Where to begin

[viktor18]

Hello, Community!

Lately, I've been assigned an internship at a company that sells harvesters and forwarders.
For those of you who are not familiar with the machines, they are being used in the forest industry



My job at the moment is primarly to prepare test benches ( AKA taking out most of the electronics from these machines ) so we can simulate operations and test components that needs to be tested before they're being sold. (They sell spare parts).

Also, I'm supposed to repair broken electronics from these machines.

I come from a computer repairing background but not electronics. I'd like suggestions how I could learn to repair electronics effectively.
The biggest concern is that there is no available schematic for any of these products. So basically, all I could do is measure my way around. ( Keep in mind that most of the time, there is an identical device in stock that I can use as a reference point )

Currently I have an old analog 10Mhz Oscilloscope, two multimeters, LCR meter (with ESR), rework station (soldering iron, desoldering iron and hot air), infrared reworkstation and a bench power supply.

I appreciate suggestions for any device that I might need, and/or resources and tips to learning to repair broken electronic devices.

[halexa]

Hi,

Do you know what type of interface is used between ECU:s in these machines. I'm guessing CAN and/or LIN like in the automotive business. If that's the case some type of diagnostic tool for these networks might be a good ide, like  Kvaser or Vector.

And since first rule of trouble shooting is voltage check you will get far with a basic multimeter.

[viktor18]

Hi,

Do you know what type of interface is used between ECU:s in these machines. I'm guessing CAN and/or LIN like in the automotive business. If that's the case some type of diagnostic tool for these networks might be a good ide, like  Kvaser or Vector.

And since first rule of trouble shooting is voltage check you will get far with a basic multimeter.

They're using a CAN network. Hmm, think I saw something from Kvaser in a box.. Perhaps there's some diagnostic tools laying around!
But that only answers the question so much.

There's multiple devices in these machines, you have the ECU, MCM, HHM, KCM, DM, PSU and so on..

The most problematic seems to be the MCM and HHM. (Main Computer Module, Harvester Head Module respectively)
I've been able to troubleshoot two MCM's so far.. and one of them showed a broken diode shortly after one of the positive pins that gets fed from the PSU module. (My Multimeter with Diode Test/continuity on the same function) actually gave a beep when I tested the diode in circuit! (So it acted like there was no diode) Now I just need to figure out what kind of diode I need to replace it with.. as there is no schematic, I only have another MCM module as reference point with a functional diode.

Another MCM, had a burned component, which was L marked on the PCB.. I searched a long time to find a similar looking component, until I found an EMI filter with a built-in ferrite.. I don't know if that's the component and don't know exactly how to figure that out yet..

But ye.. Just need to learn to troubleshoot better.

[halexa]

Hi,


Kvaser dongles are for either CAN or LIN diagnostics, depending on the dongle.

The diode was it in series with the circuit or to GND? If it was i series it's probably a reverse voltage protection diode. If it was to GND it was probably a TVS diode. Do you have schematics for the ECU, that will help you a lot in your trouble shooting. So you don't have to reverse engineer every component.

[viktor18]

Hi,


Kvaser dongles are for either CAN or LIN diagnostics, depending on the dongle.

The diode was it in series with the circuit or to GND? If it was i series it's probably a reverse voltage protection diode. If it was to GND it was probably a TVS diode. Do you have schematics for the ECU, that will help you a lot in your trouble shooting. So you don't have to reverse engineer every component.

I don't remember fully, but the diode was connected shortly after one of the input pins for power. So possibly a reverse voltage protection diode. I have to get back at that.

And no, there is no available schematics for any of these modules. It's just not possible to get your hands on them, too exclusive.
I'm very new to electronics, so don't know a whole lot about this.

Would you recommend reverse engineer the board? After all, these are common products that has to be repaired occasionally, so it's not just a one-time product most likely. I just read about reverse engineering boards now.. Do you think it would be worth it since this is not just a one-time fix product?

How would you go about reverse engineer a board? Is it doable for a total electronic noobie by using a guide on how to do it and doing some research on the components?

