HELP!! PCB REPAIR on early 90's boards

[robertlopez925]

i have been tasked on finding a repair solution for some older PCB... The company who designed these boards has sense closed there doors... and they are not replaceable so we need to find a repair... The PCB are motherboards for some older model trains. The PCB are through hole double sided the last service station successfully burned out the pads more specifically the little metal rings that allow you to solder for though hole.... does any one know how to best repair these PCB i will post micro pictures tomorrow... i left my camera flash card at work with the photos on it.

[Stonent]

Some stores like Fry's Electronics (in the US) sell solder pad/trace repair kits. You can also buy them online.

[PaulAm]

A repair that works fairly well in repairing plated through holes is to use a small tin plated brass eyelet.  There are repair kits available for way too much money, but it doesn't take much to set them correctly.  You just need a pointed punch to carefully upset them and then flatten them with a small punch and anvil.  Solder on both sides and you have a through-hole repair. You can find small quantities on ebay.

For traces and pads you can get copper traces and pads that glue down.  It's a limited market though, so prices tend to be pretty high.  Depending on the amount of damage, you can just use small wire to replace the lifted traces.  You can also get copper foil, cut to size and glue it down for repairing larger areas.

[WarSim]

The only stable way to repair through hole is an eyelet.  The proper eyelet punch curls the edges down for better contact.  There are common shortcuts that all have trade offs.  Using a swaging punch will cause a weaker connection.  Using the components leg for connection will cause a resistance increase and a weaker connection.  If either of these trade offs are critical will depend on the substrate material, hole size, trace purpose, vibration and temperature differential. 

Trace repair is performed many ways.  In the past preformed traces and pads were common.  Now circuit board repair skills are less common and the so are the supplies.  Now the most common method for the people still doing it is to use copper foil tape.  It requires good blade skills.  IMHO if anyone doesn't have the required skills they shouldn't be excavating a circuit board anyway.  Make sure you properly prepare the surface where a trace will be place or it will be a hassle when you position the trace. 

Luckily 1and 2 sided boards are the easiest to repair. 4 - 6 layer boards are much harder and require several move skills.  In general the layers are too thin in boards over 6 layers.  Of course there are exceptions to the rule, like semi flex able 4 layer boards and very thick 8 layer boards. 


Sent from my iPad using Tapatalk

[Joule Thief]

Is it possible you could post a picture of the damaged area of the circuit board?

Many additional possibilities might open up if we are able to evaluate the damage more closely.

Running an appropriately sized jumper wire component to component might save time and unnecessary expense.

[ignator]

Do you need to repair the thru hole? Can't you just dead bug the part i.e. glue the component to the board in the same location as the one that used the PWB mounting holes, then attach the smallest jumper wire to the nearest connection for the PWB route.
This method is used for avionics, the jumper wire is epoxy glued to the board every few centimeters (TAC-PAK). RTV is used to glue the part to the PWB.
This assumes the board material is not burned or chard from excess current of glowing component. Unless your dealing with some RF issues that can't support this rework why do you need to have the mounting plate thru holes?
The jumper wire is wrapped around the leads of the parts for connecting, then soldered.

[jamesglanville]

In the past I have soldered on both sides turned pins (like http://www.rapidonline.com/catalogueimages/Module/M082700P01WL.jpg but without plastic) through the holes, and soldered thin wire if necessary to replace any missing connections. Then you can just plug the replacement part into the holes.

[pickle9000]

I've used these, not for dips of course. Machine contact dip sockets (for IC's) can be soldered from the side or you can buy just the pins. Kynar needs to be in the kit, copper foil from a stained glass shop is handy, as is epoxy.


http://www.ebay.ca/itm/50x-Copper-Alloy-Brass-Eyelet-0-9x3mm-for-Soldering-Connection-e-/161309868798?pt=LH_DefaultDomain_0&hash=item258ed142fe

[WarSim]

This method is used for avionics, the jumper wire is epoxy glued to the board every few centimeters (TAC-PAK). RTV is used to glue the part to the PWB.

Avionics requirements in North America is 7g impact and resonated vibration immunity.  Bonding every few centimetres is not sufficient, if you are one of the avionics companies supplying to North America please stop this.  Overhauling is annoying when found on airworthiness conformance inspections. 
 


Sent from my iPad using Tapatalk

[M0BSW]

A repair that works fairly well in repairing plated through holes is to use a small tin plated brass eyelet.  There are repair kits available for way too much money, but it doesn't take much to set them correctly.  You just need a pointed punch to carefully upset them and then flatten them with a small punch and anvil.  Solder on both sides and you have a through-hole repair. You can find small quantities on ebay.

