Show us your ugly repair

[elex_enthusiast]

Fixing an intermittent shutdown problem of an old Toshiba laptop of a friend which turned out to be a defective proadlizer cap (well it's been a common issue with proadlizer caps failing on old Toshiba laptops).

[Whales]

vze1lryy: How did you conclude a whole rail was shorted on the mobo?  How did you even trace that rail to every part that needed it?  It sounds like you had access to the laptop's schematics.

Everything in this thread looks beautiful to me.  Too many people assume electronics is all about final, clean PCBs and mass-production techniques; because that's all they have seen in their lives.  Hell this thread is teaching me lot.

I was trying to fix an encoder on my HP 54600A scope about a week ago.  I pulled it apart, cleaned it, inspected the wipes and regreased it to no effect.  The day after I that I discovered that I had pulled and cleaned the completely wrong encoder.  Note to self: don't start repairs late at night.

Unfortunately the encoders HP used are not made to be serviced.  Instead of having nice bendy metal tabs like some pots do, they have these cast triangular spikes holding on their back plates:



For the first encoder I painfully bent the back plate until it would lift off.  Putting it back on was difficult too and it did not look very nice after.

For the second encoder I gently snapped the triangular stakes off the casting with light squeezes of my pliers.  This left the backplate intact, but no method to secure it back on.

I initially hot-glued it back on, but this was not satisfactory.  Especially since this backplate would fall onto the PSU inside the scope if it popped off during use.  The solution?  Press on the control until it fails, try again with 2x the hot glue, rinse and repeat until we are several times above the failure point for severe control pressure:



I need to track some replacement encoders down.  Even if I use a better adhesive/clip, my grease was too thin, and the controls are noticeably different.

Oh, and the problem with the dodgy encoder was not too hard to spot.  Here are the wipers attached to the shaft you rotate:



Keep these images and stories coming!

[vze1lryy]

vze1lryy: How did you conclude a whole rail was shorted on the mobo?  How did you even trace that rail to every part that needed it?  It sounds like you had access to the laptop's schematics.

Maybe! Boardview and schematic definitely help. I removed every component from the line and it still had a dead short to ground.

[T3sl4co1l]

Even without schematics, it's not hard to tell, from the components and layout, what's connected to the main power rail.

IIRC, my laptop has five supplies connected to a common 14-19VDC rail.  One's for battery management, the others are VCORE, 1.8V, 3.3V and 5V (or presumably something like that; I didn't measure).  YMMV, but it's a pretty good approach, and likely many laptops follow a similar pattern.

All those converters have pairs of transistors (they're all synchronous buck type, in my case), and you can easily measure the high side source being connected to that common rail.  Or, in the above repair case, you'd have the peculiar problem that PMOS sources are grounded...

Tim

[Stefan Payne]

So I needed a PCIe x1 video card. And I wasn't willing to pay a premium for one - or wait for shipping, for that matter.


Some dremelling later - and discovering that I cut a couple somewhat important traces - and voila! Like a bought one.

Well, yeah, that's one way to do it...
The other way to do it is to open the PCIe Slot, that's what I did on my HP/compaq thingy with Q963 Board (forgot what it was) - and plugged a HIS Radeon HD5750 into it...

[ovnr]

So I needed a PCIe x1 video card. And I wasn't willing to pay a premium for one - or wait for shipping, for that matter.


Some dremelling later - and discovering that I cut a couple somewhat important traces - and voila! Like a bought one.

Well, yeah, that's one way to do it...
The other way to do it is to open the PCIe Slot, that's what I did on my HP/compaq thingy with Q963 Board (forgot what it was) - and plugged a HIS Radeon HD5750 into it...

As I said in a later post, I was more attached to the motherboard than the video card. Also, cutting near the somewhat fragile socket pins did not fill me with a great deal of enthusiasm.

[v8dave]

After my client stuffed 110Vac into the board instead of 24Vdc, this is a result of the repair (one regulator still to be installed). It seems that the failure stopped at the regulators which were all damaged and the 3.3V and 5.0V devices on the board still work. There used to be a poly fuse where the copper is missing so I fitted a new Littelfuse to the side of this.

[samnmax]

Fixing an intermittent shutdown problem of an old Toshiba laptop of a friend which turned out to be a defective proadlizer cap (well it's been a common issue with proadlizer caps failing on old Toshiba laptops).

