justin1985 wrote on 2024-01-27, 11:57:

I got this little mATX VIA Apollo Pro board with soldered on C3 processor recently. I remember having one just like it in a seco […]
Show full quote

I got this little mATX VIA Apollo Pro board with soldered on C3 processor recently. I remember having one just like it in a secondary system in the early 2000s, but it seems pretty unique now.

It does boot, but I only just noticed what looks like 2 SMD components next to the front USB header seem to have blown up! One component looks badly damaged and is hanging off, and I think another to it's right is totally missing!

IMG_20240127_114722.jpg

If I'm reading the silk screen correctly, these are C73 and L50.

Any tips on diagnosing what these were, and what they might do?

Instinct says something to do with the USB header?

L is an inductor.
C is a cap.

No idea of values. Probably 0.1µH for L and maybe 10-100µF (my guess is 15) for the cap. Absolutely don't take my word for it.

Nexxen wrote on 2024-01-27, 12:40:

L is an inductor.
C is a cap.

No idea of values. Probably 0.1µH for L and maybe 10-100µF (my guess is 15) for the cap. Absolutely don't take my word for it.

Thanks for this!

I've tried to clean up the area with IPA and taken a closer look. I'm now thinking there might not be a missing component, just the burned one is a wider format than the others?

If that's true, perhaps I was reading the silk screen wrong? Maybe TC24 is the through hole capacitor, C73 is intact, and it is just L53 that's burned?

I guess the other question is whether this is the kind of thing that would have been caused by a short or something with the front USB connection?

only one way to be sure, measure if its in line with usb power

This brings me back to the scary days of early USB.. looks like that inductor worked as a nice fuse.

Looks like the one that blew on one of my boards.
That one I couldn't figure out the value, sorry.

justin1985 wrote on 2024-01-28, 14:47:

IMG_20240128_143900.jpg

If that's true, perhaps I was reading the silk screen wrong? Maybe TC24 is the through hole capacitor, C73 is intact, and it is just L53 that's burned?

I guess the other question is whether this is the kind of thing that would have been caused by a short or something with the front USB connection?

Your new interpretation of the silk screen is correct. L53 is an interference suppression inductor on the +5V line of the USB connector. It is connected in series with the component labelled "FS5" (which should be a fuse. If that component were a fuse, it should have blown before L53 went up in smoke, but the component installed at FS5 suspiciously looks like a zero-ohm resistor (that is a wire jumper in a resistor package for easy placement with automated assembly equipment). So something grossly overloaded or shorted the +5V supply of the front USB connector. This might be due to connecting something with a shorted USB cable, or connecting some stupid gadget like a 10W cup warmer, or this might be the result of using an incompatible front USB connector. There are incompatible pinouts, and mixing them up might short out the supply rails.

mkarcher wrote on 2024-01-28, 23:06:

Your new interpretation of the silk screen is correct. L53 is an interference suppression inductor on the +5V line of the USB connector. It is connected in series with the component labelled "FS5" (which should be a fuse. If that component were a fuse, it should have blown before L53 went up in smoke, but the component installed at FS5 suspiciously looks like a zero-ohm resistor (that is a wire jumper in a resistor package for easy placement with automated assembly equipment). So something grossly overloaded or shorted the +5V supply of the front USB connector. This might be due to connecting something with a shorted USB cable, or connecting some stupid gadget like a 10W cup warmer, or this might be the result of using an incompatible front USB connector. There are incompatible pinouts, and mixing them up might short out the supply rails.

Thanks so much!

Checking the manual on RetroWeb, the pinout is quite different to normal USB2 front connectors - pins 1 and 10 are VCC rather than 1 and 2, pins 2 and 9 are ground, etc., when 9 is usually the key pin. Pin 9 is the one that on the board had been bent right over and actually snapped off when I tried to straighten it - so almost certainly someone had tried to fit a standard USB-2 connector!

A bit of Googling suggests a USB socket should have a ~1.5A polyfuse - does that sound about right to replace the 0 resistor?

First thing will be to see if I can actually get the burned inductor off cleanly enough to leave enough pad to solder a replacement to! Getting soldering iron and tweezers in between the PCI slots looks incredibly tricky ... (I don't have a hot air station, just a temp controlled iron with changeable tips)

If I do manage to get it off, this seems like a fair guess of a value? If inductors are there to filter noise, instinct says precise values shouldn't be totally critical?

Nexxen wrote on 2024-01-27, 12:40:

Probably 0.1µH for L and maybe 10-100µF (my guess is 15) for the cap. Absolutely don't take my word for it.

justin1985 wrote on 2024-01-29, 08:05:

If I do manage to get it off, this seems like a fair guess of a value? If inductors are there to filter noise, instinct says precise values shouldn't be totally critical?

Nexxen wrote on 2024-01-27, 12:40:

Probably 0.1µH for L and maybe 10-100µF (my guess is 15) for the cap. Absolutely don't take my word for it.

I looked at another thing that blew and value was 1µH, not 0.1.
Having limited knowledge it's probably good enough for +5V USB. As explained, it serves the purpose of filtering.

mkarcher wrote on 2024-01-28, 23:06:

... It is connected in series with the component labelled "FS5" (which should be a fuse. If that component were a fuse, it should have blown before L53 went up in smoke, but the component installed at FS5 suspiciously looks like a zero-ohm resistor (that is a wire jumper in a resistor package for easy placement with automated assembly equipment). So something grossly overloaded or shorted the +5V supply of the front USB connector...

Unless, with wrong cabling connected, the first connected device was a powered USB hub =)

justin1985 wrote on 2024-01-29, 08:05:

A bit of Googling suggests a USB socket should have a ~1.5A polyfuse - does that sound about right to replace the 0 resistor?

