Chinese-sourced vintage chips, real or fake? 

Send the dies to Sam Zeloof as a challenge. If anyone could repackage them back onto a DIP, he would be the most likely to succeed :)

11:45 I guess he wants to keep the original hardware, but wouldn't it be possible to emulate the clockchip output bits correctly with programmable logic?

@@manuell3505 Not many programmable chips matching the peculiar voltage and power characteristics. That said, back in the day Ti had an official digital clock chip with LED drivers, apparently based on a mucrocontroller. It was used in at least one magazine published "build your own digital clock" project.

Yes, and I for one would love to see the "Donald Duck" drawings, and the like, on the edges of some of those chips. Bored engineers.

Exactly. I really wonder why they did not do it. Maybe worried about the extra leakage current and wanted to keep it at an absolute minimum?

@@CuriousMarc Possibly. Since that lack seems to be endemic to this sort of stuff no matter who made it, I more think that it somehow just didn't occur to the engineers that the battery could fail, or maybe it got cost reduced out by a manager. On the other hand, low voltage zeners have really sloppy regulation curves, and might well draw significant standby current at 2.5V. A stack of forward-biased diodes might be a better choice. What's the charging voltage on those batteries? It must be more than 2.5V. Hopefully a clip around a half volt over the charging voltage would be safe.

l wilton I reckon it was assumed by the time the problem occurred the product would be end of life.

@@CuriousMarc I expect the charging circuit might feed the battery with a diode, so it doesn't drain it while off. In which case you could clamp the other side of the diode and not increase the leakage while off.

Low voltage zener diodes are effectively useless for this kind of protection as they have such a soft knee. This means it either clamps at a higher voltage than you want or you have to accept a considerable leakage current at the nominal voltage.

Did you happen to be involved with making any of the VLSI chips that Tandy used?

@@BlackEpyon , sorry no I wasn't. I was a research assistant at the University of Waterloo's VLSI group. We were figuring out how to design them. The chips were following Moore's law so the capability to make lots of transistors was outstripping our ability to make designs with lots of transistors ...

@@ksbs2036 The reason I asked, is because the VLSI chips on those retro computers are pretty much black boxes. The service manuals told you what was in them, and the pinout, but not the schematic diagram internal to the chip itself. They were pretty reliable (I've never seen one go), but if one happened to go, you'd have to find a donor board to salvage from, 'cause there's no way to replicate them.

These are more early-mid 1970s vintage, right?

Working voltage too high .. Wasn't that the case for MOS in general at first? I mean, early NMOS and PMOS chips ran at basically 17V, the +5V/-12V for PMOS and +12V/-5V for NMOS were only to get (somewhat) TTL-compatible logic levels? The 4004 for example doesn't even have a "GND" pin? The couldn't implement depletion mode FETs on the same die as the enhanced mode back then, so that the current source FET used considerable voltage if I understood it correctly?

Not necessarily fast. It depends on the speed of the clock.

Well you got me. You're right!

Now we know that Ken is hardware-optimized to identify unusual 1970s IC technology.

@@UnitSe7en I was going to say the same thing

Ugmmmm ......... Ken seems to be a Schottky fast dude. .....!!!.

AvE........😆😆😆😆😆

Practice, and patience...

You can try sandpaper and a lot of patience too. Start coarse and work your way down, like polishing. Then scrape off the last part with a plastic spudger. I've got a video on my process on my channel.

This playlist has the videos about the Fridens: ru-vid.com/group/PL-_93BVApb5_GbzkFwdKpuyPo8Ty83h8b

Is there any chance you reveal your trusty vendor?

I don't believe so. It is a bipolar process so it doesn't have the gate insulator that gets destroyed when E-field go to a million volts per metre. (1 volt across a micron)

It was a lot more sensitive to process errors though, especially in the early days of IC manufacture, as it relies on barely turning diodes on and off to do logic. Power hungry as well, so those early wristwatches had a very short battery life, even if you did not turn the display on.

