What is this PCB component

The photo is of such poor resolution that it it is impossible to see the component, the tracks to it or any of the legend in the IC to which it is connected.

The component may be a power supply decoupling capacitor so on no account short out the pads. If it is a capacitor then i) it probably will not make any difference to the operation of the circuit ii) shorting the pads may short one of the power supplies and destroy other components on the board.

As has been mentioned I don't think there is sufficient detail to be sure but if it's only two legs it should only be a cap, resistor or diode. ;-)

Probably definitely not. ;-)

Why don't you (try to) measure the value of the one you still have. If you are lucky it will be a resistor and read a preferred value, like

You sometimes have to be careful measuring stuff in circuit, less you damage anything (sensitive circuitry) around it.

One with a reasonable fine tip yes. If your eyes are good enough, you can sometimes put some flux on the area and (quickly) flood the component with solder (so it bridges it) and then you can (quickly) lift it off with some fine tweezers.

Yeah, not generally an issue re which way round if it's a resistor but might be if it's a cap and deffo a diode. ;-)

We will have to wait and see. ;-)

Cheers, T i m

I had to delete the -- from the URL, but by the looks of it that's the firmware flash chip for the processor which has the thermal pad on it (which obscures its part number).

Assuming the flash is the standard SPI NOR flash pinout, that would be pin 7 which is HOLD# - see eg:

Theo

Usually on such boards resistors have values marked but caps don't. Your component is certainly not a link as there is no track underneath that requires a link to bridge it. It looks like the same kind of cap as is across the narrow pair of tracks under your arrow that go into the corner of the square IC. Probably you could use that to replace the missing one.

The pinout seems consistent with a SPI Serial Flash memory where pin 3 is a nWP or write protect pin.

There might be a internal block protection bits that is used in conjuction with this pin.

If the device is pre-programmed then the level shouldn't matter, otherwise the pin has to be held high and I would expect the resistor is passively holding the pin low to prevent inadvertent programming.

Without more details of the device the above is an educated guess.

It looks to me like this. Your second picture shows 3 is joined to 7, so /WP is joined to /HOLD, and it's like the functions are not being used. They would be going to a pullup to put them in the negated (not used) state.

This is just to show the basic idea.

I can't find an exact match for the part number on Winbond, so this is close to it but not the precise item.

Typically, people seem to remove some ROM from controller boards, and transfer it from the old board. There's some kind of check that the controller and disk are matched somehow. Forcing people who do controller swaps, to bring over the old chip. Now, if the chip was dead, I don't know what they're supposed to do with regard to recovery. Each SPI flash has a custom (Winbond factory) written serial number, so it is possible for a check to be done.

The code in the flash, isn't the whole code load. The service area has code too. The code in the flash is sufficient to move the arm over to the service area and read the rest of the code content. It's something like that.

If you copied code from one flash chip to another, the serial number of the second chip would still be wrong. It wasn't always necessary to swap those chips, but it's become a "thing" on the more modern drives.

This is information I've been able to glean from a couple data recovery forums on the web. They don't come right out and explain stuff, so you have to collect enough tidbits from what they say, to spot a pattern.

Not all the controller boards are recovered from dead drives. Some of the controller boards are available as brand new items. But that still does not remove the need to swap SPI, if that's what the design era calls for. The 4GB WDC I have here from the year 2000, would not need that. A straight up controller swap would work there. The controller board is likely around 6 years old (just a very rough guess).

Paul

It's not unusual for zero ohm resistors to be used as links, but the photo isn't good enough to make out any markings, a camera with a macro lens and some strong lighting might help ...

No don't short it, it might be a capacitor. I hate surface mount. Brian

I've always found "zero ohm resistors" a rather wierd concept, surely that's either a link or just a piece of wire. Why call it a 'resistor' when it's not?

It's a surface-mountable wire, which fits into the whole manufacturing process. It can be placed at high speed, and doesn't slow down the pick-and-place when used. The SMT item can be affixed with a glue dot like any other resistor, and won't fall off going through the line. What's not to like ? You put a reel of 3500 of those on the machine, press the button and off it goes.

Generally you try to use those, in places where it won't affect circuit performance. It can be used on a Bill Of Materials (BOM) to control the addition or subtraction of optional circuit blocks.

And just because it's zero ohms, you don't go around running 20 amperes through it on purpose :-/ Although some of my colleagues have probably thought about doing that. Because they are zero ohms and its magical. They're like tiny superconductors, right ? I can think of at least one engineer who would fall for that. And no, I've never ohmed one out to see what its real resistance is. I'm sure there's an estimate around somewhere.

As another example of a function, I could put a zero ohm today, and if I get in the lab and discover a logic signal needs series damping with a 33 ohm, change the BOM to read

You can even do zero ohms like this, but manufacturing is going to be phoning you if they spot stuff like this. The rules set for PCB design, is several feet thick. I'm sure that breaks some rule or rules. You can do stuff like this if the land is square, and rotating a component is footprint compatible.

R1 --------X X--------- destination #1 R2 X | | destination #2

If in the BOM, I set R2 to zero ohms, then my signal is going off to destination #2. And then I depop R1 so that leg is open circuit. Voila, circuit change after schematic capture is finished. I can put R1 or R2 in the BOM, but not both.

The opportunities for abuse are endless. Especially if the engineer doesn't document what all that crap is for.

On things like computer motherboards, you'd remove stuff like that for the third and final design cycle, before high volume manufacturing starts. Because a zero ohm resistor costs money and "it has to go". But on "material cost is no object" circuit boards, where you're only making 300 boards a year, you'll see stuff like that. Like, I make a logic board with $10K worth of logic chips, and I charge the customer $100K for it. The cost of a zero ohm really doesn't matter then.