[sokoloff]

Though you say that there's no available schematics, I would urge you to develop and mine the relationships and knowledge with the original engineering team(s) and especially the old-timers there.

As an intern, you have the wonderful luxury that very few people will tell you to "piss off and stop bothering me". Mine that gold.

It's good to also be thinking about how to do your task without the support of engineering, but you will make more progress, learn more, and have a better experience if you take advantage of the existing knowledgebase at the company.

[tautech]

My job at the moment is primarly to prepare test benches ( AKA taking out most of the electronics from these machines ) so we can simulate operations and test components that needs to be tested before they're being sold. (They sell spare parts).

Also, I'm supposed to repair broken electronics from these machines.

I come from a computer repairing background but not electronics. I'd like suggestions how I could learn to repair electronics effectively.
The biggest concern is that there is no available schematic for any of these products.

Key to understanding will be the functional block diagrams of how these work.
There'll be a large # of sensing circuits and control circuits but each will be reasonably simple in design with interfaces between each of data transmission, processing, control (operator input), interlocks and electro-hydraulic implementation.
Break it down into all the little blocks and the electronics of it all won't be awfully hard.

Pieces of plant like this will likely suffer from sensor, servo and connectivity failures (lost or shorts), the connectivity being the most risky if it introduces another voltage or short into a circuit.

Most field repair will likely be tracing connectivity issues and some knowledge of values of what should be on each connection with the machine in a "known" state should be very useful to know if it can be fixed in the field or if it's a back to the shop job.

You're likely to hit the wall with usefulness of your old scope especially if you want to build a library of screenshots of what waveform "should" be each connection so a DSO that can save screenshots to USB and can do basic serial decoding should be considered in the next short while. CAN can be done with many entry level DSO's.

[julian1]

My job at the moment is primarly to prepare test benches ( AKA taking out most of the electronics from these machines ) so we can simulate operations and test components that needs to be tested before they're being sold. (They sell spare parts).

The biggest concern is that there is no available schematic for any of these products.

Is this actually a commercially viable strategy? I would ask management to try and get a NDA with the equipment manufacturers for better information about their products and maybe training.

[viktor18]

Thanks for your replies. IMPORTANT INFO BELOW

There's a possibility that I've mislead you with all this information.
The company is NOT the manufacturer, they just primarly sell used machines and spare parts. As far as schematic goes, the only thing we can get our hands on is those coming with the machines, which is just a regular schematic how everything electronic is connected, but not how components inside the actual modules are wired up.
So unfortunately, nobody at this company knows about the inside and the information is just simply not available for the public.

Simply, I need to repair modules sometimes, and therefore need to know how to troubleshoot a device. Perhaps reverse engineering the board is needed? As I never will be able to put my hands on schematics of the actual PCB inside modules.

[Brumby]

You haven't mislead anyone who has read and understood what you have written.  I know I understood your situation.

You have a challenge none of us would envy.

Reverse engineering is something that requires a keen eye, attention to detail and patience that grows exponentially with the size and complexity of the board.  It is difficult enough with traces that disappear under components, but if you have more than 2 layers in the PCB, you will have fun.

Certainly, I would be getting copies of any documentation you come across.  Those block interconnect diagrams are something - and they could give you somewhere to start.

[viktor18]

You haven't mislead anyone who has read and understood what you have written.  I know I understood your situation.

You have a challenge none of us would envy.

Reverse engineering is something that requires a keen eye, attention to detail and patience that grows exponentially with the size and complexity of the board.  It is difficult enough with traces that disappear under components, but if you have more than 2 layers in the PCB, you will have fun.

Certainly, I would be getting copies of any documentation you come across.  Those block interconnect diagrams are something - and they could give you somewhere to start.

Block interconnect diagrams are not a problem. We have hundreds of books on those.

I was looking through the web and found this company http://www.protelpcbs.com/ .. and apparently they could reverse engineer the board for me and provide a full view of the board and it's components. Perhaps that's a viable solution? Most modules can probably just be fixed with a multimeter, looking for burned components and shorts.. but other than that, it might be useful to have a company reverse engineer the module for me? After all, if I can't fix a module because a lack of schematic, it a $3500 product thrown in the trash.