@@MikinessAnalog IRFZ44 likely has, in addition to humongous internal capacitance in the gate to absorb ESD current, a set of zener diode clamps to limit gate voltage. The clamps are around 20V, so they do not otherwise affect the decice in operation, and the capacitance is why the drivers need to be so robust if you want fast low power dissipation switching. DC to a kHz you can use a regular logic gate drive, but for higher frequency the current peaks required increase quite a bit.

Actually, I thought exactly the same ;)

That was my first thought. Since the chips appear real and the fault it so common it would make sense. We will have to wait for them to make a test board to find out but my money is on the chinese chips being faulty.

I was thinking the same as they're probably pulls from recycled hardware. Too bad HP didn't add over voltage protection circuitry for the chip. But then they probably figured the battery would last longer than the service life of the equipment.

@@Membrane556 Exactly what I thought. They pulled the chips from faulty HP RTC moduls or other HP equipment. Maybe they simply didn't know about the problem and assume the chips were working.

Very high, it's my understanding all these vintage chips out of China are at best pulled from scrapped gear and sold with no testing, maybe erase the EPROMs. At worst they remark them as a faster or more favorable variant.

@CuriousMarc were the chips zapped? We couldn't tell from the microscopic images.

they sometimes use low-grade opamps marked as precision ones. They also make functional copies of popular chips - search for "FTDI chip gate"

@@gkasprow Yes, and a lot got caught out by that too, bricking their devices.

So you measured the behaviour of existing circuitry to refine your own product? If it was from a different manufacturer who was unaware of it, isn't that reverse engineering?

@@manuell3505 The company had an in-house simulation tool. Which could simulate for example the influence of supply voltage, temperature or deviation in manufacturing accuracies, in the performance of I2L logic. The accuracy of the simulation results depend of course on how accurate the simulation model is. So the characteristics of the pn-junctions and the transistor behaviour were measured using special test structures and then used as parameter settings in the simulation models. This is the work I had to do: measuring those test structures, finding the correct parameter values, and then verify the simulation results with actual I2L performance. It had nothing to do with reverse engineering chips from other manufacturers.

@@windmill1965 Allright, that's just an independent research/production circle.

Great idea, Ti is keeping lots of old designs available as normal stock numbers, so they may well be keeping other historic chips on file or in stock, but on request only.

I have my doubts.

Yes, silicon cleavage does get one’s attention.

hehehehe , love the "Master Ken". comment ... the guy deserves respect !

@@thebigpicture2032 Google gets many search requests for this particular thing, though. Mostly from us Germans ;).

I was thinking that too, it would be interesting. But it might not even be on the die where it melted.

No, could not see anything obvious.

Failure analysis is hard.. dealing with it at my day job. There a whole host of testing and analysis methods including thermal imaging and SEM. Fascinating stuff though.

@@CuriousMarc, thank you for your reply.

I heard in an _Adrian’s Digital Basement_ episode that many of the authentic chips from China were acquired through e-waste processing.

Handwritten voltage of only 2.5v... wow!

Maybe blowing a BGA LED driver to save the fuse?

I get fake transistors too. Also I got fake resistors about a year ago - literally they'd put correctly coloured plastic on a piece of wire. I was like "what the actual fuck".

The interesting ones are fake FETs I've bought that are fake, but still functional and similar characteristic. Must be made from a Chinese manufacturer to similar specification and just remarked for higher price.

@@medes5597 i bought a 2500pcs pack of 1/4W and they were not metal film and not inside the 1% stated obviously, more like 2-3%, but i use em

@@jkenny1 if they are fake they quite likely withstand less current, they put a smaller die inside and that's how they scam you, buying from a more costly seller seems to not solve the problem, weirdly they sometimes provide you good parts, so the price may be near to the correct one, i would be perfectly happy to pay 1.5X times the price but have good parts, another funny fact is that they CAN provide you LEGIT parts but without fancy brands, infact inside their gadgets they usually fit legit unbranded parts, i also found a custom made transistor, if you buy an high voltage module like the 15kV one, between the components there's a BJT with package TO126, but to copy the pinout of a TO220 that has been replaced for cheapness they made an unmarked, unknown TO126 transistor with pinout b-c-e like a TO220 that as far as i know doesn't exist, TO126 are all e-c-b, so they can make a custom transistor just to sell some junk modules

the leakage current of the voltage regulators back then were well over 10uA which was way higher than what the clock chip drew

@rustybuttpate I had nearly the same experience. Damn....we are old now aren't we?

