Mystery power supply

[PeteS in CA]

Re: Mystery power supply

Chances are the base drive design is what is called proportional drive, which feeds back the collector current to provide base current. This means MOSFETs almost certainly would not work without significant modification. For power BJTs a beta of 5-10 at full current is quite typical, not a sign of a poor part. With proportional drive, the base current can be set to 1/10 or 1/5 of the expected full load collector current, ensuring the base current is adeqate to keep the BJT in saturation at that current. OTOH, base current is not part of the transformer primary current, so too high a base current lowers efficiency.

[eccerr0r]

Re: Mystery power supply

Usually you'd need to modify the circuit considerably to use FETs (specifically enhancement MOSFETs) instead of BJTs, the drive requirements are quite different. Especially high side MOSFETs need to be driven quite high to make sure they turn on (usually half bridge use two NPN transistors). A lot of designs add a zener to make sure the gate drives don't exceed breakdown voltage which may also be needed. Even if the transistors blew and my spare parts cabinet only had MOSFETs I'd still go look for BJTs to replace it, unless I really wanted to replace the drive circuit along with them.

The hFE of 8 seems very low but for high voltage operation, it's not that bad. Now I'm going to cheat a bit and assume hFE=beta because I don't want to go through small signal analysis. So, say we want 120W to go through with 240V DC source due to the doubler, only a half amp is going through the collector. To drive 500mA we'd only need a base current around 60mA to saturate - that's no more than 100mW or so burned by the base junction. Now if the transistor is in saturation at most it'd probably be dissipating 1W or so on the collector so really it's not burning way too much to pass 120W, despite a low gain.

I have not seen many high voltage transistors with high gain... when I looked up all the transistor pulls that I have, they all have very low gain. Only the high frequency low voltage, low current transistors have gains of over 100. But these are mostly for amplification where gain is crucial - for PSUs it's just a high voltage switch.

[everell]

Re: Mystery power supply

I replaced the fuse and fixed something.....several months ago. No, I did not put FETs in the circuit.....yet......

[c_hegge]

Re: Mystery power supply

I didn't say this power supply had first rate semiconductors! I am surprised it works at all...

So you've tried it and it works? Nice.

[Behemot]

Re: Mystery power supply

Larger currents than mosfets, for the whole "on"-duration (2SC2553 hFE is just 8-12 ), and lower voltages (just a couple volts - 2SC2553 has a Vbeo of 7v max) - FETs need 10-12v to fully "open up", and just short current bursts, to charge/discharge the gate.

That's corresponding with what I said - both are using different ways to open from their principle.

You need current to flow through bipolar to open. FET on the other hand uses electric field to open, hence you just bring voltage to the electrode. But currents through it are very low as they are only parasite (there is insulation, usually silicon dioxide, between the electrode and the very conduction canal; but no insulation is perfect).

[everell]

Re: Mystery power supply

2SC2553 hFE is just 8-12

Low hFE

I didn't say this power supply had first rate semiconductors! I am surprised it works at all...

[Khron]

Re: Mystery power supply

Larger currents than mosfets, for the whole "on"-duration (2SC2553 hFE is just 8-12 ), and lower voltages (just a couple volts - 2SC2553 has a Vbeo of 7v max) - FETs need 10-12v to fully "open up", and just short current bursts, to charge/discharge the gate.

[Behemot]

Re: Mystery power supply

I don't think they neccessarily have to use bipolars, but shouldn't it be designed for that? I mean, bipolars use current for opening, FET's use voltage, right?

[LDSisHere]

Re: Mystery power supply

Now I'm getting curious.......if I was to remove those 4 amp 400 volt main switching bipolar transistors (C2553) and install a pair of 6 or 7 amp 600 volt FETs, would the switching speed and efficiency improve, or would it blow the crap out of the power supply? Every half bridge power supply I have seen used bipolar transistors, but does that mean that half bridge design requires bipolar transistors?

I have to say I find this question most intriguing.

[c_hegge]

Re: Mystery power supply

You could try it but my guess is it won't work. I think half bridge might require bipolar switchers

[everell]

Re: Mystery power supply

Now I'm getting curious.......if I was to remove those 4 amp 400 volt main switching bipolar transistors (C2553) and install a pair of 6 or 7 amp 600 volt FETs, would the switching speed and efficiency improve, or would it blow the crap out of the power supply? Every half bridge power supply I have seen used bipolar transistors, but does that mean that half bridge design requires bipolar transistors?

[Pentium4]

Re: Mystery power supply

What are the specs on the output rectifiers?

[c_hegge]

Re: Mystery power supply

If it were a cheap L&C I would expect the pwm chip to be a "chip of the year" 2003 or 2005. The pwm chip on this one is KA7500

Sun Pro =/= L&C. they are two totally unrelated companies.

The output rails do not have pi filters, only output capacitor. No coils. I don't know if Sun Pro would stoop this low!!!

They would indeed stoop that low. I've seen one cap with no coils on Sun Pro branded PSUs.

... made by HongKongFlyApart, LLC.

HKFA, LLC would have been my second guess

[everell]

Re: Mystery power supply

Can you tell if it has been recapped?

As with many of these cheap Chinese power supplies, the soldering is of such "quality" it is hard to say what has been done or not done.

[lti]

Re: Mystery power supply

Can you tell if it has been recapped? It seems strange to use such small primary caps and Nichicon HM and what appears to be Chemi-Con KZJ on the secondary.

I think your guess of the manufacturer is correct.

[everell]

Re: Mystery power supply

I think it is probably WhoFlungDung brand made by HongKongFlyApart, LLC.

[Behemot]

Re: Mystery power supply

I am interested too. Eurocase here got many models lately like this one. There is the 2003 PWM, but, they got also similar, more full and even 550 W rated models. They can really do arround 400 W, never stressed it more. Posted one here.

You think only L&С does these? I have seen some which used green transformers, totally different heatsinks, but other than that were almsot same. They looked like some chineese CWT scam made according to this topology…

[everell]

Re: Mystery power supply

Leadman or Sun Pro maybe?

If it were a cheap L&C I would expect the pwm chip to be a "chip of the year" 2003 or 2005. The pwm chip on this one is KA7500. The output rails do not have pi filters, only output capacitor. No coils. I don't know if Sun Pro would stoop this low!!!

[Behemot]

Re: Mystery power supply

Or maybe any generic chineese one? They all share almost the same design. You can see e.g. empty position for second fan connector, characteristic heatsinks, +5V SB diode, 2 small transformers and so.

[c_hegge]

Re: Mystery power supply

Anyone ready to take a guess at who the manufacturer might be?

Leadman or Sun Pro maybe?