Dell The AC power adapter type cannot be determined and shield cablefor id chip??

  • Thread starter Thread starter marslee
  • Start date Start date
M

marslee

I recently got a free dell d600 laptop. The laptop is fine but it has
the notorious "The AC power adapter type cannot be determined"
problem.

I borrow a new ac adapter and the problem is solved.

I found out from a website that the dell ac adapter has some design
flaws.

http://www.laptop-junction.com/toast/content/dell-ac-power-adapter-id-chip-died

I don't want to keep buying AC adpter because of a bad id chips.
Is there any way to modify it other than using shield cable or moving
the power supply id chip to laptop?
 
I recently got a free dell d600 laptop. The laptop is fine but it has
the notorious "The AC power adapter type cannot be determined"
problem.

I borrow a new ac adapter and the problem is solved.

I found out from a website that the dell ac adapter has some design
flaws.

http://www.laptop-junction.com/toast/content/dell-ac-power-adapter-id-chip-died

I don't want to keep buying AC adpter because of a bad id chips.
Is there any way to modify it other than using shield cable or moving
the power supply id chip to laptop?
Click to expand...

So what is the design flaw ?

The one wire network is intended to support networks with a lot
of wire in them. With slew rate control, it is possible to make
decent looking waveforms, without upsetting the FCC. The devices
also have a high ESD rating. The communication rate is 16Kbit/sec
or so, sort of like the rates you might see on an RS-232 cable.
In the laptop adapter case, the network is relatively small,
compared to the device capabilities.

http://pdfserv.maxim-ic.com/en/an/AN148.pdf

The reliability report for the DS2501, shows some failures in high
voltage testing. The "HBM" entries would stand for "human body
model", which is a particular combination of RC network, charged
to the indicated voltage. So it is meeting "8KV HBM". (An Intel
USB port is 6KV rated, for comparison.) I don't know what the
"IEC 61000-4-2 CONTACT" test above it is all about, so cannot
comment on whether failures of that test can be ignored
as the report seems to be doing. It is almost as if they expect
the device to fail that particular test.

http://www.maxim-ic.com/reliability/dallas/DS2501.pdf

I couldn't get a datasheet for the DS2501 from Maxim, so had
to get one from datasheetarchive. It is a DS2502, but bulk
programmed. The bus voltage, according to this datasheet, is
2.8V to as high as 6.0V. So if the laptop-junction guy is
seeing 16 volts delivered, that is the amount of voltage
used for programming. What the effect would be on the
device, I haven't a clue (because programming is
probably another hard to find datasheet). If I was the laptop-junction
guy, I'd be comparing voltage level on the 1-wire, to
the range allowed by the datasheet. I'd feel happier if
I could get this sheet from Maxim.

http://www.datasheetarchive.com/search.php?q=DS2501&sType=part

The 1-wire family may exist with different bus voltages. Or
it could be, that the higher bus voltage is intended for
larger networks or something. In any case, I'd want to be
absolutely sure first, whether the application (Dell laptop),
is treating that poor chip properly. If the laptop has a problem,
and is abusing the ID chip, then it might kill them one
after another.

As for "moving it inside", that sounds like a good idea. Would
the chip be drawing power from the battery at all times ?
You'd almost need a schematic for the laptop, to get some
idea of how they're probing the thing, and what power source
is supposed to be used. Or what power source could be used
in a retrofit.

Any amount of external protection could be added to the 1-wire.
Like external ESD snubbing networks, a zener to limit bus
voltage, or whatever. It might not be just an unarmored
bus. Lots of fiddly little details...

As for the comment about "2 meter long UNSHIELDED wire",
unshielded wire is fine, with attention to details. The
bus is unterminated, but has slew rate limiting. So you
would not expect active signal radiation. It would be
possible for conducted emissions to travel down the wire,
but you could remove those with a ferrite bead, or a
filter cap etc. The filter cap could even be part
of the slew rate control network. For the 1-wire to work,
it needs a ground reference, and the ground wire traveling
in the cable next to the "antenna", reduces the loop size.
We don't know if the wires are twisted inside the cable
or not. Again, lots of little details.

So right now, I have only one question.

"What is 16.5V doing traveling to a 2.8 to 6V chip?"

HTH,
Paul
 
I recently got a free dell d600 laptop. The laptop is fine but it has
the notorious "The AC power adapter type cannot be determined"
problem.

I borrow a new ac adapter and the problem is solved.

