I got a 40+40 W Class B amplifier kit. The vendor told me that he caught hold of old stock. The ST symbol and transistor number etc are printed on a silver paint coating easy to scratch with a fingernail. It came with the pins bent inside, i.e., towards each other and the metal case had traces of rust also. Please comment whether the transistor is an STMicro original or not. I would like to buy 20 pcs or so to make a 5 channel amp for a start and I am afraid if it were not a fake I wouldn't be happy to lose that. Sorry if the picture is not of a high resolution/clarity as it has been taken with a mobile phone but i assume I have explained the morphology sufficiently.
Thanks in advance.
Thanks in advance.
Attachments
Parts like those from Malaysia could be up to 25 years old, dating from the Muar plant establishment. A metallic paint finish is rare now but possibly that is plating instead - difficult to tell and such details aren't described in readily accessed datasheets but the fact they have survived that time, with corrosion of the case steel says it's likely the original finish and these are genuine parts.
Personally, I wouldn't use rusty parts, whether they are genuine or not but that is your risk, probably more than the possibility of fakes. It would be hard to make profits from faking 2N3055, since they have been the cheapest common TO3 part for even longer than that. For a 40 watt amplifier, I think the risk of low performance due to a fake part is unlikely.
Take a look at another example of your brand and the internal size of die, and the heat spreader, which should be comparable if genuine parts, remembering there are many manufacturers with different internal construction designs: 2N3055 - Wikipedia, the free encyclopedia
Really, these should have been scrapped by the seller. Use them only if you get them for nothing or buy new parts from a genuine source.
Personally, I wouldn't use rusty parts, whether they are genuine or not but that is your risk, probably more than the possibility of fakes. It would be hard to make profits from faking 2N3055, since they have been the cheapest common TO3 part for even longer than that. For a 40 watt amplifier, I think the risk of low performance due to a fake part is unlikely.
Take a look at another example of your brand and the internal size of die, and the heat spreader, which should be comparable if genuine parts, remembering there are many manufacturers with different internal construction designs: 2N3055 - Wikipedia, the free encyclopedia
Really, these should have been scrapped by the seller. Use them only if you get them for nothing or buy new parts from a genuine source.
Last edited:
I recently updated a couple of 1980's 150WRMS Maplin amplifiers.
I replaced the output transistors 2N3055 MJ2955 with modern versions.
The modern versions had much better HFE and breakdown voltage.
I found I had to increase the VAS capacitance to stop oscillation.
I replaced the output transistors 2N3055 MJ2955 with modern versions.
The modern versions had much better HFE and breakdown voltage.
I found I had to increase the VAS capacitance to stop oscillation.
Yes; old designs counted on the huge capacitance and poor ft. It's getting hard to make transistors as bad as 2N3055s.
Last edited:
I think I found out about the batch number of my concern.
http://www.utsource.net/ic-datasheet/2N3055-679114.html
This is actually my first DIY and post too howsoever unwisely it may seem. Thank you very much Ian for imparting so much. Also fpitas, as your comment urges me to do a better project after completing this first.
Regarding your good suggestion Ian, I purchased the pcb/parts for 4 more channels just 3 hours earlier, a two channel pcb having been completed few days earlier. I was happy to even see the transistor long gone out of production in my city, hence the curiosity. But as suggested by Ian, as to go cheap, I got the total material for 6 channels for Rupees 1500 (30 USD) sans the heatsinks (10 USD). A 30-0-30 v EI type transformer, the best brand in my country, plus a 2x2200uF per rail dual supply (4 diodes, 4 caps) for 45USD. Elec caps are Vishay BC, ceramic/poly caps and 5%resistors are Philips. So totalling about 85USD, I think a 6 channel amp (5 to use) built up in nearly 12 hrs free time, ready for 85USD is a fair bargain.
Regarding the 14 2N3055s, got them for about 10USD. I'll post pictures of the bunch.
A LM3886 gainclone project/DIY at a future time also interests me. Are there any reports/cases of those being fake too? I fear that India sharing border with China might one of the most susceptible places for fake parts in Asia.
Any comments?
Regards to all.
http://www.utsource.net/ic-datasheet/2N3055-679114.html
This is actually my first DIY and post too howsoever unwisely it may seem. Thank you very much Ian for imparting so much. Also fpitas, as your comment urges me to do a better project after completing this first.