For traces and pads you can get copper traces and pads that glue down.  It's a limited market though, so prices tend to be pretty high.  Depending on the amount of damage, you can just use small wire to replace the lifted traces.  You can also get copper foil, cut to size and glue it down for repairing larger areas.

I've found the best place to buy copper foil is garden centres actually, gardeners use the foil to stop slugs climbing up and eating their plants. I've used some copper foil which is sticky on one side it work really  well.

[woodchips]

Are these multilayer boards? If so then repair isn't really possible, assuming you want it to be reliable. How do you make connection with the buried lands after the plated hole has been ripped out?

If just two sided then cut all the components off, high resolution photography and get some new replacements made. If multilayer then you do have problems, some investigations will be needed to see if the middle layers are just power and ground. If six or more layers then real problems, possibly only by striping components and then trying to separate the layers of the PCB. A hot air gun is useful here, heat the board up until it gets so hot it delaminates, but you will have to drill out all the through holes to stop them holding the layers together.

[ignator]

Quote from: WarSim on Today at 12:05:36 AM>Quote from: ignator on Yesterday at 10:02:28 PM

This method is used for avionics, the jumper wire is epoxy glued to the board every few centimeters (TAC-PAK). RTV is used to glue the part to the PWB.

Avionics requirements in North America is 7g impact and resonated vibration immunity.  Bonding every few centimetres is not sufficient, if you are one of the avionics companies supplying to North America please stop this.  Overhauling is annoying when found on airworthiness conformance inspections. 
 


Sent from my iPad using Tapatalk

This survives 20G impulse tests (per Section 7 RTCA DO-160). And RTCA DO160 Section 8 vibration to Robust random and sine curves (C₁R) (intended for fixed wing turbo prop, turbo jet environments). Understand this is 30AWG teflon insulated wire, low mass. And most jumpers are less then 1 cm in length. Larger wire, has shorter distance between glue globs.
And it's per standard rework standards that were invented long before me.

And we supply to the world.

[WarSim]

Quote from: WarSim on Today at 12:05:36 AM>Quote from: ignator on Yesterday at 10:02:28 PM

This method is used for avionics, the jumper wire is epoxy glued to the board every few centimeters (TAC-PAK). RTV is used to glue the part to the PWB.

Avionics requirements in North America is 7g impact and resonated vibration immunity.  Bonding every few centimetres is not sufficient, if you are one of the avionics companies supplying to North America please stop this.  Overhauling is annoying when found on airworthiness conformance inspections. 
 


Sent from my iPad using Tapatalk

This survives 20G impulse tests (per Section 7 RTCA DO-160). And RTCA DO160 Section 8 vibration to Robust random and sine curves (C₁R) (intended for fixed wing turbo prop, turbo jet environments). Understand this is 30AWG teflon insulated wire, low mass. And most jumpers are less then 1 cm in length. Larger wire, has shorter distance between glue globs.
And it's per standard rework standards that were invented long before me.

And we supply to the world.

Hmm apparently I missed the 30AWG stipulation in your previous post.  The wire mass is only part of the formula but I assume we both know that. 

Over 60 years of recorded maintenance history, these minimum standards have proven to be the week point.  Bridge wire problems are the most common failures.  These failures do not include environment and incident events.  The records do not specify if it was an implementation problem or not. 
What I know from my 20 years of experience is if there is a trace repair check it first because it is by far the most likely failure.  The reasons for failures vary:  Broken anchor point.  Internal conductor failure.  No isolation of the replaced trace.  Short due to jacket abrasion.  It has became so common that we reworked these repairs whenever found.  Our goal is a 10 year fix rather the recorded average of 4 to 24 months. 

Obviously you feel justified in using the current minimum standards.  I just attempted to let you know that it is insufficient, at least in the avionics area I worked in.  When a Search And Rescue Aircraft has to stop looking for someone because of an avoidable issue I find it very disappointing so is say something. 


Sent from my iPad using Tapatalk

[Fraser]

I was trained to use the PACE funnelet and eyelet method of repair. This also offers the capability to replace the PCB tracks using track repair kits. PACE produce the specialist parts to carry out this work. It may not be cheap but when the boards are high value it pays to do the job right.

 http://www.paceworldwide.com/products/support-products

http://www.paceworldwide.com/products/support-products/nonthermobond-cirkits/cirkit-basic-version-

[ignator]

Quote from: WarSim on Today at 08:51:00 AM

What I know from my 20 years of experience is if there is a trace repair check it first because it is by far the most likely failure.  The reasons for failures vary:  Broken anchor point.  Internal conductor failure.  No isolation of the replaced trace.  Short due to jacket abrasion.  It has became so common that we reworked these repairs whenever found.  Our goal is a 10 year fix rather the recorded average of 4 to 24 months. 

