1
\$\begingroup\$

I'm trying to repair a broken DC-ATX power supply that I have. After a bit of searching for faults I believe I have tracked the problem down to one of the MOSFETs.

enter image description here

I would like to try replace this MOSFET and see if doing so fixes the supply. The problem is that I'm not 100% sure how to identify the correct replace If I search 50N03 I find various different manufacturers components often with a "-11" or similar appended, though a few without, and all of them seem to have slightly different specificaitons.

How can I ascertain the correct part to buy here, also I'm not sure anyway local (nowhere I've looked so far seems to) has any 50N03 at all - are there any possible substitutes that anyone knows of that would do the same job?

Can anyone tell me what the "64B8F" code means, I can't seem to find much if I search for this?

\$\endgroup\$
  • \$\begingroup\$ 64B8F are more than likely going to be date and/or batch codes. \$\endgroup\$ – Doodle Apr 4 '16 at 11:16
2
\$\begingroup\$

The device is apparently one of the switches in a SMPS.

The base part is a 30V (Vds) rated N channel device in a TO-252 (DPAK) package.

If you cannot get one of these, then look for a device that has the following parameters close to the original (these are the primary parameters in SMPS device choices):

\$V_gs\$ (max) : keep within the range of the original part

\$Q_g\$ (total) : as low as possible, no higher than the original (but not too much lower either - this affects transient performance significantly and too low could cause interesting issues).

\$R_ds\$ (on) : as low as possible, no more than the original part.

The diode conduction parameters are also a key concern, as many cheap supplies rely on this diode rather than fit an external device for cost reasons.

Make sure you find a part with very similar \$I_d\$ parameters.

Ensure that the absolute maximum rating for any replacement are at least as good as the original part.

Many manufacturers have parametric search capabilities. Try Vishay, International Rectifier, Fairchild and Diodes Inc for starters.

[Update]

I looked at the IPD050N03L from Infineon (looks like an original IR part) and the only specification that is significantly different is diode avalanche pulse energy; interestingly, the original datasheet does not even have a safe operating area graph.

The key parameters are marginally better (and the gate charge and input capacitance are similar, so switch turn on and turn off time should be the same as the original).

Overall, it looks like it should do the job.

[Update]

Based on the latest comment, I suggest using the parametric tables to find a device with headline parameters of:

\$I_D\$ 50A or higher

\$V_gs(th)\$ <2.4V

\$Q_g\$ around 10 to 20 nC

With a filter on the package of DPAK (TO252)

Then look at the other parameters. Something with these specifications should do the job.

\$\endgroup\$
  • \$\begingroup\$ Is it necessary for \$I_d\$ to be similar, or just larger? I'm eyeing CEU85A3 as a possible replacement as it seems to be almost identical in specs - but it has an 80A \$I_d\$ vs the 50A of H50N03J \$\endgroup\$ – Malcolm MacLeod Apr 6 '16 at 17:54
  • \$\begingroup\$ I will have a look in the morning - a larger current capacity is probably ok \$\endgroup\$ – Peter Smith Apr 6 '16 at 18:18
  • \$\begingroup\$ Thanks, I'm now also looking at the IPD050N03L infineon.com/dgdl/… - It has a slightly larger Vds (30V vs 25V) than the H50N03J, and differs in other ways (faster turn on/off ) by amounts that I am hoping are small enough not to matter, but has the advantage that I can actually source them here in South Africa. From what I can tell the two are relatively closely matched, hoping I haven't missed something. The CEU85A3 is going to have to be posted which may take a long time. \$\endgroup\$ – Malcolm MacLeod Apr 6 '16 at 19:03
  • \$\begingroup\$ Just as an update - za.rs-online.com have disappointingly just informed me that they can't supply the IPD050N03L after all - they suggested the IRLR7821 instead, which to me seems that it is too different. I'll continue hunting and update with the final results. \$\endgroup\$ – Malcolm MacLeod Apr 14 '16 at 12:21
  • \$\begingroup\$ You might like to look at this: irf.com/product-info/datasheets/data/irfr3707zpbf.pdf Qg is a little lower, but otherwise it has similar performance to the original. \$\endgroup\$ – Peter Smith Apr 14 '16 at 14:34
3
\$\begingroup\$

I've had a little search around for you and it looks like you've got a H50N03J made by a company called Hi-Sincerity. I've found this datasheet here. If you go down to page 4 in the datasheet it explains what the marking on the chip is very nicely.

You might not be able to get this exact chip anymore but at least this will give you the parameters of the MOSFET you currently have

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.