U6 component

[Markos Vakondios]

Hi,

Could anyone identify component U6 on IO board? I want to replace it on by blown bb (from overvoltage).

Thanks

[Andrew Scheller]

> Could anyone identify component U6 on IO board? I want to replace it on by
> blown bb (from overvoltage).

I've got no idea what the component is, but on one of my bifferboards
it looks like it says "G18I" and on another BB it looks like it says
"618c".

I guess it *might* be the 1v8 LDO for the J1 connector? (which would
imply that U4 is the 3v3 LDO)
http://www.bifferos.co.uk/pinouts/

Lurch

[Rilhas]

> Could anyone identify component U6 on IO board?

I'm absolutely 100% sure that I think it is a switched mode voltage
regulator. :-)

About 1 or 2 years ago me and Biff were brainstorming some ideas for
the Bifferboard and I remember analyzing it in detail up to the point
of determining the operating frequency and such. So I thought I would
have an easy answer for you, but it turns out that I checked a new
board and the power supply design seems to have changed. The one I
studied had 2 inductors and generated 3.3V and 1.8V, but the more
recent version has only one inductor (so it generates only one
voltage) and is acompanied by what looks like a linear voltage
regulator. To me that makes sense, since the old dual-voltage
regulator was very noisy, especially regarding the 1.8V to the CPU.

But there are some good news. The simple answer I could give you - had
the power supply been the same I studied - would be that you were all
out of luck because I searched for that chip and it seemed to be a
custom design which was not documented and was not available for sale.
The good news is that - since the power supply has changed - hope is
not all lost yet.

First you need to determine the pinout of the Bifferboard's regulator.
I tried to determine it but not all is clear: pin 4 is connected to
+5V from the power supply, and pin 2 is connected to ground, so these
power the chip. Pin 3 is connected to inductor L4, so that might be
the switching pin (but for that to be true the other side of L4 should
be connected to +5V, and it doesn't seem to be, but it may come from
somewhere else in order to reduce noise). The regulation feedback is
usually acomplished with a resistor divider, which seems to be the
pair R2 and R5. If so, then the feedback pin is pin 5. One pin is
left, pin 1, which is also connected to +5V. Usually switchers have an
enable pin pulled high to enable the device, so this might be it.

Some more good news then: having an enable pin permanently connected
to +5V suggests this is not a custom switcher, so maybe it can be
found on sale (or adequate replacement). And since pins 1 and 4 are
both connected to +5V you can replace this switcher with, at least, 2
alternate pinouts (swapping enable and chip power).

All this is very pretty, but it should be confirmed before you start
buying replacement chips. Pin 3 should be switching between 0V and the
output voltage (3.3V?) at a frequency of 50KHz to 500KHz. Pin 5 should
be at some rather stable low voltage level, like 1V or 1.2V, but this
depends a lot on the chip. This is an important parameter, you must
find a replacement chip that has this same feedback voltage. The
output voltage should also be confirmed, and I think it will endup on
one of the capacitors C18 or C19, but there should be a diode going
into it (although my multimeter does not indicate D1 being connected
to either, nor to L4).

Maybe someone can confirm all this for you using a working I/O board?

When you get confirmation you can check Mouser, Farnell, DigiKey, or
any other supplier with a good search engine and filter, for switched
mode voltage regulators (A.K.A DC/DC regulators) in a SOT-23-5 package
taking an input of 5V and outputting 3.3V (or the confirmed output
voltage). This should not yield too many results, but you may need to
search more than one supplier. Then you can try to filter out chips
that do not contain 618 in their part number (although that is a long
shot). By now the list of matching chips should not be that long, so
you may try to open a few datasheets to see if the pinouts and
feedback and output voltages match. If so voilá!

Now I can imagine all this yielding good results, but I don't see it
being worthwhile. Not only all the deductions I wrote are very
fragile, it is likely that searching through suppliers and checking
datasheets will be time consuming. And if you find a solution you must
order it, wait a while, and pay a lot to get the piece delivered
(expect to pay £10 or £20 for a £0.5 chip). And when you finally get
it you have to desolder the burned out chip (easy enough) and then
solder in the new one (with all good joints without shorting it nor
overheating it). And if you succeed at that you should pray that all
the assumptions and confirmations are correct. And if they are correct
then pray that no other component was blown together with U6, because
the +5V goes to multiple places on the board (like the USB port, for
example). And then pray asking that your prayers be answered, because
a wrong chip can blow the whole thing and another malfunctioning
component can blow out your freshly-soldered brand-new chip.

... maybe you should consider buying 1 or 2 working Bifferboards that
I have on sale...? :-)