Regarding your good suggestion Ian, I purchased the pcb/parts for 4 more channels just 3 hours earlier, a two channel pcb having been completed few days earlier. I was happy to even see the transistor long gone out of production in my city, hence the curiosity. But as suggested by Ian, as to go cheap, I got the total material for 6 channels for Rupees 1500 (30 USD) sans the heatsinks (10 USD). A 30-0-30 v EI type transformer, the best brand in my country, plus a 2x2200uF per rail dual supply (4 diodes, 4 caps) for 45USD. Elec caps are Vishay BC, ceramic/poly caps and 5%resistors are Philips. So totalling about 85USD, I think a 6 channel amp (5 to use) built up in nearly 12 hrs free time, ready for 85USD is a fair bargain.
Regarding the 14 2N3055s, got them for about 10USD. I'll post pictures of the bunch.
A LM3886 gainclone project/DIY at a future time also interests me. Are there any reports/cases of those being fake too? I fear that India sharing border with China might one of the most susceptible places for fake parts in Asia.
Any comments?
Regards to all.
Date code is 2009. But the logo has a stripe across the bottom left corner - does not look right. I've got a 2009 device with a new format date code taking four lines with MYS on the bottom. Now ST may have changed the date code formats during 2009 and you just happen to have the last of the old but I'd be suspicious.
John
John
I agree that they look a bit unusual, but why on earth would someone go to such lengths (rust etc) to fake a rather recent 2N3055?...
Well, might be possible after all: since fakers have already exhausted everything else, why not go for something really unexpected.
They have to be really desperate....
We are probably going to see fakes of fakes appear in the near future
I just buy from reputable dealers instead of trying to save a few pennies here and there.
You can still buy 2n3055 from RS and Farnell (good dealers)
And Farnell still do the MJ2955. I bought a few of each recently and the HFE was very high, around 140 and the VBE breakdown voltage was around 100v (spec says 60v)
I bought a batch of irfb4227 off ebay and they blew far too easily.
They were way out of spec.
When I put in genuine parts it worked like a dream on full power all day.
The pic titled fake 3055 is my transistor cut open. The 3 sq.mm die can be seen held to the body by sort of epoxy and is not even mounted on a heat spreader, except the emitter side connection wire is ingeniously thicker. A better seeming version is pictured, a 2N3055 made by Bharat Electronics Ltd. (BEL) (Bharat means India) which seems at least relieving to look at as it has a proper heat spreader. And Bharat Electronics makes 2N3773 too. At least they dont fake but make their own originals, which implies people get what they are looking for, IF they are looking for a BEL product. Another company CDIL makes a lot of semiconductors. Also a person visiting my vendor told me that CDIL made/has made something for STMicro, which i doubt a bit, but it may be fine for some if the results match the specifications rigorously.Do manufacturers outsource parts??
I tested one channel of the amp without heatsinks for 15 seconds on 29volts 1amp on cheap 4 ohm woofer and the transistors got lukewarm. I'll connect the 30volt 5amp one, heatsink the transistors and will see if any fireworks happens. If all is fine then good else i'll resort to BEL or CDIL(if i can find it).
BEL India is defence oriented actually. Link to a relevant CDIL page-
CDIL Product Catalogue - Power Transistors - Power Transistors - Device Table
Please advise if TIP3055 is a good substitute except the heat dissipation etc, now that i could never get a ST genuine 2N3055.
I tested one channel of the amp without heatsinks for 15 seconds on 29volts 1amp on cheap 4 ohm woofer and the transistors got lukewarm. I'll connect the 30volt 5amp one, heatsink the transistors and will see if any fireworks happens. If all is fine then good else i'll resort to BEL or CDIL(if i can find it).
BEL India is defence oriented actually. Link to a relevant CDIL page-
CDIL Product Catalogue - Power Transistors - Power Transistors - Device Table
Please advise if TIP3055 is a good substitute except the heat dissipation etc, now that i could never get a ST genuine 2N3055.
Attachments
You should be able to buy ST or ON semi genuine devices from RS or Farnell at reasonable prices as has been mentioned in an earlier post.
I would recommend the metal can devices, by preference. The TIP 3055 is not a bad device, and usually has a 3MHz ft spec. which is adequate for good hifi. My concern is that the second breakdown limits begin at only 30V whereas the 2N3055 power limit begins at 40V. So the metal cans have more reserve power dissipation than the plastic versions.
If you want to build amps with power output of, say, 25-30W then the TIP's would be fine. If you want to run your amps higher - for 50-60W, this means using rail voltages of +/- 35-37.5V and in this case the metal cans, as people have mentioned, would be better at handling these voltages.