Paul

This is often a problem with 'technical people' in general.

2) Because they are often immersed in that field they forget how 'most other people' wouldn't have a clue when stuff is seen in abstract.

All this was reinforced when I moved from a lifelong role in 'IT Support' to 7 years as an IT Trainer, seeing how many people hadn't really used a screwdriver or a mouse, let alone understood static electricity or binary.

Where this appears at it's worst (the assumptions re baseline understanding) is websites offering fairly technical solutions where unless you knew what the software did, you wouldn't be able to work it out from the site.

It tells you that V3.5 is now available and that it does XYZ better than the old version but with no mention of what it does, unless you knew what all the acronyms meant. DAW, OBD etc.

This is why I can rarely make any practical progress with Linux related things.

With Windows / OSX you generally download a 'Setup' file of some sort, run it and away you go. They might have included some installation instructions because there are likely to be millions of users.

For Linux the chances are there will be fewer users and (and 'what Linux' exactly) so you are often left installing / building / finding / installing dependencies yourself and the only chance of me being able to do that with any hope of success is pretty low. Unless I can find a walkthrough, written by someone similar to me who actually covers *all* the steps in a form that someone not into the details can use, even if they don't understand.

For many the point is the end, not the means, as really should be the case for an ordinary user and an OS.

Cheers, T i m

I'm now a fan of surface mount. Far easier to remove a component off without taking the lands with it. Multi-pin components come off with a hot air gun.

What I do hate is BGAs.

Hi all,

I have taken a few more photos with our camera (had to find it and charge it first as haven't used in a couple of years). They are in the same location as the others

Chip 2 - Winbond

In terms of the circuit board. As has been mentioned above, it looks like it goes pin 7 -> missing component -> pin 3 of same chip pin 7 -> hole in circuit board but nothing the other side.

There is no writing on the missing component that I can see so I assume it is a resistor? Checking the resistance with a DMM I get 1880 in both directions.

Based on this and from what I understand from the various replies, it sounds like this is to ensure the chip doesn't go into reprogram mode and therefore safe enough for me to test the board without the component. Is this correct?

Thanks again for all your help

Lee.

Or a cap and so you are measuring 'resistance' of whatever else appears across those two connections. The only true way to tell would be to disconnect that component and measure it out of circuit. Or, all other things being equal, measure across where the component was on the other board and see what that reads (may not be possible if you have removed associated components).

That may confirm it's not a diode at least, and that it's not a link. ;-)

Cheers, T i m

Aha yes good point. With the chip removed, those pads read 1.7M is one direction and 2.4M is the other. Not that helps us though?

Which is a Sanyo Serial Flash memory as suggested:

Which I haven't looked up but assume it is a volatile memory.

Which does suggest a 1.8k resistor. Possibly a bit lower in value than I would have used and higher tolerance. 1% resistors are the norm. It costs more to put them on a PCB than the actual cost of the component.

What it does do is ensure the default level is write-protect, so that if there is a corruption in the micro the memory won't be overwritten.

Writing to a Flash memory is non-trivial, with specific codes being sent to enable writing, even when the nWP pin is high.

I would say it is normally safe, but would be safer with the resistor fitted.

Well, a bit, in that 'nothing else' (other than the component or chip) is impacting your resistance reading to the 2k level. Although that may have been know by those who have checked out the chip and looked at the tracks and concluded it goes nowhere else (other than the chip) that has a 2K resistance between those points.

Another way could be to cut a track to isolate that component in a way that it would be easy to reinstate afterwards (easier than removing and replacing the component, other than you then have one to attach to the other board ;-)

Cheers, T i m

One side of the *resistor* should go to the VCC rail (as long as the logic type is still CMOS/TTL like). That's how you can get some idea of the function, is to see what the <unknown component> is joined to. Placing a cap on something whose state is steered by leakage, isn't going to be much of an improvement. Whereas a resistor is a cinch for the job (of defining state in a safe way).

Some really high speed diff interconnect are AC coupled, so as to not upset the DC bias on the receiver. And the coding of the signals can be DC balanced dynamically (the link encoder tracks DC balance so that an "ideal" signal is seen at the other end).

But around a chip like that one, there's no reason for super-high I/O speeds.

25040BW607

25FS40 is a 512KB flash-able code chip. Someone was taken advantage of by a data recovery firm here (chip swap into another board by the looks of it). You can see a resistor next to the 8 pin DIP, for pin 3 and 7.

Now that a 1.65V chip has entered the arena, I haven't a clue what they'd tie that signal pair to. Certainly you could ohm from one end of the resistor, to the VCC pin on the flash, and identify that the hot end of the resistor is tied to the most positive supply on the flash chip. I think that one end of the resistor goes to pin 8.

The chip leakage is a couple microamps (leakage goes worst case at high temp), and the pullup used needs to be "strong enough" to override that. The resistance value is immaterial as long as its low enough - only the leakage current will be flowing in the resistor. Resistance times leakage-current-value gives voltage drop across it, and the voltage cannot be so large as to not allow the input signal to be recognized as a logic level.

You might be able to use a 100K resistor, but it would be a pointless application of such, when 10K swamps it nicely.

Paul

Thanks very much Paul the first link is talking about exactly the same HDD ?? To be honest you have gone a bit more technical than I can follow sorry - not very experienced with this sort of thing . My goal for the drive is just to get it up and running long enough to get the data off it If this link is about switching the chip into program mode given it has already been programmed, am I safe enough to give it a try to see if the drive is detected / works or am I likely to cause further issue to the PCB / chip?