Fake Components

[ppppenguin]

All (certainly most) single chip linear regulators from the 7805 onwards need decoupling close to their input and output pins. I think I first discovered this in 1974 when a 7805 was behaving strangely. The 50MHz scope I had at my disposal couldn't see the 300MHz+ oscillation at the output. 100n caps cured it.

As for buying components from secondary suppliers, my time and effort are worth more than the extra cost of going to RS or other recognised distributors who have a reputation to uphold. I would only go to unknown suppliers if it was the only way to get a part.

[BusyBee]

It looks like you have it covered but I also tend to scope inputs and outputs when I get these sort of issues. A dicky rectifier on the input may limit the supply? My main finding here when I do things is to find I tend to overestimate the capabilities of the transformer I use. Either that or I overdo it. As the Americans would say (in another field) 'There is no substitute for cubes' lol

Tracy

[Crackle]

I did other tests including checking input volts to the IC, and there was still plenty of headroom there, the volts on the smoothing cap had dropped to just over the RMS volts from the transformer, which was 16 volts
I used a massively overrated transformer for these tests to ensure there was no possibility of the transformer voltage sagging.

No I am certain the current was being limited by the regulator at just over 1 amp, in other words they are\NOT 5 amp regulators

I have ordered 2 more LM338T regulators, these look different to the first lot, They are branded as National Semiconductor and have V4 stamped in a small circle on the body just over the ADJ pin.
The faulty/fake ones have MJC stamped in a small circle in the body just over the OUT pin.
   
Mike

[Mark Hennessy]

So where did the 2 new ones come from?

[Nick]

For the sake of a few bob, just buy from a reputable source. The amount of time chasing dodgy eBay components far outweighs the cost of getting it right first time...

[Crackle]

The problem was the RS Texas Instruments ones look identical to the first ones I bought, and it is highly likely the Farnell ones are the same imported stock as they are also branded TI.
I have chosen some which are a different stock (NS) in the hope they will work.
No problem if it doesn't I just demand a refund, easy.

Mike

[Mark Hennessy]

RS and Farnell won't be selling fake parts.

The fact the first ones look like stock parts from RS and Farnell is a credit to the effectiveness of the counterfeiters.

Did you put a 'scope on the output of the first one when testing earlier?

[Nick]

The major suppliers spend a great deal of time and money ensuring the quality of what they offer - they buy directly from the manufacturer and batch test what they receive. The source of each component is traceable (allowing auditing) in case of problems, allowing end users peace of mind. They don't buy from eBay.

The guaranteed quality of their stock is one reason they are more expensive than the knock-offs, but the financial and reputational cost of shipping a product which subsequently has to be recalled is huge by comparison.

Seriously, if you can avoid it, just don't buy active components from eBay or similar. It's a false economy.

[Crackle]

Hi Mark
No I have not scoped the output, yet.

Mike

[Crackle]

Just put a scope on the output and nothing other an a nice flat line was detected.
It works very well as a 1 amp regulator  

The replacements arrived from the other supplier, but were not the same as the one in the photo on the ebay listing.
They were the same type as the already failed ones, and guess what they also failed. That's 2 refunds now received.
I seem to be building a collection of 1 amp variable voltage regulators.

I am waiting to hear if the last supplier has any of the original ones he was selling left.

Regarding buying from Ebay, well somebody has to clean up all the fakes, it may as well be me, time is not a concern.
The reason for this thread was just to inform and warn others who may not be testing these devices fully, they would sit there working happily at 1.3 amps if that is all you want.

Mike