Be careful that the plastic types may also be faked!
John
I would recommend the metal can devices, by preference. The TIP 3055 is not a bad device, and usually has a 3MHz ft spec. which is adequate for good hifi. My concern is that the second breakdown limits begin at only 30V whereas the 2N3055 power limit begins at 40V. So the metal cans have more reserve power dissipation than the plastic versions.
If you want to build amps with power output of, say, 25-30W then the TIP's would be fine. If you want to run your amps higher - for 50-60W, this means using rail voltages of +/- 35-37.5V and in this case the metal cans, as people have mentioned, would be better at handling these voltages.
Be careful that the plastic types may also be faked!
John
Your other questions - I have not used any CDIL transistors but they appear to make a range of oldish standard devices. Probably OK.
It has been rumoured that some manufacturers have purchased parts from outside suppliers and put their own badge on, but if they have done that then they will (usually) have run qualification programs to verify the quality. As ON Semi and ST have had plants manufacturing transistors for decades (although in other guises perhaps such as Motorola and SGS) I would have thought they would not have needed to do this. But I do not know.
John
It has been rumoured that some manufacturers have purchased parts from outside suppliers and put their own badge on, but if they have done that then they will (usually) have run qualification programs to verify the quality. As ON Semi and ST have had plants manufacturing transistors for decades (although in other guises perhaps such as Motorola and SGS) I would have thought they would not have needed to do this. But I do not know.
John
The copper heat spreader must be more expensive than the die, because on fakes I see normal or larger dies mounted on small pear sized spreaders, or a 5mm copper "button" embedded in the steel body or the die straight soldered to the steel base, whcich simply does not work, even although some fakers make it double thickness (around 1/8").
So in decreasing order of quality:
* 3.5mm die soldered to full size copper "coin": good.
May be a fake, but it will probably work, they spent some effort and $ on it.
* 3.5mm dies on a smaller pear shaped less than half sized copper spreader: typical of red script "Toshiba" fakes: very good dies, measure well, but don't dissipate full power. Good for home audio or , say, up to 40W guitar amps, but not higher power.
* 5mm dies, straight soldered to the iron base, typically have a drop of a white substance on them, looks like some kind of epoxy paint. Useless, except maybe for a 12V regulated PSU or similar light work.
* 3.5mm die straight soldered to a double thickness base: useless.
* what you show, I had never seen something so ludicrous: a 3.5mm die ***epoxied*** to the iron base ... what were they thinking?
FWIW I used for ages in my Guitar amps 2N3055H made in India, USHA brand, labelled UR with a small arrow between them: *very* strong, durable, almost impossible to destroy even under grueling stage conditions.
Unfortunately the Importer had brought 5000 in and since I was the only one buying them, never again.
I'd like to hear something about that factory.
So in decreasing order of quality:
* 3.5mm die soldered to full size copper "coin": good.
May be a fake, but it will probably work, they spent some effort and $ on it.
* 3.5mm dies on a smaller pear shaped less than half sized copper spreader: typical of red script "Toshiba" fakes: very good dies, measure well, but don't dissipate full power. Good for home audio or , say, up to 40W guitar amps, but not higher power.
* 5mm dies, straight soldered to the iron base, typically have a drop of a white substance on them, looks like some kind of epoxy paint. Useless, except maybe for a 12V regulated PSU or similar light work.
* 3.5mm die straight soldered to a double thickness base: useless.
* what you show, I had never seen something so ludicrous: a 3.5mm die ***epoxied*** to the iron base ... what were they thinking?
FWIW I used for ages in my Guitar amps 2N3055H made in India, USHA brand, labelled UR with a small arrow between them: *very* strong, durable, almost impossible to destroy even under grueling stage conditions.
Unfortunately the Importer had brought 5000 in and since I was the only one buying them, never again.
I'd like to hear something about that factory.
john_ellis, JMFahey thank you for elaborating.
Since I am not a tech, let me put it in layman words (for my ease of mentioning)
I don't think I might be right, but are the u-turn shaped lines on the die like fingers interwoven, for increasing the p-n junction perimeter? On the pic posted by john_ellis and the Bel3055 pic posted by me it has less fingers i.e., perimeter between the p-n junction as compared to the pic of fake 3055, cut open by me (pic alongside the Bel3055 pic).
Although this reply might not matter to anyone as in todays time, except the humblest diyers or a fanatic, would like to make/revive a 3055 class B amp. But still irritated by the fakery and concerned with my first project, I would like to have my say howsoever nonsensical it may seem as I have forgotten the book knowledge now.