[Brumby]

Sounds like an idea.  Wonder what the cost would be?

As for chasing down burned components - you need to be careful.  Without a diagram or a good idea of the function, you could end up replacing a component with something inappropriate.  This could be a rectifier diode when it should be a TVS - or a resistor ... what value?

[viktor18]

Sounds like an idea.  Wonder what the cost would be?

As for chasing down burned components - you need to be careful.  Without a diagram or a good idea of the function, you could end up replacing a component with something inappropriate.  This could be a rectifier diode when it should be a TVS - or a resistor ... what value?

Yes, that's the biggest concern right now. I have been able to identify a few broken components, but a lot of research goes into finding replacements (as many are not marked)..

And for the price, I'm supposed to email them pictures and information on Monday, and they'll forward it to the engineering team and give me a price range. It totally depends on the price and if my boss is willing to spend the money.

Alternatively it's possible to use one board as a scrap yard for components, so I can take components that are working on one board and put them on another board to make at least one functional board. Hopefully the price will be somehow acceptable. It's a one-time cost that can save us a lot of modules and time.

[TerraHertz]

Are you able to post photos of some typical PCBs here? Or is that proprietary information?

Also, do you have a copy of Photoshop? I ask that because using photoshop image layers as aids while reverse engineering a PCB to the circuit diagram, saves a lot of time and confusion.

First take the best contrast possible, reasonably square hi-res photos of both sides of the board.
Then import front and back into Photoshop as separate layers, make the top one 50% transparent.
Use the 'distort' tool to get both images nice and square and lined up. Blink the top one on and off, to see any misalignment of holes.
The you add some more layers, to draw tracks and stuff on as you work out the connectivity. Some by eye, hidden ones by multimeter. Also a layer to allocate part numbers, where the PCB silkscreen doesn't. And one more to dot parts 'done' as you add them to a tentative schematic.
Here's an example:  http://everist.org/NobLog/20160331_lightning_luck.htm#tla614

Note that was only a 2-layer PCB. Boards with 4 layers where the inner 2 are just power planes are not much worse. Boards with multiple internal trace layers - personally I don't even try.

[james_s]

The way I learned to repair stuff was hanging out with someone who worked at a repair shop, he taught me the basics and then I took it from there. If there's a good tech in the company spend as much time shadowing them as possible.

[nctnico]

The way I learned to repair stuff was hanging out with someone who worked at a repair shop, he taught me the basics and then I took it from there. If there's a good tech in the company spend as much time shadowing them as possible.

And if such a person isn't around then try to retrace the schematic from the board and try to determine what the signals look like going in & out the board (the big picture). What also helps is to look at datasheets for special chips. Usually there is a typical application circuit which shows how the circuitry around the chip is supposed to behave. Also don't be afraid to apply power to a board using a DC power supply (with a low current limit) and just measure around to see if you can spot something. I repaired lots of CRT monitors this way and even ones other people gave up on trying to fix them.

[tggzzz]

I come from a computer repairing background but not electronics. I'd like suggestions how I could learn to repair electronics effectively.

To what granularity of the unit-under-test (UUT): box, board, component?

Your first task is to diagnose which UUT is faulty. That requires you can stimulate it, observe its response, compare the actual response to the expected response.

If you can't find a way of stimulating a UUT, all you can do is look at the power supply.
If you don't know the expected response (comparing with a known-good UUT is acceptable), then you cannot determine if it is faulty.
If you can't determine which UUT is faulty, you can't repair it - but randomly swapping units might be acceptable.

[onesixright]

You have a challenge none of us would envy.

Depends. In these cases, if (I should say when) you do fix it, the reward is bigger (ego)  Well that's what I like to believe. As long you love what you do!

[MatteoX]

There is a book about reverse engineering that was discussed in the forum: Ng "The Art of PCB Reverse Engineering: Unravelling the Beauty of the Original Design".
I haven't seen it but forum member Fraiser wrote very nice review on Amazon.