Sent from my iPad using Tapatalk

The bigger problem I dealt with was production not having he ability to solder jumpers where these are circuit changes/mods that were not incorporated into the PWB. Production hates them, as they can't hire the skill to implement. So test failure or worse occurs. The problem with just implementing the changes into the glass is justifying a certified (FAA TSO) product without retesting.  Some circuit changes can be justified by engineering analysis, but many require minimum qual testing (qual is generally a government systems buzz word, in my commercial world it means retesting per RTCA DO160). I don't know what your experience is, your username implies to me you work military products.
At any rate the service center NEVER has a problem with reworks. The skill level is different as that person is a real technician with FAA credentials, vs the factory assembly person just having minor training as a layperson. They (factory) used to be fairly skill persons, but for the last 10 years getting skilled operators has proven impossible. And also of late the service centers just scrap assemblies as their bonus is based on quick turn around. However that seems to be different in the many world wide and state wide service centers. It's a function of how fast they can get the factory to build a replacement CCA.

I retired last year from what used to be Collins Radio. It was fun for the first 15 years, but the last 14 were hell (dilbert arrived, retired 6 years before I wanted, best thing I ever did, but still bitter).

[WarSim]

Quote from: WarSim on Today at 08:51:00 AM

What I know from my 20 years of experience is if there is a trace repair check it first because it is by far the most likely failure.  The reasons for failures vary:  Broken anchor point.  Internal conductor failure.  No isolation of the replaced trace.  Short due to jacket abrasion.  It has became so common that we reworked these repairs whenever found.  Our goal is a 10 year fix rather the recorded average of 4 to 24 months. 

Sent from my iPad using Tapatalk

The bigger problem I dealt with was production not having he ability to solder jumpers where these are circuit changes/mods that were not incorporated into the PWB. Production hates them, as they can't hire the skill to implement. So test failure or worse occurs. The problem with just implementing the changes into the glass is justifying a certified (FAA TSO) product without retesting.  Some circuit changes can be justified by engineering analysis, but many require minimum qual testing (qual is generally a government systems buzz word, in my commercial world it means retesting per RTCA DO160). I don't know what your experience is, your username implies to me you work military products.
At any rate the service center NEVER has a problem with reworks. The skill level is different as that person is a real technician with FAA credentials, vs the factory assembly person just having minor training as a layperson. They (factory) used to be fairly skill persons, but for the last 10 years getting skilled operators has proven impossible. And also of late the service centers just scrap assemblies as their bonus is based on quick turn around. However that seems to be different in the many world wide and state wide service centers. It's a function of how fast they can get the factory to build a replacement CCA.

I retired last year from what used to be Collins Radio. It was fun for the first 15 years, but the last 14 were hell (dilbert arrived, retired 6 years before I wanted, best thing I ever did, but still bitter).

My name refers to an AI I coded in 92, to prove it could be done on commercial hardware. I was annoyed that I kept being told that I couldn't be done. 
But yes I did repair military equipment, where the term "missile grade" was the daily buzz-word. 
No idea who in the chain performed the flaky repairs, just saying it was a problem.  Once we got the equipment it never got sent back for repairs.  If an issue was raised it got the bean counters up in a tizzy.  When that happens the equipment goes into the void never to be found nor replaced. 
I remember working on a widely used VHF radio from Collins in1990-93.  I am sure there was others but that one I remember.  Good thing about that one most of the repairs where due to antenna and lightning compatibility issues.  For some reason lighting preferred scrambling HF or VHF.  I personally had about a dozen MOD related failures on those years.  In comparison pretty low. 
Yeah the bean counters argue about who pays what for months then at the eleventh hour there is a big panic for action. 


Sent from my iPad using Tapatalk

[ignator]

Quote from: WarSim on Yesterday at 09:48:10 PM

Yeah the bean counters argue about who pays what for months then at the eleventh hour there is a big panic for action. 

