Warning fake components

[sideoiler]

Hi all,
Just a warning to you all about fake components brought off the net and e-bay
I am trying to repair my 1970 wharfedale 100.1 amp, and I was not having much luck as I was blowing the output transistors like they was going out of fashion. In the end I had some new 2N3055"E" from RS and fitted them and now have one of my wharfedales up and running. [ sort off ]

I have now found out I brought off e bay fake 2N3055 transistors, and after doing some research it looks to be a big problem out there for fake components so be warned before you buy.

http://www.diyaudio.com/forums/solid...nterfeits.html

pic 1,2 fake
3-5 real thing from various manufacture like RCA from the early 70s

[Brian R Pateman]

Buying components from unknown sources is always a risk.

I tend to stick with suppliers I know, Cricklewood Electronics for example. They are reasonably priced and delivery is usually next day, even in my remote corner of the country.

I have no connection apart from being a regular customer for over 30 years.

[paulsherwin]

2N3055s are notoriously variable in quality.

['LIVEWIRE?']

I always buy semiconductors (and most other components) from reputable sources such as CPC, Cricklewood Electronics, CHS, and formerly from SEME-Nedis (Nedis UK) That way you are unlikely to get fake parts.

[McMurdo]

I use alot of diverse semiconductors in my day job. I'm not going to name names in public but I've unfortunately had a few fake parts from a few of the suppliers mentioned in this thread. I suspect the suppliers weren't even aware they were fake. In one case, one of the suppliers sent some audio transistors that I thought had been obsolete for a few years, yet they had a seemingly endless supply. When they came and I fitted them, they blew after a few seconds. I opened one up and it had a tiny die compared to the original parts, so I sent them back for full credit.
In another case I had some ICs which when I used them were all out of spec. They were also supposed to be obsolete and again the supplier seemed to have lots in stock; I suspected they were rejects that had somehow found their way onto the market. I never got a refund for those and managed to source a couple of NOS from an ebay seller instead which, ironically, were perfect.
Yet another case, I was pleased to find a supplier who said they could get an obsolete type of International Rectifier power thyristor bridge. When it arrived, I put it in the equipment (a dc motor shunt field rectifier) and a quick test showed a short across the 415V AC supply. The new thyristor module, when tested, was found to be internally wired wrong with a flywheel diode across the AC terminals!!! I got a refund that time.

It's a minefield.

[Nightcruiser]

I've had a lot of trouble with CPC Farnell recently so I'm not going to use them so much in the future.

Extremely poor sorry about that attitudes coupled with very poor stock control resulting in being let down time and time again and partfilled back orders. I'll take my business elsewhere.

[Restoration73]

Be aware of the differences of 2N3055E and 2N3055H types. Whilst it is informative to
remove the can to examine the junctions, do not do this to specialist radio frequency
transistors which contain beryllium oxide, as contact with this material may result in
severe health implications.

[TonyDuell]

I have perhaps been lucky...

Worst I had was from a company that has been mentioned in this thread (and who, to be fair are normally very reliable and I have dealt with them many times). I bought some 1A thyristors in TO5 cans. When they arrived they were unmarked (!). I fitted one to an ancient telephone answering machine I was repairing, the control system of which consisted of many thyristors with turn-off capacitors between them trigger by things like incoming ring, end of announcement, etc Anyway, with said 'thyristor' fitted it didn't work at all, turns out it would never turn off. In the end I removed all drive from the gate to find it was still 'on' at power-on. To cut a long story short it turned out I had been supplied with PNP transistors in TO5 cans and the gate pull-down resistor was turning them on (OK, emitter and collector were technically the wrong way round, but when you have a 470ohm base resistor and a 15k collector load you don't need much gain....).

Of course the supplier refused to replace them (they have a 'no returns' policy) but to be honest the cost of the components wasn't the issue. It was the several hours I had wasted...

Perhaps the most amazing thing is that I was once looking for a custom ROM/RAM IC for an HP calculator. I found a supplier (you can guess where) who said they could supply it, I contacted them to ask for price and order quantity and they asked me what package I wanted it in. Given it only ever existed in one type of package I am guessing they would print the number I was asking for on any IC with the right number of pins and then when it didn't work would fall back on the 'you must have damaged it' line.

[evingar]

Originally Posted by McMurdo

I use alot of diverse semiconductors in my day job. I'm not going to name names in public but I've unfortunately had a few fake parts from a few of the suppliers mentioned in this thread. I suspect the suppliers weren't even aware they were fake. In one case, one of the suppliers sent some audio transistors that I thought had been obsolete for a few years, yet they had a seemingly endless supply. When they came and I fitted them, they blew after a few seconds. I opened one up and it had a tiny die compared to the original parts, so I sent them back for full credit.
In another case I had some ICs which when I used them were all out of spec. They were also supposed to be obsolete and again the supplier seemed to have lots in stock; I suspected they were rejects that had somehow found their way onto the market. I never got a refund for those and managed to source a couple of NOS from an ebay seller instead which, ironically, were perfect.
Yet another case, I was pleased to find a supplier who said they could get an obsolete type of International Rectifier power thyristor bridge. When it arrived, I put it in the equipment (a dc motor shunt field rectifier) and a quick test showed a short across the 415V AC supply. The new thyristor module, when tested, was found to be internally wired wrong with a flywheel diode across the AC terminals!!! I got a refund that time.

Ditto here - We repair a lot of old switched mode PSU's and occasionally forced chance it with suppliers that we normally wouldn't touch with a barge pole. Having said that, like you I've also had problems with "blue chip" suppliers, most recently I had a batch of PN2222A transistors that simply didn't work in a power supply - measured out OK and "kind of" worked. Of course, once you have replaced an obviously failed device with new one (from a good source) and the fault signature changes, you start to look elsewhere for the remaining problem - especially if you don't have service data for the article you are trying to repair !

