What is this PCB component

We've established this is a 1.65V to 2.10V device, so likely running on its recommended VCC of 1.8V

I would expect a connection of one side of the resistor to nWP pin 3, and the other side to pin 8 (Vdd)

This IC is SPI, and doesn't use any form of coding to restore DC levels.

SPI interfaces will typically run at max 50MHz bit rate. This one gives an operating frequency of 30MHz, though on reading the datasheet suggests that is a max for some operations and a max 25MHz for others.

Firefox tells me that link is a honey-trap.

You generally have little option with a typical multimeter. The best way is to increase the ohms range until the the polarity of the leads makes little or no difference to the measurement.

I feel we have already established the resistor on the populated PCB was ~1.8k

The pin seems to go elsewhere, so we don't actually know the final intention nor what the leakage might be?

The chip power supply will likely be 1.8V. I'm sure the mcu, which judging from the heatsink material will be an high speed ARM and have a

1.8V VCCIO to match.

Someone must have chosen the original 1.8k value for a good reason in the first place? Probably an aim for a 1mA current rather than just the IC leakage? 10k would normally be fine.

Lee Nowell wrote:

ogic type is still CMOS/TTL like).That's how you can get some idea of the function, is tosee what the <unknown component> is joined to. Placinga cap on something whose state is steered by leakage,isn't going to be much of an improvement. Whereas aresistor is a cinch for the job (of defining state ina safe way). +-----+----- 10K ---- 3.3V | | 3 7 /WP /HOLD I/O I/O <=== outputs could go quad SPI mode, so only want vanilla SPI operation while wired this way. Otherwise, I/O would fight with one another during quad SPI read mode.Some really high speed diff interconnect are AC coupled,so as to not upset the DC bias on the receiver. And the codingof the signals can be DC balanced dynamically (the link encodertracks DC balance so that an "ideal" signal is seen at theother end).But around a chip like that one, there's no re ason for super-highI/O speeds.25FS40625040BW607*******25FS40 is a 512KB flash-able code chip. Someone was taken advantage ofby a data recovery firm here (chip swap into another board by the looksof it). You can see a resistor next to the 8 pin DIP, for pin 3 and 7.

?id=7461&sid=0d9f536f05c832a225ae2897f8db8523&mode=viewThe 25Q40BW datasheet is here (the zero above is a Q). Note theextremely low operating VCC voltage. Which means you'd have to checkthe datasheet carefully for levels allowed on the outside.https://www.winbond.com/resource-files/w25q40bw%20revf%20101113.pdfWhen ohming the resistor, if present, you'd want to usea meter with "low power ohms" (I'm saying this in the general-practicessense - this has nothing to do with VCC value). This means, a multimeterwhich when placed on the ohms scale, applies no more than1V max to the circuit. This is intended to avoid causingsemicondu ctors to start conducting. My old analog multimeterfor example, applies up to 9V (from its 9V battery) to thecircuit. You can either consult the multimeter manualfor details/specs of the ohms function. Or, you can connecttwo harbor freight multimeters red to red, black to black,place one on ohms, one on volts, and discover the open circuitvoltage specs that way.Now that a 1.65V chip has entered the arena, I haven't a cluewhat they'd tie that signal pair to. Certainly you couldohm 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 themost positive supply on the flash chip. I think that one endof the resistor goes to pin 8. +-----+----- 10K ---- 1.65V??? ---+ | | | 3 7 8 /WP /HOLD VCCThe chip leakage is a couple microamps (leakage goes worstcase at high temp), and the pullup used needs to be"strong enough" to override that. The resistance valueis immaterial as long as its low enough - only the leakagecurrent will be flowing in the resistor. Resistancetimes leakage-current-value gives voltage drop across it,and the voltage cannot be so large as to not allow theinput signal to be recognized as a logic level.2ua * 10K = 2e-6 * 1e4 = 2e-2 = 20mV, and is only a fraction of 1650mV.You might be able to use a 100K resistor, but it would bea pointless application of such, when 10K swamps it nicely. Paul

If the resistor pulls /WP high, that's the deasserted state. That means "we don't want to write protect". The chip will be writeable.

The ATA spec allows firmware to be updated, but that's probably intended to be the Service Area firmware store. I don't know if there's any intention to overwrite the bootstrap. A Seagate updater might have the capability or need to update the bootstrap. Not every drive has a file for download from Seagate, so some never need any sort of help at all.

Some bits of disk insanity are handled by separating the controller PCB from the metal assembly, and there's also a TTL level serial interface (TX,RX,GND) for issuing commands. At least one Seagate drive that used to brick itself after a month, you could issue a few cryptic commands, then bolt down the controller board, and the drive would recover. And that was caused by a corrupt data structure that the drive would load.

