Author Topic: Reverse engineered open-frame 12-24v -> 100-250v DC/DC converter (Read 7247 times)

robertbaruch · « **on:** May 18, 2016, 11:55:51 pm »

So I reverse engineered a small power supply I got from China. It's a boost converter. It seems to work pretty well (I tested it at 170VDC @150mA out, efficiency was around 75%).

At first I figured the chip would be one of the usual 8-pin converter suspects. But I can't seem to find a match. I suspect that the chip could be custom, or at least not intended for individual sale. The markings on top of the chip just read "Strong Electronic Technology" in Chinese (the name of the company), same as the markings on the PCB.

There's also a circuit I haven't seen before on the input. It seems that the way it works is, assuming no charge on the input bulk capacitor, when power is applied, current flows through from plus, to gnd, then through the MOSFET's protection diode, through the fuse, and out the negative.

Eventually the capacitor charges up and the gate voltage rises, but even so, Vds still has to be negative, so this thing doesn't seem to do much.

So two questions:

1. Can anyone figure out the purpose of each pin?
2. What's the deal with that MOSFET circuit on the input?

ataradov · « **Reply #1 on:** May 19, 2016, 12:13:33 am »

The MOSFET on the input is a typical protection against polarity reversal.

robertbaruch · « **Reply #2 on:** May 19, 2016, 01:11:28 am »

Yeah, but then why not just use a diode and forget the MOSFET? Seems like one part is better than three...

ataradov · « **Reply #3 on:** May 19, 2016, 01:13:17 am »

Quote from: robertbaruch on May 19, 2016, 01:11:28 am

Yeah, but then why not just use a diode and forget the MOSFET? Seems like one part is better than three...

You eliminate diode voltage drop when powered correctly and have combination of closed diode and transistor when powered incorrectly. It is a win-win, except for the price, probably.

tautech · « **Reply #4 on:** May 19, 2016, 01:23:15 am »

Isn't the controller just one of the venerable UC3842 series or similar? $:-\$

Pin 7 VCC
Pin 8 Vref

robertbaruch · « **Reply #5 on:** May 19, 2016, 02:23:15 am »

Ooh, UC3842 definitely looks like a match, thanks!

tautech · « **Reply #6 on:** May 19, 2016, 02:28:10 am »

Quote from: robertbaruch on May 19, 2016, 02:23:15 am

Ooh, UC3842 definitely looks like a match, thanks!

Don't discount similar IC's like: 3844,3845

robertbaruch · « **Reply #7 on:** May 19, 2016, 02:32:44 am »

Right -- it must be the '43 or the '45 because the others have a minimum Vin of 16v, and this power supply is rated for a minimum of 12v. Probably the '43 because from 12v to 170v needs a duty cycle above 50%, which the '45 cannot do.

robertbaruch · « **Reply #8 on:** May 19, 2016, 02:37:13 am »

Updated schematic.

robertbaruch · « **Reply #9 on:** May 19, 2016, 02:41:04 am »

Also, in case you're wondering, the eBay listing is for "DC12V 24 to DC 100-250V 70W high voltage converter boost step up power supply".

(US listing, USD12.49)(AU listing, AUD17.37)

A nice little supply for a nice little price.

robertbaruch · « **Reply #10 on:** May 21, 2016, 06:34:02 pm »

Someone on Hackaday pointed out that the HY1707 is only rated for VDS up to 70V. Which is disturbing, since the switch needs a VDSS of at least 238V. Going to see what happens if I replace it with an IRF630.

This is why I bought two -- one for experimentation and reverse engineering, and one original.

robertbaruch · « **Reply #11 on:** May 23, 2016, 04:27:05 am »

So I replaced the (switching) HY1707 with an IRF630, put 1.7kohm load on it and... the power supply failed after about 5 seconds of operation. The voltage just drooped and died. As if some part was heating up and going out of spec. Sure enough, the IRF630 seemed to be heating up the heatsink quite a bit. And I heatsinked and thermal pasted the IRF630, too. Then for good measure I put back the HY1707 and, flawless operation. Not as much heat. WTF!

