Suspected fake/counterfeit NEC V20 [uPD70108HCZ]

[The_YongGrand]

Hey there,

I'm testing those chips I bought from eBay chinese retailers, and I fed the clock signal (4MHz) into the processor itself, and there is no ALE/ASTB strobing at all.

From what I read in the http://helmpcb.com/electronics/8088-computer, the author said that to make sure if it's running, the ALE pin must be checked in the oscilloscope. I have followed the rest of the wiring as seen in that website and the www.homebrew8088.com, but still the pin doesn't strobe or do anything.

I begin to suspect that this chip is fake. Here's the pic if you need to see it:

I would like to hear the opinions.

 

[MicrocomputerSolutions]

I'm not surprised that it doesn't work as it's supposed to.

A lot of Chinese chips are defective rejects, pulls from salvaged equipment. I refuse to buy any chips from China after buying (300) Hitachi static ram chips that were supposed to be new from two different vendors. Almost 30% were defective and it took me two weeks to sort out the working chips from the defective ones. When I messaged the Sellers about the defective chips, they sent me more chips and the process started all over again with the same result 30% of the new chips were defective.

Counterfeit chips are often refinished, painted, sand blaster, etc, and then remarked. There is no telling what chip is under the new finish or even who made the dies or encapsulated the chips.

The price of Chinese chips is very tempting, until you get them and find that they contaminate your chip supply making your supply of chips unreliable, if you allow the chips to mix.

Often times it will cost you more to send the defect things back than you will receive. The Chinese Sellers know this, and send defective parts knowing that there is little chance that they will have to make good or refund.

I would not expect work a like chips like the NEC V20 to work exactly like the real thing.

Compupro tried using the V20 and V30 in place of real Intel Processors on a couple of their S-100 boards. They found that the V20 and V30 are not 100% compatible with Intel Products and they ended up replacing all the V20s and V30s with Intel processors on board that left the factory. AMD processors had problems too, and Compupro replaced the AMD processors that they supplied on Disk-3 hard drive controllers, and CPU-8085/8088 and CPU-8086 processor boards.

Take the time to email the Seller and tell him that the part is defective, and you want your money back. He/she will probably offer you another instead of a refund. Insist on the refund, and tell them that they should pay for the return shipping as well since the part is defective. With any luck they will return your money and tell you to keep the thing. Those people do not want to catch any attention, and most will go out of their way to shut up the few people that bother to complain.

 

[The_YongGrand]

I'm not surprised that it doesn't work as it's supposed to.

A lot of Chinese chips are defective rejects, pulls from salvaged equipment. I refuse to buy any chips from China after buying (300) Hitachi static ram chips that were supposed to be new from two different vendors. Almost 30% were defective and it took me two weeks to sort out the working chips from the defective ones. When I messaged the Sellers about the defective chips, they sent me more chips and the process started all over again with the same result 30% of the new chips were defective.

Counterfeit chips are often refinished, painted, sand blaster, etc, and then remarked. There is no telling what chip is under the new finish or even who made the dies or encapsulated the chips.

The price of Chinese chips is very tempting, until you get them and find that they contaminate your chip supply making your supply of chips unreliable, if you allow the chips to mix.

Often times it will cost you more to send the defect things back than you will receive. The Chinese Sellers know this, and send defective parts knowing that there is little chance that they will have to make good or refund.

I would not expect work a like chips like the NEC V20 to work exactly like the real thing.

Compupro tried using the V20 and V30 in place of real Intel Processors on a couple of their S-100 boards. They found that the V20 and V30 are not 100% compatible with Intel Products and they ended up replacing all the V20s and V30s with Intel processors on board that left the factory. AMD processors had problems too, and Compupro replaced the AMD processors that they supplied on Disk-3 hard drive controllers, and CPU-8085/8088 and CPU-8086 processor boards.

Take the time to email the Seller and tell him that the part is defective, and you want your money back. He/she will probably offer you another instead of a refund. Insist on the refund, and tell them that they should pay for the return shipping as well since the part is defective. With any luck they will return your money and tell you to keep the thing. Those people do not want to catch any attention, and most will go out of their way to shut up the few people that bother to complain.

Thanks for taking the time to explain about this. Basically, I have been quite lucky purchasing bulk items from Chinese retailers in eBay which are modern items and bulk supplies of modern components.

However, this is the first time I got my hand burned buying vintage ICs from those places. Some years back I purchased clones of arcade sound chips from Taobao (clones of GI AY-3-8910 and YM2413) and they worked perfectly, but the retailers are very hard to communicate, and I do not speak Mandarin at all (and even that my Cantonese is limited), but I got these all-okay.

I bought these V20 last year, and I do not wish to get a refund since all the warranties are already expired.

