Tayda 4558 vs others

[matmosphere]

For both their 4558s and TL072s, I've gotten both legit looking ones and off-looking ones (maybe counterfeits). When I was placing orders regularly, I would sometimes throw in one or two of each  just to see what they were currently shipping and if what I got was good, and then I would immediately place another order for more to stock up. Seemed like I would usually then get a the good ones. They were cheap enough that doing it this way saved quite a bit of money over buying from Mouser, even if I wound up getting a few large orders of fake looking ICs.

This gets exactly to the point I was asking about. How could you tell fake vs real ones and did you hear a difference between the two?

The four or five pedals I have tayda 4558's in don't sound terrible, I'm just wondering if they'd sound better if I replaced the ICs.
I hadn't intended this to be another referendum on tayda parts. I've used them for almost all of my passives and very little is ever out of spec other than a few shoddy diodes. I've actually gotten to the point that I don't bother checking my box caps and most of the diodes anymore.

[wgc]

I have had the same experience with the tayda 4558 and tL072s, but fwiw I use the smd versions.  I had a number of TL072s that were simply bad- I think I bought 10-20 and after 4 were doa, I threw out the rest - likely an ESD issue on that particular batch.

The 4558s worked, but sounded dramatically different than the ones I got from mouser/digikey.  My first batch were from Tayda, and I figured I just didn't like the 4558s in my builds.  Once I got them elsewhere, I was pleasantly surprised.

I think they buy B stock when they can and (most) people, who aren't intentionally overdriving them, probably wouldn't notice any difference.

[reddesert]

Some of you are electronics professionals and know more about the industry than me, so correct me where this is wrong.

The industry has a practice called "second sourcing." Because product designers don't like to use a part that relies on one source only, component manufacturers will license a design to each other. So you can buy an IC from makers X and Y where both are totally legitimate, though the manufacturing process might be different.

There's also copying. The original 4558 appears to have been made by Raytheon in 1974, and I imagine any patents are long expired, so other manufacturers could make a work-a-like. I don't know if the designation "4558" was or is trademarked, but would guess by now, it doesn't require a license to call your work-a-like a 4558.

Mouser has 4558's from 4 different sources and Tayda has 3 sources, and I expect they're all "real" by these categories. The question of what specs they meet would be harder to answer, although both provide datasheet links.

Finally, there's actual counterfeiting. This is like when somebody sands or blacktops and re-labels ICs, whether to forge date or spec codes or the entire designation. This is a real problem in the electronics industry because it circumvents parts certification, in addition to the possibility of things being DOA. I assume that this is where ebay fake CA3080s and other such things come from - they could just be some random chip relabeled CA3080, and be totally unusable.

The electronics industry is very concerned about faking high value or unusual parts. It seems unlikely many shysters would counterfeit a 4558, because they'd be turning a 20 cent IC into another 20 cent IC. They could be diverting parts that failed QA but were not destroyed into the supply stream, I guess, but have no idea how likely that is.

[chris weaver]

I think it's a get what you pay for thing, I like the mouser/small bear/arrow quality, but there are times I buy in bulk from tayda.

[alanp]

There's also copying. The original 4558 appears to have been made by Raytheon in 1974, and I imagine any patents are long expired, so other manufacturers could make a work-a-like. I don't know if the designation "4558" was or is trademarked, but would guess by now, it doesn't require a license to call your work-a-like a 4558.

IANAL, but I strongly suspect that you can't trademark a number.

It may be urban legend, but I heard that the reason why Intel changed the name for it's processors in the upgrade from the 486 era, to the Pentium era (and not the 586), was because it could not stop competitors building dirtyroom/cleanroom workalike processors, and calling them 486's. My family used to have a PC with a Cyrix 486DLC processor.

[pickdropper]

Some of you are electronics professionals and know more about the industry than me, so correct me where this is wrong.

The industry has a practice called "second sourcing." Because product designers don't like to use a part that relies on one source only, component manufacturers will license a design to each other. So you can buy an IC from makers X and Y where both are totally legitimate, though the manufacturing process might be different.

There's also copying. The original 4558 appears to have been made by Raytheon in 1974, and I imagine any patents are long expired, so other manufacturers could make a work-a-like. I don't know if the designation "4558" was or is trademarked, but would guess by now, it doesn't require a license to call your work-a-like a 4558.

Mouser has 4558's from 4 different sources and Tayda has 3 sources, and I expect they're all "real" by these categories. The question of what specs they meet would be harder to answer, although both provide datasheet links.

Finally, there's actual counterfeiting. This is like when somebody sands or blacktops and re-labels ICs, whether to forge date or spec codes or the entire designation. This is a real problem in the electronics industry because it circumvents parts certification, in addition to the possibility of things being DOA. I assume that this is where ebay fake CA3080s and other such things come from - they could just be some random chip relabeled CA3080, and be totally unusable.

The electronics industry is very concerned about faking high value or unusual parts. It seems unlikely many shysters would counterfeit a 4558, because they'd be turning a 20 cent IC into another 20 cent IC. They could be diverting parts that failed QA but were not destroyed into the supply stream, I guess, but have no idea how likely that is.

I've purchased a lot of components over the years.  As a hobbyist and at my day job in engineering.  Typically, places like Mouser, Digikey, Arrow, Newark, etc.. have many different variants of a typical part.  It's necessary when you are designing a product.  Nobody wants to shut down production because a specific 100n cap has a 16 week lead time.  In our PCB layout software at work (Altium) the BOM for each board lists approved alternates for this very reason.  For something like caps and resistors, it's not unusual for there to be 5 approved variants that are all basically equivalent.

All of those vendors listed above are primary distributors and generally their stock comes directly from the manufacturers (although I'm sure there are exceptions).  But all of it is vetted and comes through known channels.

Vendors like Tayda often deal with back channel parts of vague provenance.  You see this with a lot of eBay vendors as well.  They get their parts through brokers at the lowest prices they can.  The parts might be good, they might not.  They are definitely cheaper, so one can give em a shot and deal with the results.  It's certainly not a crime to use them.

Tayda can be good for hobbyists as it allows people to build up a bank of parts for a very low price without having to buy quantities to get price breaks.  Once people step up, I always recommend Mouser because it's often. not much more expensive for things like passives if you are willing to buy 50-100 of something.  Of all the above vendors (Mouser, Digikey, etc..) I've never seen anything like the issues I've run into with Tayda.  They are fine, but definitively a tier below the real suppliers.

As far as counterfeiting, who knows?  As you've mentioned, there isn't a huge financial incentive to copy ubiquitous parts.  There is far more incentive to dump out of spec parts to secondary vendors as a means of recovering some yield loss.  As long as they are honest with the vendors/distributors, there's no problem with it.  At that point, the onus is on the vendors to inform their customers that the parts are out of spec.  Some do, some don't.  Some may not even know because their distributors don't inform them (I actually suspect this is the case with Tayda).

[pickdropper]

I have had the same experience with the tayda 4558 and tL072s, but fwiw I use the smd versions.  I had a number of TL072s that were simply bad- I think I bought 10-20 and after 4 were doa, I threw out the rest - likely an ESD issue on that particular batch.

The 4558s worked, but sounded dramatically different than the ones I got from mouser/digikey.  My first batch were from Tayda, and I figured I just didn't like the 4558s in my builds.  Once I got them elsewhere, I was pleasantly surprised.

I think they buy B stock when they can and (most) people, who aren't intentionally overdriving them, probably wouldn't notice any difference.

I can't say I've noticed a huge difference between 4558s, but I did have a Tayda batch that sounded decidedly different.  It was a number of years ago, so that may not be the case anymore.