That was one of the reasons for retiring, they did not want to fix a very bad implementation of what is the latest modular system. They way underestimated the development cost, and did minimal corrections to problems found during qual test. This all occurred because one very incompetent manager, that had a very incompetent project engineer that thought he and the incompetent ME were 'gods gift' to packaging (they both did suffer from fundamental religious syndrome, which was my 'litmus test' to very stupid people). I was never included in any design reviews, as they never held them until the design was cast in stone. It's hell to watch.  25 years ago, if I found a problem during test, I didn't need anyones approval to fix it (I was the PE). But now it runs up the flag pole, where directors who have no technical knowledge of the issue make 'will not fix' edicts. This is not "better is the enemy of good" type problems, this is make it so it works, and ensure the flight line mechanic does not break it installing/replacing it type problems (you can't hand install a 150 pin connector that has 4 oz. insertion/extraction force per pin, and has ball grid packages that do not like the PWB being bent, it was screaming for insertion/extraction cams/levers).
It's a real problem designing avionics, you really only get one chance to do it right. Your not making a million of any part number, your lucky if you make 200 of a design. They just don't design and deliver that many biz jets of one type that does not have improvements in their avionics (and airframe) during their manufacturing run.  And as the companies development did not start without a signed contract with the airframer, your schedule (and budget) does not have time for iterating the design. Our avionics have a 30 years support life, and we copied previous designs. So once a bad implementation is injected it's kept there for ever.
In this case, this modular design is going in many aircraft, and the gov. sys. folks are using it as well.
RAZZLE-FRAZZLE

[WarSim]

I hear your pain.  6 months ago I started a 15 month leave of absence.  Best decision ever, and it's unlikely I will be returning.  Soon I will be starting a replacement job search.  On these interview I will be paying close attention to all the people in the room to decide if I want to work for them.  My last/current position my immediate boss was good but the two above him where the idiots.  Ruined the whole job.

BTW I have run into very similar connector problems.  What we ended up doing is manufacture skids for the equipment.  On replacement the skid stays with the box when removed from the rack.  Of course it didn't eliminate the problem because connector replacement was common at repair depot and vibration would eventually cause connection problems anyway. 


Sent from my iPad using Tapatalk

[SeanB]

Oh yes, mod step 3 after serial number xyz11, which is pin compatible connector wise with revisions below, but where the wiring loom was changed slightly so that plugging in an earlier version would either not work, or would cook something. Or the same part number, but a different mod state depending on the particular airframe number. Oh yes, the part number is the same for both units, you just have to look at all in the stores to see which are which mod state, or send one away to get changed to the alternate mod state.

Was going to rant a little.........

[ignator]

Sorry for theft of this thread.
Warsim:
I'm convinced every high tech company (at least in the USA, maybe world) has drank the 'koolaid'  (for those wondering, this is reference to the jonestown massacre) of dogbert consultation. No managers need to have IP knowledge of any thing for the engineers they manage.  So you end up with managers changing jobs every few years as this was good for their career.  And it was encouraged.

My last group head (direct report for me), came from the auto industry in factory operations (circuit assembly). He didn't have a clue about avionics safety.
My sad point is, I left at a low in my career, and do not have any proud thoughts that I made the product better. The company changed into something that did not allow this. And getting new talent that can do this won't happen as the company is not managed in a way to keep people for a life career. They froze the pension in 2004. And at that time, killed it for all new hires. And they advertised they could buy any talent they needed. Ya right, e.g. design a fail operational auto-pilot that you can show by analysis, has a probability of less then 10e-9 fails/hour (regulatory requirement) of a common mode design error that will cause a catastrophic event. You don't get that talent from the consumer electronics, auto industry, appliance design industries (or newbees out of school). You will only get that from a company that has designed that equipment in the past (has years of IP doing this, by evolution of many smart engineers).

Good luck finding a competent place of work.

For the connector issue, I could not get anyone to listen during the proto type build stage. Then we delivered for flight testing. The Cusomers complained, but this fell on deaf ears. There were so many software bugs that were high on the list of things needed for flight test. I tried getting the new ME interested in doing something to modify the LRMs (line replaceable modules) and cabinet lead in structure, but too new of a kid to understand that it was going to be a long term field issue. Inertia of the many fielded systems, and the lack of desire to fix. I even tried getting her interested in a special lever (a paint can opener device).  I left 15 months ago, I don't think there is a certified customer yet with a large count of airframes. First customer was just starting deliveries of a biz jet, which typically don't get high hours of flight time like a commuter or air transport customer.
I believe that I was good at designing human factored interfaces. Easy to assembly and maintain, easy to test, easy to install and remove. But you have to have a say, and that takes smart managers that know you, and have a long relationship with you, and that is what changed.

[WarSim]

"Smart managers". Isn't that an oxymoron?  .


Sent from my iPad using Tapatalk

[ignator]

Quote from: WarSim on Today at 03:10:33 PM

"Smart managers". Isn't that an oxymoron?  .


Sent from my iPad using Tapatalk

Yes we used to joke that that promoting a bad engineer to management kept them from designing junk.

[wagon]

Yes we used to joke that that promoting a bad engineer to management kept them from designing junk.

An engineer I know, who also runs a company, once opened up a competitors product and called his engineers into his office..... 'If any of you ever design something like this, your gone!'.

Wise words.