More than likely Dismantled in india and sold on to china

@@Beatnik-Manor You obviously frequent the same places in Bengaluru as I do when looking for bits for keeping old plant electronics alive :)

I think neither. They probably could not tolerate the Zener leakage.

@@CuriousMarc And if it was on a maintenance contract your friendly local HP customer engineer would replace it at every PM. If you didn't have a maintenance contract it would cost you an arm and two legs for it to be replaced and serve you right according to Bill and Dave :)

Pretty sure that would make them stop working.

@@BlackEpyon Applied Science does it with a 555 timer: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-eoRVEw5gL8c.html

@@BlackEpyon its a technique thats used to test chips. Though it depends on the technology if they can withstand it. Also, a proper electron microscope dose not destroy chips as easily as applied sciences hacked together SEM did, if you where basing that off of his video.

@@lbochtler Applied Science got himself a proper SEM too. I was thinking more along the lines of the die being bathed in electrons and just washing out. It wouldn't be able to function WHILE it's being scanned, but should be able to function afterwards. I think.

@@BlackEpyon The technique that is normally used is EBIC (Electron beam induced Current). You wont see it run in real time though, as you need to scan rather slowly. Yes, i know Applied Science has a proper SEM as well, but i dont recall him trying the microchip experiment using it

the leakage current is waay too high to use zeners or leds. .

Probably the thickness or composition of the passivation layer on top is different.

I looked at them closely again, they appear to be, new old stock, genuine chips. If they had been in some previous equipment, then they were cleaned up very well.

Here: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-ZQeHHYJYWXo.html

There are many contaminates that slowly diffuse into the semiconductor killing it over time. The contamination migrates faster at elevated temperature. It migrates even without power. Many warehoused parts are not in climate controlled conditions. When CPU's went to copper interconnects, it was discovered Copper is one of the contaminates. The process is very specific to copper plate and plane the ILD side of the chip while protecting the semicondictor side from copper contamination. Many of the vintage chips did not have the knowledge of the migration contamination failures until much later. Simple copper pipe used before water quality issues were discovered, resulted in chips with slowly migrating poison to the transistors. Contamination can be from the process with some trace copper contamination to die cutting, to packaging. For more info.. www.smta.org/chapters/files/Philadelphia_SMTA_Philly_Rev2.pdf

That’s the case for this one apparently. Real but broken.

oui. C’etait une vraie.

Géniale, je chercher une IC obsolète, Reticon SDA1024A, malheureusement le chinoise ne fabriqué même pas de faux 🙁 (Désolé je suis Anglais mais j'aime bien de utilise mon language française des temps en temps)

@@dm4728 Bravo pour le français!

That’s what I feared too, but in a later episode ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-pr6HTiWrMmk.html I built a LED watch chip tester, and they all tested good!

Probably they go open circuit and the charger / power supply is relying on them to pull the voltage down to a safe level for the rest of the circuit.

@@jg374 Ah of course, that makes more sense. Thanks.

@@jg374 You would expect hp to have done a bit better job to prevent that. It wouldn't have taken much.

@@MrFathead Just a simple zener diode to be exact.

@@nrdesign1991 cant use a zener in that circuit. leakage current is waay too high. These clock chips typically ran on something like 100nA back then. Maybe as much as a 1 uA

Was quite common. Some portable tube radios had the same problem, the tubes in it were specified for 1.4V heating voltage and there was a single NiCd cell in there, which was recharged from the mains if used at home. If this cell did go bad or was missing, the voltage shot up and blew ALL the filaments. When this was designed, its quite possible that the people who designed it were still very familiar with this arrangement .. it was just 10 years ago that those radios finally got out of favour.

@@VintageTechFan Well to see other brands do it wouldn't have been surprised, but HP have the best tech and should address it. Only a few components would have fixed it but they didn't even do that, no matter if it was 40 years ago or not. But who knows, it might back then have been and design thing to destroy it after some time.

damaged pcb is very easy to notice due to the chemicals in the battery discoloring the pcb

Zener diode