If you change controller boards, some of the more modern drives require that you move the flash chip over from the original controller, as some sort of check is done. It's not sufficient to just program the SPI on the new controller with the same firmware image. Which suggests the serial number on the firmware chip is being checked too.

Other than that, you'll have to visit places like hddguru and see what the scoop is on that drive. Whether there's a known problem with a famous solution, or whether it's just a slightly higher incidence of troubles.

Paul

<snip>

If the resistor pulls /WP high, that's the deasserted state.That means "we don't want to write protect". The chip willbe writeable.The ATA spec allows firmware to be updated, but that's probablyintended to be the Service Area firmware store. I don't knowif there's any intention to overwrite the bootstrap. A Seagateupdater might have the capability or need to update the bootstrap.Not every drive has a file for download from Seagate, so somenever need any sort of help at all.Some bits of disk insanity are handled by separating thecontroller PCB from the metal assembly, and there's also a TTL levelserial interface (TX,RX,GND) for issuing commands. At leastone Seagate drive that used to brick itself after a month,you could issue a few cryptic commands, then bolt downthe controller board, and the drive would recover. And thatwas caused by a corrupt data structure that the drive wouldload.If you change controller boards, some of the more modern drivesrequire that you move the flash chip over from the originalcontroller, as some sort of check is done. It's not sufficientto just program the SPI on the new controller with thesame firmware image. Which suggests the serial number onthe firmware chip is being checked too.Other than that, you'll have to visit places like hddguru andsee what the scoop is on that drive. Whether there's a knownproblem with a famous solution, or whether it's just a slightlyhigher incidence of trouble ________

Hi Paul

Yes that is what I have done. I purchased a replacement PCB and for my drive I have to move the flash chip from the original board to the new board. Trouble is that on the new board there is that component missing. Either because it wasn't there or more likely my cack handedness trying to get the flash chip off with a soldering iron.

Question now is whether I am ok to see if the board works with out it I am less keen to remove the component from the original board in case I do more damage and then impede a professionals ability to fix it. Concerned without the component I could damage the flash chip / PCB

Thanks

Lee

If you've managed to remove an 8 pin TSSOP without damage, a resistor shouldn't be a problem.

When I do those, I use two soldering irons, a 15w and a 25w. The 15w is used as a "pusher". The 25w gives sufficient heat to do soldering operations. SMT components tend to stick to the wetted soldering iron tip, which is why I started experimenting with two soldering irons. And I prefer working with my 15w and 25w, rather than with two 25w irons.

By applying the soldering iron to both ends, you should be able to push the resistor out of the way. If the board has a solder mask, the solder shouldn't really stick.

The 15w iron can be used to push the item around, as the resistor may attempt to stick to the 25w iron.

You can use ChipQuik to reduce the melting point of a solder joint.

A bit of solder wick can be used to clean up a land for the resistor, after it's removed. Cheap PCBs, the lands are easily de-adhered by mechanical effort, so you can't be overdoing it with the cleanup effort.

Reflow soldering can be used for replacement on the other board. After the land is cleaned off, you take some fine solder with rosin core, and apply some to the land, not heating it enough to force off all the rosin. My good roll of solder is:

0.015 <=== I call this angel hair. SN63PB37 03/24/00

Then, place the resistor onto the prepped lands.

With the 15w iron and the 25w iron, you slowly rotate one CW, the other CCW, and what happens is each iron applies a downforce to the ends of the resistor while the solder gets hot enough to melt. Then pull the 25w away while still rotating, and then the 15w. The theory being, the resistor won't stick to the 15w as the other end starts to cool.

That's how I do SMT here without hot air gear. We had hot air gear at work, but I don't have equipment like that at home.

You could practice on some other electronics, to get the hang of doing these, and then you'll be ready for the real item.

My problem would be with the 8 pin TSSOP. I do stuff like that by cutting the legs with fine cutters, which is fine if a chip is to be replaced with a new one. But not so good for a chip-swap, where the chip must not be damaged.

I'd probably have to send off and get some ChipQuik and use the dental floss method. Poisoning the solder joints with a low melting point solder, makes the solder melt at a lower temperature and then other removal techniques can be tried. One other USENET poster said he was using dental floss to slide under a chip leg, once it was poisoned and then heated to the new melting point with the iron. But I've not tried that. Once the item is free of the PCB (without ripping lands off of course), then the solder wick can be used to remove the low temperature solder, in preparation for usage of some other solder material.

SMT components are the devil to handle. They bounce. If they fall on the floor, you might never find them again, because they can bounce and bounce again. At least some of them, seem to be made of a glass-like material (I had some boards come back from the factory, where all the powder blue resistors were cracked and it looked like a temperature profile issue). Other than the bouncing issue, the vast majority I worked with behaved themselves otherwise. I won't be buying any powder blue colored ones :-)

Paul