At this point I'm wondering if they are really using an HY1707. I also am puzzled why an IRF630 wouldn't work.

station240 · « **Reply #12 on:** May 23, 2016, 11:05:08 pm »

Gate capacitance is the difference. The higher the capacitance, the more current you need to source/drain to switch the mosfet on/off.

I'd bet the PWM IC isn't rated to high enough current, hence the mosfet takes longer to switch, and gets a lot hotter in the process.

edavid · « **Reply #13 on:** May 23, 2016, 11:10:18 pm »

Quote from: station240 on May 23, 2016, 11:05:08 pm

Gate capacitance is the difference. The higher the capacitance, the more current you need to source/drain to switch the mosfet on/off.

I'd bet the PWM IC isn't rated to high enough current, hence the mosfet takes longer to switch, and gets a lot hotter in the process.

The IRF630 has lower gate capacitance than the HY1707.

OP, are you sure that's what the original part is? How is it marked exactly? Photo?

niekvs · « **Reply #14 on:** May 24, 2016, 05:46:25 am »

The IRF630 has an order of magnitude higher Rds(on) vs the HY1707. Current ratings are also very different. Find a part with similar specs, but higher Vds of 200V or so.

robertbaruch · « **Reply #15 on:** May 25, 2016, 02:32:00 am »

100% sure that the _markings_ on it are HY1707. I'll go look for am equivalent part with higher VDSS.

robertbaruch · « **Reply #16 on:** May 25, 2016, 03:07:26 am »

My bad. I had a suspicion, so I desoldered the transformer. I initially thought it was a single coil, but it turned out to be two. I think this is why they could use a lower voltage switch, because the turns ratio is quite high. I may sacrifice the transformer to find the winding ratio.

New schematic attached.

FlyingHacker · « **Reply #17 on:** May 25, 2016, 03:52:45 am »

Curious: can you build one of these for less than the price in eBay? I find a lot of the time you can't. The parts alone exceed the price, then there's the board, and the labor, etc.

I assume a lot of this is due to counterfeit chips, but economies of scale comes into play too.... Then you have the Chinese government subsidizing the postage for free shipping and it is hard to beat...

...until it bursts into flames

robertbaruch · « **Reply #18 on:** May 25, 2016, 04:47:37 am »

I think it's possible, but not by much. In the end, if the circuit is good, it comes down to time. Since I'm now pretty sure this circuit will last, I don't see why I wouldn't just buy it.

robertbaruch · « **Reply #19 on:** May 26, 2016, 02:06:56 am »

If you know of or suspect an error, please let me know. I've checked the schematic against the board, and it seems okay, but I could always have missed something.

robertbaruch · « **Reply #20 on:** May 27, 2016, 02:46:53 am »

Thanks for the feedback (lol).

* The ISEN pin 3 in the circuit has a resistor and capacitor in parallel to ground, which doesn't look like it could sense any current, right? And you were right -- I missed a resistor that looked like a jumper but wasn't, it is a low-resistance sense resistor. I compared it to a known sense resistor, and they were about the same, 0.015 ohms.

* The feedback to pin 2 does look a bit wonky. I can understand the transistor: it provides higher current (and a VBE drop) from the resistor divider. But then there's an LED from the emitter straight to ground? I triple-checked this, and that seems to be what the circuit is doing.

I've attached the updated schematic. Also I've attached a composite image showing the solder side of the board with the components overlaid on top. Maybe someone can spot something that I didn't? The LED is in the lower left. In the lower left corner, from the bottom, we have a transistor, then above that a capacitor, and then the LED above that.

Fraser · « **Reply #21 on:** May 27, 2016, 10:49:10 am »

I have been intending to buy one of these boost converters for a while. Thanks for sharing your findings with the forum.

Fraser


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Reverse engineered open-frame 12-24v -> 100-250v DC/DC converter (Read 7247 times)

Share me