It must depend on surface area of the die (the junction should work by downward /upward diffusion of electron or holes between one side collector and another emitter (n-p-n) base being the center (which is half the fingers/area on the die). That might not matter as the surface area of the base and emitter (collector being the case) would remain the same regardless of the number of u-turns/fingers but what about the increased the base-emitter perimeter on the die surface. (if it matters)
I think that the increased perimeter must change something, let's assume, keeping the quality control constant.
Then if one could, deducing from what JMFahey wrote, take care of the heat, could the same fake 3055, cut open by me, work at least nearer to an original. Then for the sake of learning, i would experiment by immersing the transistor in water, insulating the emitter and base leg connections (in heatshrink sleeve and rubber adhesive) for improved dissipation just to find the limits before fireworks happen.
Regards to all.
P.S. I will atleast get the BEL 2N3773s tomorrow
Since I am not a tech, let me put it in layman words (for my ease of mentioning)
I don't think I might be right, but are the u-turn shaped lines on the die like fingers interwoven, for increasing the p-n junction perimeter? On the pic posted by john_ellis and the Bel3055 pic posted by me it has less fingers i.e., perimeter between the p-n junction as compared to the pic of fake 3055, cut open by me (pic alongside the Bel3055 pic).
Although this reply might not matter to anyone as in todays time, except the humblest diyers or a fanatic, would like to make/revive a 3055 class B amp. But still irritated by the fakery and concerned with my first project, I would like to have my say howsoever nonsensical it may seem as I have forgotten the book knowledge now.
It must depend on surface area of the die (the junction should work by downward /upward diffusion of electron or holes between one side collector and another emitter (n-p-n) base being the center (which is half the fingers/area on the die). That might not matter as the surface area of the base and emitter (collector being the case) would remain the same regardless of the number of u-turns/fingers but what about the increased the base-emitter perimeter on the die surface. (if it matters)
I think that the increased perimeter must change something, let's assume, keeping the quality control constant.
Then if one could, deducing from what JMFahey wrote, take care of the heat, could the same fake 3055, cut open by me, work at least nearer to an original. Then for the sake of learning, i would experiment by immersing the transistor in water, insulating the emitter and base leg connections (in heatshrink sleeve and rubber adhesive) for improved dissipation just to find the limits before fireworks happen.
Regards to all.
P.S. I will atleast get the BEL 2N3773s tomorrow
JMFahey,
BEL transistors are available easily here in my city atleast (only small vendors here as it's a residential city with shops so basic as compared to bigger countries.
People are interested in doing business of what earns them most profit (Clothing, electronic/electrical trading, food, everything of basic community needs) Nobody gives a damn about hightech stuff, maybe people dont know and whoever know don't care) People are contented if they have a SUV or German car. They will pay out of their nose for that but still look for sale season for other stuff.
Not to wander off more, if you are interested, i will enquire about CDIL (i have their BD139,40 procured 3 days ago).
Hi sarkari, thanks A LOT for the update.
As of your amp, don't go as far as water cooling it, just bolt it to a good heatsink, with proper mica and grease, and test it.
I mean, *IF* yours is the one showing the full size heat spreader and the die soldered to it.
As of the epoxied ones, they are useful only to make some key holders or similar decorative use, but nothing else.
As of your amp, don't go as far as water cooling it, just bolt it to a good heatsink, with proper mica and grease, and test it.
I mean, *IF* yours is the one showing the full size heat spreader and the die soldered to it.
As of the epoxied ones, they are useful only to make some key holders or similar decorative use, but nothing else.
Hi Sarkari
The interdigitated fingers in a bipolar do increase the perimeter but the main reason is to access the base electrode more often. The base is deposited over the collector during manufacture then the emitter has to be formed into fingers to allow the base to be contacted in between.
The performance of the transistor depends on being able to connect to the base under the emitter so narrower emitters with more fingers should go faster, but it depends on a lot of other parameters as well (base doping, thickness etc)
John
The interdigitated fingers in a bipolar do increase the perimeter but the main reason is to access the base electrode more often. The base is deposited over the collector during manufacture then the emitter has to be formed into fingers to allow the base to be contacted in between.
The performance of the transistor depends on being able to connect to the base under the emitter so narrower emitters with more fingers should go faster, but it depends on a lot of other parameters as well (base doping, thickness etc)
John
- Status
- Not open for further replies.