Eventually after "going round the houses" I returned to the transistor I had replaced, did some out of circuit tests on and found it was incredibly low gained - manufacturers rejects springs to mind.

The cost of such components is trivial - time spent in a commercial situation though turned a potentially simple repair into a damned expensive exercise - not to mention negative "brownie points" from a miffed client due to late delivery

[Radio Wrangler]

Back a few decades ago the 2N3055 was a dumping ground for reject power transistors. There were many RF and SMPS types with much whizzier ratings and dramatically higher Ft which, if a batch didn't meet spec still far exceeded the very open-ended specs of the 2N3055

So the wafers got diced and the dice got packaged. The reject silicon could still be sold for some money, albeit a much lower price.

These devices still had large die in them, and could really do 60v 15A 115W. The problem was that the higher gains and Ft made all sorts of things go unstable. Quad triplets being definitely in trouble. I just stopped designing 2N3055 into anything and went for tighter spec'd devices.

But things have changed and the market is flooded with any old die in a TO-3 case. Probably the idea is that anything which will show up as "bipolar NPN' on a hand-held tester won't be spotted before the money changes hands.

I've had fake AD pampas fitted by a major league multinational contract manufacturer.

David

[kalee20]

Occasionally, fake devices do get sold by reputable distributors.

What happens is, somebody dodgy buys a quantity, and a few weeks later asks to return them. They may happily pay a restocking/admin charge, but instead send back a load of cheap counterfeit devices which the distributor accepts back in good faith. So the punter still has 'good' devices which he can sell at near full market value, in exchange for an equal quantity of counterfeits devices and the restocking charge.

2N3055's, as has been mentioned, have a fairly wide spec. I have seen some samples with a small chip inside and wire bonds to the pins (Motorola manufacture), and some with larger chips and chunky stamped leads (RCA manufacture). It doesn't necessarily mean that they are fake, but it does show the variation between manufacturers.

Sometimes, manufacturing improvements make a massive difference to performance - I seem to recall the '3055 has an ft of 2.5MHz (min) and while original devices may have barely exceeded this, newer devices could so vastly exceed it that RF oscillation starts in an amplifier.

[sideoiler]

Originally Posted by Restoration73

Be aware of the differences of 2N3055E and 23055H types.

I have never found out the difference from a normal 2N3055 and a 2N3055E any one know ?

After doing some research it looks like the MJ15003 may be a better replacement for the old 2N3055 transistor.

[BottleMan]

Originally Posted by Restoration73

Be aware of the differences of 2N3055E and 2N3055H types. Whilst it is informative to
remove the can to examine the junctions, do not do this to specialist radio frequency
transistors which contain beryllium oxide, as contact with this material may result in
severe health implications.

There should be no beryllium in either of these transistors. The "E" suffix denotes a different version of the "2N3055" that uses epitaxial construction, was introduced by Solitron in about 1964 and has a far higher transition frequency. Beware of parasitic oscillations. I'm not sure, but it might have a different safe operating area to the standard part. This is important if the amplifier drives loudspeakers with big reactive changes.
Graham

[Herald1360]

The original 2N3055 was a dormant doorstop of a device with lowish Ft and moderate gain along with good second breakdown characteristics. They used non epitaxial construction. The open ended specs, though, meant that cheaper epitaxial devices could easily meet the requirements though they introduced problems for designs that relied on some of the older device's characteristics.

RCA latterly Harris now ? renamed the original device as 2N3055H to distiguish it from epitaxial versions.

[G6Tanuki]

The big problem with subbing 'modern' epitaxial 2N3055s in situations where the designer used the original hometaxial 2N3055 is that the Epitaxial version retains its gain up into the low-Megahertz range - so what was a previously 'tame' series-regulator power supply or amplifier originally designed/manufactured in 1965 using Hometaxials becomes a 50-Watt randomly-variable-frequency sawtooth power-oscillator after you've replaced the pass-transistors with the post-1970-production Epitaxial versions (some of which are quite capable of being used as "top-band" RF-amp transistors!).

Trying to retro-fit 'parasitic stoppers' to tame this problem can be fun: wirewound emitter-resistors with ferrite beads on the legs isn't always enough.

[McMurdo]

renamed the original device as 2N3055H

homotaxial, afaik

[julie_m]

I bought some cheap 2N3055s at a radio rally about 10 years ago and they turned out to be "a bit lively", breaking into oscillation at a few MHz. Nothing a couple of small-value ceramic capacitors could not tame but the original design must have relied on the power devices running out of gain much sooner.

[Brian R Pateman]

Two posts have been deleted.

This forum is not the place to discuss the sale of goods act, dealers trading practices or other potentially litigious matter.

Further instances will result in the closure of the thread.

[Radio Wrangler]

I'm a bit surprised at 2N3055s being faked, because usually the perpetrators go for much more expensive types. Their costs are the same and the income is much larger.

The price of everything in a TO-3 package has been going up, and the manufacturers say it's the cost of the case, so that ought to be pushing the fakers to go for higher profit parts.

If they'll fake 2N3055, they'll fake anything.

David

[camtechman]

Fake Transistors are nothing new. As far back as the 1970's when there was a shortage of semiconductors, there was quite a few companies who could supply "Re-brands". One of which had all the facilities to brand up a lesser known brand to a better one, they even had Texas Instruments stamps.

In fact, during that period, I worked for a component sourcing company in Herts, and there were quite a few of them, as it was a profitable market, and one top Re-branding company tried to head-hunt me to work in east Anglia (cough..cough)