Repair of $150k Phantom Flex 4k camera - advanced soldering

[rx8pilot]

My side hustle is diagnosing and repairing high-end cinema cameras. The Phantom Flex 4k is a high-speed 4k camera that is the standard for super-slow-motion shots in movies and television.

Most of the issues on this camera are power-related and I have learned to diagnose and repair most of the component-level failures. This particular unit sitting in front of me has a number of cracked solder joints on one of the high-density PCB-PCB connectors. The straight connectors were a straightforward hand soldering fix under a microscope. The right-angle connectors are a different story. They are on a very high layer count PCB backplane with quite a few connectors on both sides. There is an inside row of pins that are well outside the range of any soldering iron. I have a good quality convection batch oven, but even that is a risk of damaging the other connectors or allowing them to shift out of position.

Really not sure how to safely attack this one. I have been considering adding vapor phase soldering but that does not necessarily offer a magic fix. Under a microscope, I can see 5+ of the blind pins separated from the pads when looking through the tiny gap between PCB and the bottom of the connector.

Anyone care to chime in with suggestions on this one?

[dr.diesel]

Is it possible to wire around the backplane on those pins, bodge wire fix?

[james_s]

Can you get in there with a hot air pencil? With the board preheated that might do the trick.

[rx8pilot]

Is it possible to wire around the backplane on those pins, bodge wire fix?

I don't see an option for wires but the good news is that the pins on the edge appear to be primary incoming power and they are all parallel routed on the PCB. Shorts/bridges are not an issue in this area.

Can you get in there with a hot air pencil? With the board preheated that might do the trick.

I do not have a hot air pencil, just one of the larger hot air machines. This sounds like an interesting idea, however. I could do a basic pre-heat in the big convection oven, take it out on use the larger hot air system to keep it pre-heated from below while attacking the actual area of interest with the hot air pencil. It could work.

Another element I am considering is machining some connector guards out of teflon to reduce the risk of damaging the good connectors. That may allow me to pump more heat into the PCB with less risk of melting the adjacent connectors. Hmmmmmm......

An interesting point is that this PCB is not available for replacement if it is damaged. Nuts.

[AngusBeef]

https://www.allaboutcircuits.com/projects/tools-techniques-hot-air-soldering-surface-mount-components/

Sent from my SM-T820 using Tapatalk

[ddavidebor]

What is the problem? the highlighted pins seem perfectly fixable by hand

[Psi]

Yeah,  any pin/pad is fixable with a soldering iron under the microscope if you have the right tip.

[Mr. Scram]

You buy a $150K camera and they won't sell you parts? Or is the thing EOL?

[rx8pilot]

What is the problem? the highlighted pins seem perfectly fixable by hand

The exposed pins are no problem with a soldering iron, it is the hidden second row of pins underneath the connector.

Take a look at the image attatched. The hidden row of contacts is what is throwing the big party.

SAMTEC:
https://www.samtec.com/products/lsem-140-01-f-dh-a-n-tr

[rx8pilot]

You buy a $150K camera and they won't sell you parts? Or is the thing EOL?

No, they will not sell the parts to an end-user. They will only offer direct service where they replace sub-assemblies at eye-watering prices in addition to taking a very long time. That is largely why this has become a common camera to land on my bench. Most of the issues/failures are about $30 for parts and 10-20 hours of labor. It is a fun side hustle to do nights and weekends.

[Psi]

I find it hard to believe the inner row of connections will be damaged given that it is anchored from each side.
One side from the TH mounting pins and the other side with the outer row of pins.

But if one of them is cracked, then yeah, you could try carefully reflowing the entire pcb.

If that doesn't work and you're sure that's where the issue is you'll need to remove and resolder the connector. (preferably a new connector)

That's totally doable but very easy to stuff something up while doing it.
Getting connector off with reflow oven and hot air shouldn't be too bad.
Cleaning the pads is easy.
But getting an even coat of solder paste back on each pad correctly will be tricky.

[rx8pilot]

I find it hard to believe the inner row of connections will be damaged given that it is anchored from each side.
One side from the TH mounting pins and the other side with the outer row of pins

I can see the inner row pins freely moving when I look through the tiny opening between the PCB and the connector body. Sadly.....the anchors are not soldered at all putting all the stresses on the signal pins. Kind of ridiculous since this connector is subjected to a lot of pressures from the mating part.

But getting even coat of solder paste back on each pad correctly will be tricky.

I think I will make a stencil for this assuming I can get to that point.

[ddavidebor]

What is the problem? the highlighted pins seem perfectly fixable by hand

The exposed pins are no problem with a soldering iron, it is the hidden second row of pins underneath the connector.

Take a look at the image attatched. The hidden row of contacts is what is throwing the big party.

SAMTEC:
https://www.samtec.com/products/lsem-140-01-f-dh-a-n-tr

I would guess your best line of action is dumping a litre of no-clean flux on that thing and have it reflow in a professional reflow oven in some assembly house.
With all of their work being eaten by Shenzhen fabs they're never picky.

[Psi]

ah, i see.
My guess is that there was very little solder holding them on to begin with.

