Replacing damaged chips/microcontrollers; Which components in consumer electronics need to be programmed?

[Duderino]

I'm just starting to get into electronics as a hobby and I was wondering if someone here could help me understand the process of replacing damaged microcontrollers. With the right tools and practice, I could probably replace most components, even SMDs. However, from what I understand, I cannot simply replace a faulty microcontroller with a new microcontroller.

If my understanding is correct, microcontrollers need to be programmed for the device to work correctly. Specifically, code needs to be written to the device's memory (NOR Flash?). In order to do that not only do I somehow need to obtain the code/programme to write onto the new microcontroller, but it seems that different microcontrollers require different programming equipment and/or procedures. All this would pretty much make fixing devices with dead microcontrollers near impossible for a hobbyist. It seems to me that just about the only way to do this would be to find/buy a donor device with a working microcontroller, which is likely to be expensive.

I can only guess that most professional workshops specialise in fixing only specific devices. They either (1) have access to new microcontrollers and to dedicated programmers, which are necessary to dump the code from working microcontrollers and to write it onto new microcontrollers; or (2) have a stock of donor microcontrollers. I doubt that anyone other than the manufacturer and authorised repair centres have access to the source code for the microcontroller and, possibly, also to pre-programmed microcontrollers.

Q1. Is the above correct?

Q2. What other components may require programming and cannot be replaced by a generic component when fixing consumer electronics? From what I understand microprocessors do not need to be programmed after/before replacement, even though they use microcode. Perhaps NAND Flash?

I would appreciate as much detail as possible. Thanks very much for your time and help.

 

[rjenkinsgb]

Yes, you are correct.
If you can buy replacement component parts for a specific device from the manufacturer or a service agent, then the part for that equipment would come pre-programmed for that function.

However that is getting rare, most places now only supply complete PCBs, if they supply any parts service at all.

Equipment manufacturers do not normally supply program data for their products; the unique program is often the key to making a product difficult to copy or counterfeit, so is heavily protected.

Scrap / salvage equipment is your only option.

A lot of newer gear uses fully custom ICs or programmed logic arrays as well as MCUs, and they fit the same category - only the equipment manufacturer (or a repair agent) can supply them.

Generically, fully custom devices fall in the class of ASICs - Application Specific ICs; parts unique to a product or product range, by one manufacturer.

 

[Nigel Goodwin]

Programming isn't a problem, because you can't get the code to do it.

If a microcontroller (or the EPROM/PROM for a microprocessor) fails then you simply get a ready programmed replacement from the manufacturer of the unit - if available.

However, both devices are EXTREMELY reliable, and very rarely fail - I suspect the vast majority that do get changed, aren't actually faulty, and don't cure the fault.

So pretty well it's not a cause for concern, as they don't really fail.

The main exception I ever came across was in a particular series of Sony CRT TV chassis - the same PCB was used in both 4:3 and 16:9 sets, and the LOPTX's used to fail, killing the line output transistor. If you replaced the transistor it blew again immediately, and you heard the transformer arc internally. On the 4:3 set replacing both cured the fault, but on the 16:9 set the microcontroller (a large SM Philips device) would be destroyed in probably 80% of cases.

So even those weren't 'failing' they were been destroyed by external problems.

 

[Duderino]

Programming isn't a problem, because you can't get the code to do it.

So in the extremely unlikely case that a programmable component fails (or gets destroyed) and I can't get a pre-programmed replacement from the manufacturer, there is no way to fix the device other than perhaps with donor parts, right?

Also, I noticed that in the footnote to your comment there is a link to WinPicProg, which seems to be a piece of software developed by you for programming various chips. Since you can't use it to fix/replace/programme chips in consumer electronics, as you can't get the code to do it, is the only purpose of it to programme chips for your own personal projects?

 

[Nigel Goodwin]

So in the extremely unlikely case that a programmable component fails (or gets destroyed) and I can't get a pre-programmed replacement from the manufacturer, there is no way to fix the device other than perhaps with donor parts, right?

Correct, although as I said it's VERY rare such a device would fail. But in any case, it's no different to any other component no longer been available - and if you repair electronics (as I did for 46 years) you come across that on a weekly basis.

Also, I noticed that in the footnote to your comment there is a link to WinPicProg, which seems to be a piece of software developed by you for programming various chips. Since you can't use it to fix/replace/programme chips in consumer electronics, as you can't get the code to do it, is the only purpose of it to programme chips for your own personal projects?

Exactly.

However, assuming you can get a blank device, and a programmer to upload to it, you could always try writing your own software for the faulty device - but I suspect that's likely to be extremely difficult.

 

[unclejed613]

usually, what seems to be a chip failure is one of the following: 1) POL (Point of Load) regulator failure (usually the caps go bad)... 2) solder failure, particularly under BGA chips... 3) tin whiskers shorting between pins... i've seen all 3 of these often on the DSP boards of AV receivers, and 99% of the time, they do not cause permanent damage to the chip... replacing the caps, reheating the chip, or brushing the "fuzz" off of flat-pack pins (whichever applies) fixes the problem without replacing the chip. replacing modern controllers, especially BGA chips require hot-air soldering equipment, which is rather expensive.

 

[unclejed613]

However, assuming you can get a blank device, and a programmer to upload to it, you could always try writing your own software for the faulty device - but I suspect that's likely to be extremely difficult.

a lot of Samsung smart devices use open source software, but it's really hard to navigate Samsung's various websites in search of source code... i found the source code for a blu-ray player once, because i wanted to see what was changing between versions... but each device has it's own source code, and it's difficult to track down.... since it's open source, they have to make it available... but nothing in open source licensing says it has to be easy to find....

 

[Nigel Goodwin]

a lot of Samsung smart devices use open source software, but it's really hard to navigate Samsung's various websites in search of source code... i found the source code for a blu-ray player once, because i wanted to see what was changing between versions... but each device has it's own source code, and it's difficult to track down.... since it's open source, they have to make it available... but nothing in open source licensing says it has to be easy to find....

Generally only portions of the code are open source, so it's usual that you don't have all you need to compile it - assuming you could even source (or afford) the exact correct compiler.

 

[crutschow]

However, both devices are EXTREMELY reliable, and very rarely fail

Generally true.
But I did have a motorized TV lift stand in a cabinet which was electrically programmable for the lift height.
It started lifting erratically, and a search came up with a statement that this lift had the common problem of the EEPROM losing random memory bits after a few years, causing the observed failure.

 

[Nigel Goodwin]

Generally true.
But I did have a motorized TV lift stand in a cabinet which was electrically programmable for the lift height.
It started lifting erratically, and a search came up with a statement that this lift had the common problem of the EEPROM losing random memory bits after a few years, causing the observed failure.

Presumably that would just be an external EEPROM?, used to store adjustments and settings, it's been a VERY common fault in TV's since they started using them. Not a problem though, you either replace the chip (always a good idea - and it's just a standard EEPROM) - either with a pre-programmed one (which you can easily program yourself - I had a collection of EEPROM files I had read out of different TV models, as most engineers did), or the replacement part from the manufacturer comes ready programmed. But in many cases, simply fitting a blank EEPROM triggered the micro to download the default settings to it - a rather nice sensible feature

Sharp were particularly troublesome, and in their later sets added a 'voting system', where the data was stored three times, and before use the three were compared and if one was different it was rewritten to be the same as the other two. This didn't help, and those sets were just as bad, if not worse, than the previous models.

So an entirely different fault, and cause, and nothing to do with the microcontroller or EPROM/PROM program storage.