I found out from a website that the dell ac adapter has some design
flaws.

http://www.laptop-junction.com/toast/content/dell-ac-power-adapter-id-chip-died

I don't want to keep buying AC adpter because of a bad id chips.
Is there any way to modify it other than using shield cable or moving
the power supply id chip to laptop?
Click to expand...

http://www.buy.com/retail/product.asp?sku=209135292&listingid=29860431&dcaid=17902
 
I found the schematics.
http://img16.imageshack.us/img16/2240/psidl77.jpg
http://img16.imageshack.us/img16/4444/iotopsidl.jpg

The DS2501 data pin is connected to power jack center pin and then -
The AC adapter jack center pin has no voltage when not connected to
laptop.
When connected, center pin has 3.23V.

According to a laptop repair website, some D600 motherboard has burnt
L77 and D108. In this case, I/O must be replaced.
Click to expand...

L77 BLM11B102S is a ferrite inductor (possibly murata.com). It appears
as a resistance at high frequencies, and is used to attenuate high
frequency noise. It consists of a "piece of wire wound one or more
times, through a piece of ferrite". When used in a data circuit, they
can distort the data signal, but in this case, 16KHz will likely be
quite transparent to its effects. It is intended to prevent RFI from
escaping the laptop, by traveling down the adapter cable. (So electrical
noise from the laptop, won't spoil broadcast TV reception.)

D108 is a dual diode configured in "signal clamp" mode. Dual diodes
are available in various configurations, and this head to tail
style of dual diode, allows the two ends to be connected to "rails".
In this case, one rail is +3VALW and the other is ground. If
the signal on PS_ID goes more positive, by 0.7V, then +3VALW,
then a large current can flow through pin 1 of the dual diode,
burning it. If PS_ID were to go more than -0.7V (below ground
voltage), the lower diode would conduct. My guess is, the upper
half of the dual diode, is what is getting burned.

The intention of D108, is to protect LPC47N254 from damage.
Sometimes a signal will have a little overshoot or undershoot,
and rather than have a protection diode inside LPC47N254
go into conduction, they use a beefier three legged
D108 type device outside, to take the strain. The
overshoot and undershoot time period and current
magnitude, are normally handled by devices like D108,
without it even getting warm. So if all D108 is doing,
is handling logic signals, normally there would not be
a problem.

But D108 is not designed, to stop say, a power rail from
the adapter (16.5V leaking from the main power lead). If
the 16.5V main DC from the adapter makes contact with
D108, without a current limiting resistor in place, then
D108 will be a "smoking hole in the ground". The level
of current may even be high enough, to melt the tiny
wire (depending on how it is formed) inside L77. L77
will have a low resistance to DC flow, but perhaps not
when a power supply is placed right across it :-(

I would guess that pin 2 is somehow coming in contact with
pin 1 on the power connector. I would examine the connector
style, to see if it has a mechanism to prevent rotation and
incidental contact. Or, look to see if the pins stick out
far enough, to allow the wrong contacts to touch.

As well, notice the interface actually consists of four signals.
Safety_Ground (connector metal shell?), DCIN-, DCIN+, PSID.
The laptop eventually connects DCIN- to ground. That means
DCIN- is zero volts with respect to metal on the chassis.
And DCIN+ would be +16.5V or whatever, when the adapter
is providing power to the laptop.

If I was picking the connector, I would make sure that
PSID had the shortest pin in the connector, such that
PSID made contact the last, when the connectors mate.
I would try to ensure that ground was properly established,
before PSID could get in trouble. It is possible there
are other fault modes - for example, you could still
damage LPC47N254, but without leaving burn marks on
L77 or D108. If the engineer in charge of this
project hadn't included PSID in the design, the
design would be closer to being bulletproof. Adding
a data signal puts the design in a whole different space
(more engineers should be called in for consultation).

We had some issues like this at work. A lot of our products
were designed for hot insertion (safely adding an electronics module
to a running machine), and at least one module had the pins
making contact in the wrong order. This caused a quite negative
voltage to appear on a logic signal, blowing the chip all
to hell. So when there are power signals and data signals on
a connector, you need to involve a mechanical engineer, to
do an analysis as to how many different ways a customer
can jam the product in the hole the wrong way, and so on.
A good design, guides the connector mating, such that only
the correct pins can mate, and mate in an order that will
not allow abnormal voltages to appear on any logic signals.

Paul
 
Back
Top