ultrasonic cleaner MH-060S, replacing transistors [fixed > broken again]

[htm]

Hi,

i'm trying to fix this same cleaner
the 15L version (MH-060S)

It has 2 boards (see attached images).
I marked the bad components on the main board: transistor, some diodes and a blown fuse.(i only know to test the voltage drop using the diode-function on the multimeter)
The unmarked transistor shows a voltage drop of 0.3volt accros B&E. Maybe that one is Ok?
(The second board has the exact same issue)

Both boards having 2x  SC3320 transistors.
I've found them on ALI for much too cheap ( link )
But i'm probably better of buying proper ones.

I've tried searching them outside aliexpres, but wasn't successful.
But since i'm not so well versed in electronics,
i'm hoping someone here can propose a proper replacement part and website for the EU? Any other suggestions are welcome as well.

Thanks

[fzabkar]

AFAICT, the 2SC3320 was originally produced by Fuji Electric in the early 1990s.

https://image.makewebeasy.net/makeweb/0/u8mVxmGkK/Document/_62_Fuji_Semiconductor_2SC3320_datasheet.pdf?v=202012190947
https://pl-1.org/getproductfile.axd?id=8702&filename=2SC3320.pdf

Nowadays there appear to be numerous copies. Here are just a few:

https://pdf.direnc.net/upload/2sc3320-datasheet.pdf
https://www.unisonic.com.tw/uploadfiles/836/part_no_pdf/2SC3320.pdf
https://datasheet.octopart.com/2SC3320-Inchange-Semiconductor-datasheet-15978868.pdf
https://storage.ua.prom.st/2579499_c3320.pdf

I wouldn't know what to recommend.

Edit:

BUV48 / BUV48A looks like a suitable substitute, but it is just as difficult to obtain from trusted sources. There are versions in plastic packages, but I would prefer metal.

https://rocelec.widen.net/view/pdf/lvpyrsjj9z/ONSM-S-A0005942251-1.pdf?t.download=true&u=5oefqw
https://www.mouser.com/catalog/specsheets/buv48.pdf

[htm]

Thanks!

Yours is cheaper than the NTE2311 i found(which is a bit overrated).
Still shipment with those sites(farnell, tme, mouser) is substantial :/

When i tried looking for a replacement myself, i was comparing thise properties:
 CE-current=400 CE-voltage=15A, BE-saturation voltage and try comparing switching times.
Would these parameters be enough for checking?

Like i saw darlington and biplair come. i believe darlington internally has more components and is used to have more gain? is it advantagious to stick to bipolair here?
Did you compare more properties for this application?

For fitting that other metal-package; i wouldn't know what heatsink to use.
But what are the advantages of a metal case? much better heat dissipation i guess, or more?

[fzabkar]

I found the NTE part as well, but it cost $19.

As for parameters, I would also compare hfe, in this case 10 (min), and the switching times. The ultrasonic cleaner is a tuned circuit, so the parts would have been carefully selected. A Darlington pair wouldn't be suitable, I would think.

When you are testing the cleaner, I would let it run for a few minutes, then check the temperature of the transistors. I used to do this with line output transistors in CRT TVs. Fakes would run very hot while the genuine transistors would be relatively cool.

[htm]

On this site, selecting the parameters starting with 15A i came as only available part to your BUV48A and it ships locally, so without customs fees
https://www.reichelt.nl/nl/nl/transistor-npn-to-218-450v-15a-150w-buv-48a-isc-p6180.html?&nbc=1

Looking at the PDF, the component not having printed parts number, and the next page has 2 pictures misplaced, doesn't give a good impression.

Maybe a dumb question;Would you consider buying from this store? or would aliexpress be similarly trustable regarding authenticity?

[wraper]

LCSC has SC3320 from Chinese manufacturers.
https://www.lcsc.com/product-detail/Bipolar-BJT_SPS-2SC3320T4TL_C462223.html
https://www.lcsc.com/product-detail/Bipolar-BJT_SPTECH-2SC3320_C480012.html

[htm]

@wraper
Thanks for the links.

In these datascheets, i see Vbe(sat) = (max)1.5V , but it doesn't seem to provide a 'normal' value for it.
but the 4 transistors in my device read different for all of them;

#	Vbe	Vbc	Vce
Mainboard left:	0.001	0.002	0.001
Mainboard right:	0.3	0.5	0.5
2ndboard left:	0.001	0.002	0.002
2ndboard right:	0.3	0.3	0.001

Seems only 1 survived. i'm assuming it's better to replace all of them.

[htm]

i'm persuaded to go with the BUV48A since it seems to have double the voltage rating can handle 60 more Watts.

