit's front were  placed inside a slot in the heat sink, that had sharp corners that could damage the insulation, and that can become really hot.

The wire from the protective hearth pin of the IEC power inlet was not soldered (or at least the joint was detached by a simple pull), but  kept in place by heat-shrink tube that was kept in place  by a blob of hot melt glue (hot glue was everywhere in the unit's inside).

Yes it looks as piss poor implementation, but I can not share you opinion about Caution:SAFETY RISKS!!!
This is a tiny PSU, and the idea about melted wires and possible damage of the insulation is just a wild dream.

In the worst scenario it will just burn out it self.

Currently I feel bored to open up and take pictures from my 1000$ PSU.
But for 50 EUR, you got a treasure over there..   
The parts of it , in the free market it would cost four times up the price that you payed for it.

And no, I would not call this small device as a mini time bomb.



 

[ciccio]

You are correct,  Kiriakos-GR:
if the "glued with solder" wire connecting PE terminal  to chassis is detached, the unit will not burst in flames..
if the single insulation of the AC HOT wire, that was touching a very sharp corner (so sharp it could cut my  skin..) was cut during transport from China to here, my RCD would trip..
but if both events occurred, I could be "slightly electrocuted".

My intention was to show that for such a low price you could get a fairly performant power supply, that could cost one or two  euros more and be wired respecting the "good practice".

It was a very simple work to open it and do some rewiring....

Saturation  has shown another version  of the same unit, with better overall wiring, so it can be done...

[alm]

Yes it looks as piss poor implementation, but I can not share you opinion about Caution:SAFETY RISKS!!!
This is a tiny PSU, and the idea about melted wires and possible damage of the insulation is just a wild dream.

In the worst scenario it will just burn out it self.

I agree that wires are not likely to melt at these kind of power levels unless they're way too thin. Insulation can melt quite readily if the heat sink gets hot enough, though, I don't see how the low power makes this less likely. It depends on the dissipated heat times the thermal resistance between the heat sink and ambient. If the heat sink is on the small side or air flow is insufficient, it may get quite hot. A TO-220 device without heat sink will burn you at a few Watt, for example.

I wouldn't be so worried about the thing blowing up, although catching fire might be an issue if there is no fuse between the wires and the mains connector, you're unlucky and live and neutral short, and the wires are too thin to be protected by the mains fuse (which is designed to protect the wiring to the socket, which tends to be a much larger gauge).

I would be the most worried about shocking the user, especially with the lousy ground connection. A live wire with damaged insulation could come into contact with the case, which is floating since the ground connection is unreliable. The case will now be at mains voltage, possibly with a fairly low output impedance.

[Uncle Vernon]

but if both events occurred, I could be "slightly electrocuted".

Slightly Electrocuted? Like Kennedy was slightly shot?  Either can result in being slightly dead. 

Just like being shot are no degrees of electrocution. You are either electrocuted or not. You cannot be slightly shot and you cannot be slightly electrocuted. If there is any risk of contact with main then the risk of electrocution is real.  The is nothing about this PSU design (or lack of it) that suggests inadvertent contact with live conductors would be via a high impedance path.

[The_Penguin]

Coming in late, but after reading this, I think I'll crack open my supply (YiHUA YH-305D) that looks a lot like the ones in this thread.
Got it new for about $90.00 from a Canadian supplier.  Haven't used it much, but seems ok. 
Will report back once I inspect it.

[m1gry]

I found this page by googling "YH-305D power supply". Ironically I think that I have previously visited it! I saw the power supply on UK Ebay, wait for it, for £199.99 (4 available / 71 sold). Judging by the comments it can be picked up much much more cheaply if you "shop around". I was looking for a lab type PSU. I currently have a Wier PSU which after a few simple repairs is working fine, but only goes to a couple of amps at 30V.

I also viewed a very detailed set of YouTube videos about the Rigol PSUs. They are a little pricey, but seem really good quality, and the colour display is very impressive.

--Gary

[metalphreak]

The cases for these kinds of power supplies are sold by the boatload to all manner of manufacturers in china. Then they chuck whatever electronics they desire inside, so you really can't judge a book by its cover here. Some might be decent quality and at least safe. Others can be terrible with dodgy construction

I believe a lot of the cheaper chinese stuff can be great value for hobbyists who have no need or desire to pay $500 for a power supply. The problem is weeding out the absolute rubbish crap from the decent bargains.

I'd encourage everyone who takes a gamble on them to share their experiences. Your thoughts, impressions, teardowns as well as model, brand, and who/where you bought it.

[gxti]

I got a Mastech HY1803D, 20 volts 3 amps, about a year ago. I forget exactly how much it cost but it was in the 100-150 USD range. For the price I was expecting something like the original poster's unit but I was pleasantly surprised -- neatly wired, well-earthed, good regulation and little overshoot. It even uses good quality capacitors, Nichicon and UCC if I remember correctly. I haven't really beat on it yet but it more than suits my needs. The only concern is that the power transistor is not adequately cooled to run at max power for sustained periods of time. It uses a passive design with the transistor and heatsink on the back of the unit. I tried running it at full load and just holding a thermocouple to the transistor case it got up to 130 C before I chickened out.

[don.r]

Just out of curiosity, what regulators do these units use? Are they linear or switching?

[gxti]

Just out of curiosity, what regulators do these units use? Are they linear or switching?

You see that huge transformer in the pictures? There's your clue that it's linear. The cheap Chinese power supplies always are linear because they don't require a whole lot of expertise to design and the firms that sell these things are just shoveling them out to make use of whatever parts are cheapest at the time.

[don.r]

Just out of curiosity, what regulators do these units use? Are they linear or switching?

You see that huge transformer in the pictures? There's your clue that it's linear. The cheap Chinese power supplies always are linear because they don't require a whole lot of expertise to design and the firms that sell these things are just shoveling them out to make use of whatever parts are cheapest at the time.

Not necessarily though. I was wondering if they use a DC-DC buck or a linear transistor after the transformer. The reason I ask is that I was entertaining the idea of converting an old audio amp into a variable bench supply and looking for DC-DC regulators. They make great starting points as they have big transformers and beefy electrolytic caps as well as meaty heatsinks.

[alm]

Not much point in combining a large mains transformer with a buck converter in a commercial design: (most of) the complexity of a switching power supply with (most of) the costs and weight of a linear power supply.

[don.r]

Not much point in combining a large mains transformer with a buck converter in a commercial design: (most of) the complexity of a switching power supply with (most of) the costs and weight of a linear power supply.

I guess that's why they have the fan and heatsink then for the linear burnoff. It would still be interesting to know which linear controller they use.

[gxti]

Here's the schematic of a HY1803D: http://partiallystapled.com/~gxti/trash/HY1803D.pdf

The only monolithic regulator in it is used for just the control circuitry. The actual power path is entirely op-amps and transistors. U2B is the current limiter, U3 is the voltage regulator, and U2A seems to control the relay cut-over (changing transformer taps for lower voltages).

I haven't fully dissected the schematic but at one point I was thinking about enhancing it with a micro-controller to get encoder pots with better precision and a proportional response (finer control as the voltage gets lower). Having push-buttons for common voltages would be cool, too.

[don.r]

Very interesting. Thanks for that.

[ZeroKool]

The cases for these kinds of power supplies are sold by the boatload to all manner of manufacturers in china. Then they chuck whatever electronics they desire inside, so you really can't judge a book by its cover here. Some might be decent quality and at least safe. Others can be terrible with dodgy construction

Does anyone know which chinese manufacture made these clone cases? Just want to buy 3 cases off from them and put my own design in it for 2 of my friends and 1 for myself.

Thanks