I know it's a stupid question - how could I actually get an actual V20HL? How does this working, actual V20HL look like? The V20HL looks very powerful on the datasheet, but I do not want to gamble on the eBay again as I have a higher possibility of getting a fake one again. Or, I must stick to buying a normal Intel 8088?

 

[MicrocomputerSolutions]

Look for a US Based Source that is not an Asian Company (some Asian Companies use a US Address, but still ship the junk from China) for all ICs. Although it costs a little more sometimes, I try to buy all processors, memory, TTL and discrete components from large/major US based, Suppliers. I've learned my lesson the hard way.

There is nothing wrong with the Real "Intel" Processors.

The AMD equivalent Processors will test slightly faster than Intel Processors when using the Intel machine codes.

The NEC processor and disk controller chips also test faster than Intel chips at the same clock.

Are you thinking about playing around with unsupported AMD and NEC machine codes? You may get strange results.

The AMD and NEC chips sometimes give/gave different results than than supplied by Intel chips. Which is why Compupro stopped using NEC and AMD processor and controller chips. They actually sent replacement Intel chips to me to install in Customer Compupro Boards (free of charge) when they were in for servicing. That says something about what Compupro found out about running other brands of processor and controller chips.

I have some NEC, and AMD processors, and disk controller chips here (NEC 765, Intel 8272 floppy disk controller equivalents) somewhere in storage that were removing during service from Compupro processor and disk controller boards that were never returned to Compupro (My Bad). I know they are not the fastest parts available, because Compupro bailed out of the compatibles before the higher speed parts became available. They are Genuine Parts, if you really want to experiment. What are you willing to pay for them?

If you want to make sure that you're getting Genuine Parts, find a US based source and spend the $10 (each) that the new processors are going for, here in the US, instead of the $2.50-$3 they are offering to sell them from China for.

 

[Chuck(G)]

I don't know--I don't think the V20 is fake--the lettering looks too good. Usually, you can see something sleazy like a coat of black paint, over-lettered. V20s don't command much of a premium even when they're NOS.

It might well be defective. Do you have a working 8088 system that you can check it in?

 

[deathshadow]

the lettering looks too good.

That's....well, part of the problem. It looks too good to be a real one. It's the more modern late 90's lettering style, the text is the wrong colour, and that uniform grainy texture is a typical result of a walnut blasting. I've NEVER seen a real V20 that looked like that...

This is what a real V20 usually looks like -- they're pretty uniform in labelling and packaging when it comes to plastic.

The package is smooth, not grainy. It uses the old NEC logo in a cheap white ink. While the four in my collection are -8's with a single -10, they're pretty uniform. See CPU world:

http://www.cpu-world.com/CPUs/V20/MANUF-NEC.html

That top texture and labelling style didn't become commonplace until they started making PGA and QFP flavors like the V50's. I very much doubt it's legit. It's TOO good to be a real one.

Though the HCZ part is a bit odd... would indicate a late low voltage model meaning you plug it into a regular PC, it could go POP... in fact, are you sure you're not feeding it too much voltage? I know a few people make that mistake with the HL versions. (assuming you can find one that's not LCC or QFP)

Can't say I've ever heard of a "HCZ" model, though it appears a LOT of chinese sellers are listing exactly that...

 

[MicrocomputerSolutions]

What's that stuff hanging off the side close to the legs?

Looks like the liquid masking stuff (Magic Mask?) we used to use on the windows of plastic models to keep paint overspray off.

The Hitachi memory chips that I got from vendors with two different names looks as perfect. As a rule, the printing and finish on Hitachi chips is pretty high quality and uniform, but those chips did not look exactly like other chips of the same age, and the chip legs had a different appearance (wrong shade of silver?).

As the previous Poster noted, the writing is too modern, and too perfect compared to other NEC devices that I have here. The writing is all white in color, pretty faint (not strong or bright) on the examples that I have, and poor in quality of application. The ones I have all look similar to the Previous Poster example (but more faint) and not the OP's example.

As for the processor not being a fake because it's not a high value part? I've seen common TTL parts that are fakes, and what's cheaper than those?

 

[Chuck(G)]

I'm still not convinced of a fake. The CPU Chips collection all show bog-standard older V20s out of NEC USA (Natick, MA). This is a later chip produced and marked by "NEC Japan" and the labeling is perfectly consistent with more modern NEC chip labeling.

This is the "H" series chip--Datasheet here. You'll note that the HCZ-16 is a perfectly valid version of the chip.

If you'd like, send me one of your suspect parts and I'll test it in a working system.

 

[daver2]

The pins that I would expect to be wired up appear to be on your breadboard.