Next choice to make:
There's a store www.reichelt.nl which has the BUV84A by "Inchange semiconductor",  for €4,06  (4x + local transport fee)  Total €23.18
and there's a store www.tme.eu which has the BUV84A by "STMICROELECTRONICS" for €5,13    (4x + international transport fee)  Total €32.50

I'm not sure if "Inchange" is a reputable source, but the latter sounds familiar
What are the odds i'm getting into trouble with the "inchange" one?

[fzabkar]

I think the ST part is plastic.

[htm]

I think the ST part is plastic.

You mean, the ST part is dodgy instead of the "inchange" one?

otherwise, if you mean they both are plastic case.
That's because i'm not confident enough to try a metal TO-3 case on the board and have another heatsink connected. i'm quite a newby and unfamiliar with the options.

i guess metal one is more heavy duty(i'm a newby). but how do i get a proper heatsink fot it which fits? From what i see on mouser, those heatsink require quite some space on the board and screwholes as well.

[wraper]

Using metal can transistors in 2024 unless replacing original metal can part is quite dumb for multiple reasons.

[htm]

@wraper
Haha, ok.

[wraper]

BTW PCB footprint also allows using TO-220 package which gives way more substitute choices. Somewhat higher thermal resistance does not matter at all as heatsinks/heat dissipation are not anywhere large enough for it to play a significant role.

[htm]

@wraper
i've searched for Ice>=15A , Vce >=400 , but most casings i see are TO3 and only few are TO3P, on both TME and Reichelt. i don't find any TO-220?

Regarding both suppliers of the BUV48A for the plastic versions, would both parts be same quality? (i added links in previous message) or is ony brand preferred?

[wraper]

You're right. All 15A stuff in TO-220 seems to be discontinued. However you don't necessarily need 15A max either. Lower current rating would be fine as long as saturation voltage is low enough, gain at high current high enough and it's not too slow as what you should care the most about is heating not more than original.

[htm]

The PCB has a few big capacitors rated at 450V, and it has 2 big coils probably capable of generating a big current rush, so i'd rather have a sturdy component than another one that breaks down in a while.

i don't mind getting the BUV48A. Only thing is, i don't want to buy bad look-alikes. Hence my question if either brand has better reputation. (or maybe both shops only have proper brands. i don't know)

[edit:]I'll go with the TME website, because reichelt has the specs for BUV48 listed for its part BUV48A..

[fzabkar]

Using metal can transistors in 2024 unless replacing original metal can part is quite dumb for multiple reasons.

Why?

[htm]

.. Its designer chose a 30-year-old transistor in a metal package ..

i don't understand. The picture i posted isn't a metal package, is it?

[fzabkar]

.. Its designer chose a 30-year-old transistor in a metal package ..

i don't understand. The picture i posted isn't a metal package, is it?

Damn, I've confused myself again. Yes, you're right. Yours is plastic. If you do decide on a metal package, make sure that the heatsink is isolated from the circuit.

[fzabkar]

You're right. All 15A stuff in TO-220 seems to be discontinued. However you don't necessarily need 15A max either. Lower current rating would be fine as long as saturation voltage is low enough, gain at high current high enough and it's not too slow as what you should care the most about is heating not more than original.

The MH-060S cleaner is a current design, yet its designer chose a 30-year-old transistor. ISTM that s/he would know best what works in that particular application. Otherwise, why not choose from a plethora of more common, potentially cheaper transistors? :-?

[htm]

Yeah, confuses me as well.. current design with a component hard to get 
i can't wait to try the thing with the new components.. 

[fzabkar]

Some techs will use cheap substitutes to validate a repair, then replace them with the expensive parts once they are satisfied.

[wraper]

You're right. All 15A stuff in TO-220 seems to be discontinued. However you don't necessarily need 15A max either. Lower current rating would be fine as long as saturation voltage is low enough, gain at high current high enough and it's not too slow as what you should care the most about is heating not more than original.

The MH-060S cleaner is a current design, yet its designer chose a 30-year-old transistor. ISTM that s/he would know best what works in that particular application. Otherwise, why not choose from a plethora of more common, potentially cheaper transistors? :-?

Because they cloned some decades old design without thinking too much? On other hand it's a Chinese 2SC3320, so might be because they could get it cheap and TO-220 footprint is for using say xxx13009 instead.

[elektryk]

TO-220 footprint is for using say xxx13009 instead.

MJE13009 may be good starting point for debugging.

BTW I tried to repair several ultrasonic cleaners, but some of them worked only a few minutes/seconds and burned another pair of BJTs, probably due to broken inductor, transformer or ultrasonic transducer itself.

[htm]

@elektryk
Ok. thanks. not verry promising 🤒
i'll have to find out how to test if inductor, transformer or transducers are defective..

Regarding MJE13009 ; this one says 400V and 12A, while the capacitors on the board says 450V. Wouldn't the same voltage be smart?