.

[ddavidebor]

But even then tough cracked joints may not reflow properly if the surface tension toward the pin aided by the oxidation makes them not touch...

[ddavidebor]

So maybe it's better to remove them and reflow a new one IF you can remove it safely, which is quite hard. With than many layers you need heathing on both sides to avoid damage.

[Psi]

Yeah, i think the only way to fix it properly is to get that connector off the pcb and ideally replace with a new one, but you might be able to get away with reusing it if it looks ok. But i strongly recommend a new one.

If the TH legs are not soldered i would be tempted to bend them inwards so they don't touch the sides of the hole.
That way the connector can simply be lifted off the board once all the solder melts.

I wouldn't use hot air for this (unless you have a preheater). It's too easy to rip up a track because you're not heating it evenly. Tracks are much easier to rip off when there has been force applied to a connector that has cracked solder.

The way i would do it with the gear i have is to put the board in a reflow oven and get the entire PCB up to solder temp.
Then i would open the door, carefully pull out the tray and very very carefully remove the connector with two tweezers. One holding the pcb and one lifting the connector.

If using a oven you will have to sort out how to hold the pcb so you dont melt any connectors on the under side.
What i mean is, if the oven wire rack is touching a connector on the under side the plastic will be super soft and it will bend the connector. So you have to make up some spacers so its resting on the PCB FR4.

There is also a risk the PCB may warp/bend when heated.

[rx8pilot]

I am drawing up a custom aluminum tray that will guard the connectors. I can do a controlled pre-heat in the convection oven and then use two hot-air systems (and a lot of flux) to bring it to the melting temp.

Still thinking that through.

As for a brand new connector, there is a huge lead time so I am highly motivated to protect and use the original if at all possible.

Typet purly on my fone.

EDIT:
I have a Haas CNC machine which makes custom fixturing and trays a simple affair.

[rx8pilot]

Here are the masking fixtures I am machining to protect the connectors. 6061-T6.

The plan is to put the PCB in the fixtures and bring it up to about 210C in the batch oven and then pull it out and use two air systems - one on the bottom and one on the top to get to flow temp. Going to attempt laser cutting a stencil with my CO2 laser (not sure how well that will go). At least the final solder will be done with leaded solder so the temps will be lower.

[Zucca]

I would buy a bunch of those connectors and make some PCBs to play with them.
Once you are confident with your process then move on with the $150K golden one.

BTW having spares connectors could save you in the future.

Are those not available anywhere?

[Psi]

Going to attempt laser cutting a stencil with my CO2 laser (not sure how well that will go). At least the final solder will be done with leaded solder so the temps will be lower.

You can get a laser cut stencil for like $20 from any PCB place online,  why try to make one yourself?

You can cut it to size with a pair of scissors easy enough.

[Rerouter]

You have 2 beautiful screw studs to mount a thermal gaurd plate, I would fit a sheet of metal mounted on them, and kapton tape between the board and guard to prevent any cross flow of heat,

Preheat the board to something like 130C, even better if you have something from underneath to keep it toasty,

Then just go overkill on the existing connectors, melt them off if you have to,

Once removed, clean back the board with wick, tin any pads, lay down some flux, get the area hot enough that the solder begins to flow on the pads, switch to heating with the hot air from the underside, and place down a new replacement socket, nudge it lightly until you see it drop on both sides, give it a few more laps with hot air underside with some more nudges, then you should be set.

For those bigger center connectors, as long as you keep to one of the edge connectors at a time they should be fairly steady as the other half of the connector should be holding it in place.

Edit, OK now I see the 2 giant connectors on the other side, this makes things harder. when it comes to direct underside heating... your going to have to go a bit warmer than I would normally on the board, and may be worthwhile preheating the new connector before placing it, this way it doesn't immediately freeze the joints.

[amyk]

You buy a $150K camera and they won't sell you parts? Or is the thing EOL?

No, they will not sell the parts to an end-user. They will only offer direct service where they replace sub-assemblies at eye-watering prices in addition to taking a very long time. That is largely why this has become a common camera to land on my bench. Most of the issues/failures are about $30 for parts and 10-20 hours of labor. It is a fun side hustle to do nights and weekends.

If enough of these same boards come to you with the same defect, maybe it's worth reverse-engineering and making your own... from what I see, it only contains a few passives and connectors. Perhaps you could even improve on the existing design so that those joints don't crack as easily...

[Renate]

It may not be applicable to this case (if you can't get a new connector),
but I always try to destroy the old part to protect the PCB even in trivial cases (like an 8 pin DIP).
I'd take a Dremel cutoff disk, chop up the connector and remove the pins one by one.

[rx8pilot]

It may not be applicable to this case (if you can't get a new connector),
but I always try to destroy the old part to protect the PCB even in trivial cases (like an 8 pin DIP).
I'd take a Dremel cutoff disk, chop up the connector and remove the pins one by one.

It may become the only option. The lead time on these is fairly long but I have a good relationship with Samtec - they may have some sample qty's available if I am lucky.

Typet purly on my fone.