Your repair is not ugly at all, compared to mine 
This was an Acer 18" laptop.

[Electronoooob]

More pictures

that has some great character

[Electronoooob]

Here's pic one...

deargod

[SeanB]

Temporary adjustment so it would not leak, and today I replaced the diaphragm, along with replacing almost all the M5x12 CSK stainless steel screws that hold the assorted parts together, I cammed a few getting them out to do the quick adjustment, and got new ones along with a new 3mm hex key to make sure I did not damage the new ones.

Still had one that decided to play coy, so had to do a little "persuasion" using a hammer and a cold chisel to get it moving out enough to undo fully.

New diaphragm to place in storage as spare arriving Monday, this one lasted pretty well, replaced in 2013. The rubber ones used before never lasted that long, they typically only last a year before tearing, which can be a little messy. Polypropylene fabric ones are a lot better, and are as a bonus a lot cheaper. I can even get them in ballistic Nylon, or for really rough applications in PTFE. Looked at the ( hand written) label, and it was made in 2001, I guess they do not move very fast.

[TerraHertz]

Temporary adjustment so it would not leak, and today I replaced the diaphragm, along with ...

How the heck does that work? And what sort of flowing material is it for?

Here's something kind of ugly, not really a repair, and not electronics.
A friend brought a found large tower PC case round on Friday night. He needed it to fit a new over the top motherboard with some huge CPU cooling system. (Which I haven't yet seen.) Anyway, he wanted the case front to look 'industrial' as opposed to standard office crap.

So here it is, industrialized. Yes, that's the new front. Red button is the power momentary action, black button is reset (and is a military high-force type.) Two LEDs to go in later. He's not sure if he'll anodize it, or just leave it plain aluminium.
He didn't want any drive slots or USB ports. The machine will be all solid state drive(s), with USB hub nearby.

[SeanB]

Temporary adjustment so it would not leak, and today I replaced the diaphragm, along with ...

How the heck does that work? And what sort of flowing material is it for?

Here's something kind of ugly, not really a repair, and not electronics.
A friend brought a found large tower PC case round on Friday night. He needed it to fit a new over the top motherboard with some huge CPU cooling system. (Which I haven't yet seen.) Anyway, he wanted the case front to look 'industrial' as opposed to standard office crap.

So here it is, industrialized. Yes, that's the new front. Red button is the power momentary action, black button is reset (and is a military high-force type.) Two LEDs to go in later. He's not sure if he'll anodize it, or just leave it plain aluminium.
He didn't want any drive slots or USB ports. The machine will be all solid state drive(s), with USB hub nearby.

Anodise it to be almost white, and dye it black or blue, depending on the surrounding equipment.

The valve is a Mucon AD6.

http://muconusa.com/products/ad-series-iris-valve/

Used for controlling the flow of powder by me, though they have been used for many things, including controlling the flow of fish in a pond.

[TerraHertz]

Anodise it to be almost white, and dye it black or blue, depending on the surrounding equipment.

But that would make it look almost attractive, and that's the opposite of what he wants. I had a ride in his latest car that evening too. It's a sleeper. Looks like some crap dirty beat up small old Jap car. But it's terrifying. The guy who modded it lost his license almost immediately, but my friend is more circumspect. Same effect intended with the PC.

The valve is a Mucon AD6.
http://muconusa.com/products/ad-series-iris-valve/
Used for controlling the flow of powder by me, though they have been used for many things, including controlling the flow of fish in a pond.

That's interesting. So there's nothing exposed to the flow medium but the stretchy plastic liner?
One of the things my vacuum project needs is a large diameter diaphragm valve, for throttling the turbo or diffusion pump, to allow controlling very low gas pressure and flow-through in the working chamber. Not yet sure how to do it; thanks for the link.

[jonovid]

My ugly hack is on a poor little Chinese mini DV Camera.  so I can add an interface too it with Arduino.

[w2aew]

My "ugliest" repair, at least in terms of the "custom" work that I had to do, was replacing some IF filters in a Kenwood VHF/UHF transceiver where I had to install extra series caps to remove the DC bias from the filter.  Here's a video of the repair (real ugliness starts at about 10:54 in the video):

[Cyberdragon]

I have lots but most I'd have to take stuff apart to show them. Heres one I can show now though.