In pre-USB3 days, the specified maximum current for USB ports was 500mA, so a 1.5A fuse for a single USB port sounds excessive. Fusing 4 ports in parallel with a 2A to 2.5A fuse was very widespread. This is OK for an average current of 500mA, but also allows spin-up of 2.5" hard drives which is typically around 1A - if you don't connect 4 drives at the same moment. This arrangement also made the 10W (2A) cup warmers work in many scenarios if the other ports were not consuming a significant amount of power. In your case, the fuse will handle two ports in parallel, so going lower than 2A makes sense. To still cater for some inrush current, 1.5A seems quite appropriate.

justin1985 wrote on 2024-01-29, 08:05:

Nexxen wrote on 2024-01-27, 12:40:

Probably 0.1µH for L and maybe 10-100µF (my guess is 15) for the cap. Absolutely don't take my word for it.

If I do manage to get it off, this seems like a fair guess of a value? If inductors are there to filter noise, instinct says precise values shouldn't be totally critical?

I don't know about the inductor. To prevent something similar happening again, do not just choose the inductor by inductance, but also by rated current. If you use a 1.5A polyfuse, use an inductor specified for 2A (or higher) continous current to make sure the fuse acts before the inductor overheats. Nexxen already corrected the guess to 1µH instead of 0.1µH. 1µH inductors between 2A and 5A current seem very common, just use anything that you can easily obtain and approximately matches the footprint.

Careful work with the finest tip on my soldering iron and I managed to get the old inductor off. It came off in two parts, but the pads seem to have survived well enough - the bottom one perhaps a little loose but it is substantial enough to have survived, I think. It took quite a bit of cleaning with a small fibreglass pencil and IPA on a tiny cotton swab to see the contacts properly.

The explosion looks like it burned a hole right into the substrate of the board! Is that anything to be concerned about?

There is continuity from the surviving pads through to the 0 resistor at the top, and at the bottom to pins 1 and 10 on the header (VCC). So seems consistent with being blown by having the wrong USB header connected.

In terms of the fuse that should be where the 0 resistor is, are we talking about the kind of "polyfuse" that usually looks like a kind of yellow/orange disc type thing?

justin1985 wrote on 2024-01-29, 18:17:

The explosion looks like it burned a hole right into the substrate of the board! Is that anything to be concerned about?

There most likely never was an "explosion", but "burn" seems to be the correct word. I guess the coil was literally on fire, and that burned the top layer of the board. If you were dealing with significant voltages (line voltage would count), you should be afraid of leakage current through residual carbon (where the hydrogen of carbo-hydrates burned of, and carbon residues remained in the board), but at 5V this shouldn't be an issue. The flame retardands mixed in the epoxy of the PCB prevented further damage.

justin1985 wrote on 2024-01-29, 18:17:

In terms of the fuse that should be where the 0 resistor is, are we talking about the kind of "polyfuse" that usually looks like a kind of yellow/orange disc type thing?

Yes, exactly. I suppose they originally intended a 2A one-way fuse (link as reference, not as endorsement) on that position, but the disc of polyfuse will do, too. Mainboards designed for polyfuses typically employ SMD polyfuses, though. fitting the through-hole disc-type polyfuse is likely easier than adapting the board for SMD polyfuses.

It was a struggle but I managed to find some hopefully suitable components from RS, and soldered them between the PCI slots with minimal damage ...

The inductor is 1uH and rated at 1.2A which seemed to be the best I could find in a 1206 package (and in less than full reel quantity). The resettable fuse has a trip value of 1A, which is perhaps a bit low, but at least should protect the inductor ... I also replaced the snapped header pin.

The board still boots after all that abuse, which is encouraging. I can't actually test it until I put together a breakout board to translate the weird pin-out to standard USB. Shouldn't be too difficult with some new header pins and an offcut of Vero board, I guess.

While I'm at it though, I'm tempted to see if I can add the missing ISA slot at the bottom of the board. It does look like the through hole capacitor is the only other missing component. It does trace through to one of the slot pins. Is there any way to work out what value it might have been, or at least guesstimate something that might work?

It is a power related capacitor and will have same value as the others like it, most probably 16V 10µF. It does look like you can just solder in the ISA connector and it'll work.

Eventually got back to this, and I can confirm just soldering on the ISA slot did work!

I also added a scavenged 10uf capacitor (from an old PCI modem) in the empty hole next to the slot. The only testing I've done so far is initialising a Yamaha YMF719 soundcard using Unisound in DOS, but that worked fine 😀

Cleaning out the solder from the holes first and cleaning up the flux etc did take off quite a lot of the brown solder mask. I guess it's a good idea to paint some more on? Is it the UV curable stuff I should use?

Really pleased to have this odd little board all fixed up and working! I don't really have an immediate use for it, but I think it makes an ideal spare DOS focused system. It's like a VIA EPIA ITX system, but with floppy controller, game port on the onboard sound, and now an ISA slot for more exotic sound cards 😀

It's currently in a much too modern super compact mATX case that only has slim drive bays, but I'll keep my eyes open for something like a compact OEM mATX case with proper drive bays.

justin1985 wrote on 2024-04-21, 22:36:

Cleaning out the solder from the holes first and cleaning up the flux etc did take off quite a lot of the brown solder mask. I guess it's a good idea to paint some more on? Is it the UV curable stuff I should use?

If you don't need the property of the solder mask to be heat resistant and repel solder, but just the isolation and mechanical protection of the traces below it, clear nail polish is a sufficient substitute.