What have you wired each pin to though? I guess black is 0V and purple is +5V for starters...

My recollection of 8086's were that if they were fed 'silly' opcodes they would 'disappear off into outer space and never be seen again'. My suggestion would be to add an 8-bit tristate buffer to the ADx lines and hardwire a 'sensible' opcode (like NOP) to the input side of the tristate buffer. Only enable the buffer when a MEMORY READ is requested. This will guarantee that the CPU sees a genuine OPCODE (NOP) every time it tries to execute an instruction.

Take it from there...

I have also breadboarded up CPU circuits before as tests - and they are somewhat 'prone to working - not working - working - not working' if you get my drift... Make sure the connections are 'good' and add a decoupling capacitor or three to the 0V and +5V ins of the CPU [just seen some on the upper edge of the image].

What power supply are you using for the CPU?

EDIT: Why is pin 1 wired up? On the data sheet it seems to be classed as IC (Internal Connection).

Dave

 

[Chuck(G)]

@Dave, I tried to make sense of the wiring shown in the photo also--and failed. That, and the knowledge that the H-series V20 is a very real product led me to believe that the OP's working with the echt V20H, but doesn't quite grasp how to make it work.

I'll be delighted to be disabused of this notion, however.

 

[daver2]

By the looks of the photograph the OP has the following pins wired:

1 - IC (Internal Connection?)
17 - NMI
18 - INT
19 - CLK
20 - GND (POWER)
21 - RESET
22 - READY
23 - /POLL
25 - ASTB (OUT)
31 - HLDRQ
33 - S/LG (wired to +5V = SMALL)
40 - +5V (POWER)

The OP has wired all the inputs up (as I would have expected) - the question is "to what voltage levels" or "to what source".

I think we are saying the same thing - "give us some more details on how you have actually wired the pins up as shown so we can see if they are what we would expect".

Dave

 

[Chuck(G)]

...which is why I suggested testing the chip in a known-good 8088 system.

 

[The_YongGrand]

By the looks of the photograph the OP has the following pins wired:

1 - IC (Internal Connection?)
17 - NMI
18 - INT
19 - CLK
20 - GND (POWER)
21 - RESET
22 - READY
23 - /POLL
25 - ASTB (OUT)
31 - HLDRQ
33 - S/LG (wired to +5V = SMALL)
40 - +5V (POWER)

The OP has wired all the inputs up (as I would have expected) - the question is "to what voltage levels" or "to what source".

I think we are saying the same thing - "give us some more details on how you have actually wired the pins up as shown so we can see if they are what we would expect".

Dave

Sorry for the late reply and lack of information.

The voltages applied on the supply is +5V and already tested with +3V too since the datasheet of the V20HL said it's okay to run at +3V.

The source is from a breadboard power adapter. I will change to a better power supply unit once I get hold of it.

 

[The_YongGrand]

The pins that I would expect to be wired up appear to be on your breadboard.

What have you wired each pin to though? I guess black is 0V and purple is +5V for starters...

My recollection of 8086's were that if they were fed 'silly' opcodes they would 'disappear off into outer space and never be seen again'. My suggestion would be to add an 8-bit tristate buffer to the ADx lines and hardwire a 'sensible' opcode (like NOP) to the input side of the tristate buffer. Only enable the buffer when a MEMORY READ is requested. This will guarantee that the CPU sees a genuine OPCODE (NOP) every time it tries to execute an instruction.

Take it from there...

I have also breadboarded up CPU circuits before as tests - and they are somewhat 'prone to working - not working - working - not working' if you get my drift... Make sure the connections are 'good' and add a decoupling capacitor or three to the 0V and +5V ins of the CPU [just seen some on the upper edge of the image].

What power supply are you using for the CPU?

EDIT: Why is pin 1 wired up? On the data sheet it seems to be classed as IC (Internal Connection).

Dave

Hey there,

For the wires, I do not have a red wire for the +5V, so yeah, I had to make do with my available wires in my place.

The capacitors are already there at the power rails but not seen in the picture - I took a close-up view of the processor only.

Again, I'll try to add the remaining support chips like you have suggested and test it with the NOP first.

Also, I'll take out the pin 1 wire - on the datasheet it said about "internally connected to GND" and I had misread it.

...which is why I suggested testing the chip in a known-good 8088 system.

I have known someone who has an 8088 board in my nation. I'm speaking to him right now and see if I can lend him one of the chips and let him test it.

EDIT: Okay, I've wired this thing to the '245 and placed the 0x90 [NOP] at the other end of the buffer, and there is strobing at the ASTB. Thanks to the retro-computing veterans!

As I do not have a logic analyzer on my hand, it's impossible to know that if it will start at FFFF0. This processor is running at 8MHz, by the way.