Notice something?

[bitseeker]

Notice something?

The photo is a bit out of focus, but apart from the "X" on the caps, I see a cylindrical thing floating in midair in the upper right quadrant of the pic. What is that? A flying bodge?

[Cyberdragon]

Notice something?

The photo is a bit out of focus, but apart from the "X" on the caps, I see a cylindrical thing floating in midair in the upper right quadrant of the pic. What is that? A flying bodge?

Sorry for blurr, low res android cam pic. Those "X"s are actually "+". There are 4 floating trimmers, two at the bottom and two in the middle. The blue thing at the bottom, a small brass one on the opposite side of the transformer, the one you see in the middle, and a small brass one behind it. The trimmers kept failing open or intermittent/noisey and I don't have the right values in that style.

[lowimpedance]

An Exact function generator !?.

[Cyberdragon]

Yep, Exact model 126. Still has a bit of noise on the low frequencies, maybe more bad trimmers.

[bitseeker]

There are 4 floating trimmers, two at the bottom and two in the middle.

Wow, four flying trimmers. Fun and ugly repair!

[Cyberdragon]

You haven't seen the floating HV caps yet. Those are hidden away in some tube stuff where no one can see them. I do put electrical tape on the leads if the wires aren't tight enough to keep them from flopping around but still...

Wait...found one...

The cap in the middle bounces when poked (several hundred volts when on). It's nippon chemicon though.

[TerraHertz]

Here's another. Does this win "Tiniest Ugly Repair" ? (Ha, what am I saying? Very likely someone can beat this.)

I'd been given three Dell S500wi short-throw Projectors, all with some problems. The first was an easy non-electronic fix, this second one had a dud main fan, so would rapidly shut down with an overheat warning.
Unfortunately the fan has a custom frame, so couldn't easily be swapped with any other. Either fix it, or trash the whole projector. The second option was quite tempting, since I'd like to take the optics path apart from curiosity. Also it seems stupid to try and fix a fan, but I thought I'd have a quick try.



So much for 'quick', it turns out that to remove the fan requires almost completely disassembling the machine, just to get at the 4th fan mount screw. Brilliant design.

The fan is a standard 3-pin type; gnd, +12V and tacho output for fan fail sense.
It was trying to start, with small jerks of the rotor. Spins freely by hand, so... something electronic dead in the fan drive pcb. Like most fans the rotor comes out easily - peel off the label and remove the circlip on the shaft.
But then there's a problem - the poles and drive PCB assembly is solidly glued to the central plastic mount.

Needed some way to apply force, without damaging the fairly delicate pole pieces. I found a bit of brass tube that was a near fit to the poles, cut a slot in it, and clamped it tight on the poles with a pipe clamp. With that as a handle, NOW it comes off!




The board is pretty standard - a control IC (On semi, LB11970RV) and a typical Hall Effect sensor (marked "7", very helpful.)



Found the data for the drive IC, gave the board power and started probing around the IC. Everything seems OK, except that the two signals from the Hall sensor are nowhere near the same voltage. One is near ground, and noisy. I thought they were supposed to be the same with no B field?

I pulled apart another old fan picked from the junk box at random. With that one's board powered, the hall effect sensor outputs are the same. Putting a magnet near it produces nearly a volt differential, polarity reversing if the magnet is flipped. Yeah, I thought so. And generally they are all passive, so fingers crossed, I can substitute this one in. And with luck it will even be the right polarity, since it's just a square of semiconductor with contacts on the four sides brought out to pins, and flipping it around shouldn't make any difference.



Ha ha, except I think I'd better preserve the + and - power pins polarity. And of course Murphy ensured given that constraint, this one is a mirror image of the dead one. OK, so I will mount it upside down.



So there it is, soldered dead-bug style.




And it works!

[bitseeker]

That is a tiny repair, TerraHertz. As I was reading, I thought you were going to do some kind of ugly bodge on the projector's main PCB that drives the fan or something. Swapping in the PCB from another random fan was cool.

Although the upside-down dead bug and solder blob do qualify as an ugly repair, I think that overall it's a bit too good for this thread. Next time, duct tape an old AC fan to the exterior of the case!