Fake Components

[Mike Watterson]

Maybe mislabelled or faulty rather than fakes. I can't see the motivation for relabelling a 25c part as a 28c part when it's going to be discovered immediately and there is hardly the profit to pay for the relabelling.

It's inexplicable to me.

[Refugee]

I have found that there is a good non snake oil substitute to most simple parts out there if you are prepared to thumb a few data sheets.
I managed to find a compliment to the IRF840 on the Farnell/Element14 website.
I have not yet built my project with it though.

[Mark Hennessy]

It's inexplicable to me.

That's as may be, but it's very real indeed. Huge, in fact. It's not difficult to see how it happens - I'll resist the temptation to use lmgtfy.com.

I'm guessing Mike has an LM317T clone that's been relabelled. In theory a genuine LM317T is good for 1.5A given proper heat sinking and attention to detail, but not all the clones manage the full output current. Of course, it could be something else entirely that works in the same way - there are many possibilities.

Either way, getting higher currents from even a genuine TO220 LM338 can be challenging - a NOS or second-hand LM338 (TO3 package) is a much better bet IME.

A lot of the sellers offering LM338s target the DIY audio market, where they get away with it because the average currents are very low compared to the peak currents (which are handled by the output cap). Their use in audio started many years ago when "Gainclones" were all the rage - some bright spark came up with the (almost entirely pointless) idea of regulating the PSU rails.

[Mike Watterson]

Certainly a partial explanation. You'd need three things for it to work for a seller:
1) Getting a load of LM317T for nearly nothing. The prevalence of NOS out of production parts from China suggests large left over inventories from when a product ended. That's totally believable.
2) Ability to remark a part very easily and cheaply with almost no labour and invisibly.
3) A very large number of people buying LM338T that only need LM317T. The real LM338T is cheap anyway.

You've sort of explained (3) though considering how cheap "real" LM338T are there must be a lot of cat 3 people and (1) and (2) must staggeringly apply. Most of these sellers have a huge inventory of parts that do what is claimed. The usual kind of "fake" is actually for expensive military/Aerospace/automotive parts and the supplied parts work. What are they faking? Two aspects. One is that the part may be actually ex-equipment and the other is that the part is probably really only tested to commercial for domestic production and possibly also in a cheaper package (material or radiation levels).

It's possible that the suppliers to the sellers are doing the fraud, I suppose.

I've had problems years ago with RS, so the companies changed to Farnell. Long ago Radiospares never returned parts to the maker. I worked in a Capacitor factory in early 1970s and all the just out of spec parts were sent to them if the production targets were not met. At the end of the 1970s in another company we had sealed clear bubble packs of RS parts pined to the notice board: Pre-blown fuses, screws with no thread etc. No doubt RS has improved QA as selling to production & R&D has replaced sales to Service Depts of Radio & TV shops.
I found that Maplin in its last years was dreadful. I was getting cheaper parts at better quality from Asia.

Professional users etc are better buying from Mouser, Farnell, RS etc. But many hobbyists can't afford them, especially outside UK and with the shipping charges etc. So we buy ex-USSR surplus from Baltic States and Eastern Europe and cheap parts from Asia.

[Mike Watterson]

What's a "gainclone"?

[Nick]

What's a "gainclone"?

https://en.m.wikipedia.org/wiki/Gainclone

[Mike Watterson]

What's a "gainclone"?

https://en.m.wikipedia.org/wiki/Gainclone

So nothing special AT ALL. Simply DIY amps using high power ICs that inevitably replaced the hybrid modules that were simple for DIY amps.
The original 47 Labs had a stupid name and exploitive price for what is basically a variation of the maker's application note. Gainclone is an even stupider name. My 5.1 Amp has similar monolithic power ICs and cost €99. No decoder, however HDMI / Coax / RCA digital audio in to 5.1 + stereo out digital decoders to drive dumb amps are now very cheap.

[Mark Hennessy]

Only they're not cheap - they are many times the price of a name-brand LM317T, let along a Chinese clone.

As for Gainclones, please do some basic research before making such a proclamation: https://lmgtfy.com/?q=gainclone

The reason for the name comes from the commercial product that inspired it - the Gaincard. At the time, it was a hugely radical thing. To appreciate that, you would need to understand the hi-fi scene at the time, and think a bit wider than just the engineering, which I appreciate might not be easy (I aim that at engineers in general, not anyone specifically).

What the Gainclone craze provided was a route into DIY audio that previously wasn't available to people with limited experience - until that point, IC power amps had mediocre specs and were not considered suitable for anything beyond TV sets and music centres. But if you took the time to investigate what an LM3875 or LM3886 can do, you wouldn't be so quick to dismiss it - with attention to detail, 1kHz THD+N of less than 0.003% at 50 watts is easily attainable, and that is pretty impressive by any reference. And by no means guaranteed with discrete circuitry. Put them in parallel/bridge arrangements and you can easily do 150 watts or more. I've lost count of the number of them that I've implemented, and I've lost count of the number of commercial hi-fi and pro-audio products that use them. They measure great, sound good, never fail IME, and are cheap. What's not to like?

Unfortunately, the people who look down on Gainclones are usually the "grumpy old man" crowd who didn't have access to such high performance chips when they were new to audio electronics. But this sort of attitude does nothing to encourage the next generation of engineers, and really isn't something we'd want to encourage on this forum. We've had similar problems in the past when discussing things like Raspberry Pis, where we saw comments like "it's just another SBC". Yes, but look wider than the immediate engineering. Few products are a success because of the engineering alone.

[Mike Watterson]

Where did I dis the LM3875 or LM3886? I was dissing stupid prices and the idea that it was unexpected.
But there were excellent hybrid modules before such IC amps existed for easy DIY and such a development of a decent IC was inevitable.
I'm NOT dismissing the IC, but dismissing the stupid name and price of the original product using the IC. Decent ICs have been available for years now, so much so that one used in the Gaincard is obsolete.
I've never despised decent IC amps or Raspberry Pis, I do despise most Audio magazines and Audiophools since maybe 1980s. It's all about fashion.

I mean op-amps were once made with valves and then discrete components. Of course monolithic IC amps can be totally HiFi, since ages ago.

[Refugee]

I spent a fiver on a small HiFi for the CD player from the big rummaging table at our local flea market.
The intention was to tap into the audio before the amp and use it like a separate CD player.
I got it home and obtained a scan of the manual. The manual even shows the block diagrams of the chips. It is a JVC set.
That was when I realised why it began to feel heavy on the way back to the car.
I looked up the audio amp chip and the distortion figure was more like that of a preamp. You bet that was for running the reduced load with the headphone resistors in circuit.
I wired up a 3.5mm plug with 33 ohm resistors hidden in the other end of the lead.
The amp chips are TDA7294. They look the same as the others we are chatting about but with a MOSFET output stage in the same package style.
It pluged and played and the cassette even worked in both directions. It is auto reverse.