You might find some helpful information about reverse engineering in the following thread: https://www.eevblog.com/forum/eda/pcb-reverse-engineering/

[xrunner]

Are you able to post photos of some typical PCBs here? Or is that proprietary information?

...

Boards with multiple internal trace layers - personally I don't even try.

I'd like to see some pics too. Much good has already been said, but the thing is repairing a board that has an obvious issue - like a burned resistor (not to mention what caused it but ...) is one thing. If you get a board that looks fine to the eye, well you have another problem. Unless you can find the problematic component by testing with a DMM or other type of device, you have a real hard problem. Because if you can't stimulate the board with a good input and follow a signal path, you won't have a clue what's supposed to be going on without a schematic, especially if you haven't repaired electronics before. And if it has multiple layers - ouch. And if it has an embedded controller with firmware and so on ... double ouch.

I mean I don't want to get you down, but it's possible that you have gotten yourself into a very difficult situation. But I'll be quiet now because if I've brought you down somebody else will pick you back up. But post some pics and we can at least say if it's easy or hard on a scale from 1 to 10.

Good Luck! 

[james_s]

There is a big advantage though if you are working on a specific type of equipment. It likely won't take long before you encounter a bunch of similar boards with the same problem and once you figure it out once you can repair most of them pretty easily. I remember when I was repairing a lot of CRT monitors they all started to look the same. I started seeing one of the same 5 or 6 faults in almost all of them.

[viktor18]

Thanks for your replies.

In instructions manuals, there's some information on how to troubleshoot most modules by measuring output and reading error messages. Generally, if there's a problem with an input or output channel, it's going to send an error message that there's a trouble at that specific pin.. so that leaves you with some clues where the problem might be on the board, at least you could trace your way around there and make sure nothing is wrong.

I'm including some pictures on a board here.



I understand that most of you really try to remind me what a hard task I've taken.

But understandable, some boards are just simply not going to be repairable without a schematic atm.. but as long as I can repair those with obvious faults there's not a problem! They understand my limitations.

[Harb]

Good luck with it.

I take on a lot of projects that are of similar nature.....Equipment that is no longer supported, or was never meant to be repaired by a 3rd party and some even potted or de-numbered (or both) boards........it can be a very frustrating exercise thats for sure.

All you can do is look at the interconnection diagrams you have, observe the machines operation and get a clear understanding of that much and then break the larger circuit into smaller more understandable blocks.......which are usually very simple sense and react systems that will go on to send a warning or trigger another reaction.
Do the interconnection diagrams at least show pin designations ? ........ie what each pin represents ? If so, its shouldn't be impossible to see a probable circuit in there.
The boards you have pictured look like there is lots of similar circuits in there to perform similar functions.
if you are selling lots of similar units, or repairing lots of the same "boxes" it may be worth every cent to get them reverse engineered........

You will also need to make a test Jig if this is going to me regular........an easy way to trick the box into thinking its in the actual machine.
If you are in a machinery wreckers situation, you should be able to "find" enough parts machines to build a test jig out of parted out sensors etc, or build them up yourself......... most will be simple make/break switches , resistance value or pulse counting setups etc

Probably the first thing I would do is search every component and its purpose.......its going to be near impossible if the parts are proprietary  numbers...........if they wont give you a circuit, you can forget getting info on what part is what.

again good luck, I know and understand your frustration......and can only imagine the job you have ahead not coming from electronics and being given this task........always, and I mean AlWAYS make sure they understand the issues you face, so they know they may end up with a "Deader" unit than they arrived with.

[viktor18]

It's actually happening. We have disassembled a functional machine and I have taken everything electronic, including all modules, sensors etc.. And I'm installing most of it on a test bench so we can simulate.

It'll be helpful for the sales team as they can test units before selling them, and for me so I can troubleshoot more easily.

This is the type of Schematics I can get my hands on, this is for the module I posted pictures on above.

[Harb]

Thats a good starting point......at least you have that much.