Messing around with notebook power supplies.

[Meaker]

As you likely well know notebooks tend to use custom power plugs or a range designed to handle different loads. There are of course margins to exploit in these connectors so long as higher quality are used and this lead me to create a modded 330W dell brick for my MSI, it has a very similar output voltage (19.5V vs 19V) and is capable of supplying almost twice the stock 180W brick

Our vic... *cough* I mean brick of choice today is the big daddy at the bottom of the pile.



Manufactured by delta for use with the M18X this bad boy packs 330W at 19.5V.

Flipping it over you will see 4 removable feet covering the screws.



It uses torx anti tamper screws so you need one of these:



T10 security torx screwdriver (hollow 6 point star).

Let's plug it in and test the voltage shall we? NO WAIT A SECOND ACTUALLY.



DO NOT SHORT THE CENTRE PIN WITH THE INSIDE OF THE BARREL

It will kill the brick. So measure it but be careful!

Now let's take that cover off after removing the 4 screws:



You have two choices, unsolder the 4 tabs holding the shield in place, gently pry off the glue holding it all on or....

cheat:



Nice chunky leads for obvious reasons:



So you are going to have to unsolder each of the 3 connectors, ground (outside of the barrel) is next to the small data pin connector. Solder in your harvested cable from another brick into the holes.

Pop the case back on and measure the no load voltage (should be around 19.5V)

[DRT]

If you do this to use with a Dell, remember to transplant the 1-wire bus ID chip as well (easily found - it connects between the 3rd wire and 0V on the lead). On mine it was a TO92 style device. On my Inspiron the BIOS  throttled the CPU speed and prevented battery charging until I re-fitted the ID chip.

[c4757p]

On my Inspiron the BIOS  throttled the CPU speed and prevented battery charging until I re-fitted the ID chip.

Are you fucking kidding me??

I swear, the reasons not to buy a Dell just keep piling up.

[Jon Chandler]

Many different laptops are using this scheme in the chargers.  Even another model of charger from the same company with adequate capacity may not work!

[c4757p]

That's just shady. I know the ostensible reason for it - there are a lot of dangerous, crap chargers out there - but that shouldn't be the manufacturer's responsibility! Hyundai isn't responsible if I decide to put something nasty and explosive in my gas tank, why should Dell be responsible if I decide to use a Wun Hung Lo PSU? But of course, one look at the price tag on a new brand-name charger reveals the true reason....

Thankfully I haven't run into this problem yet - and as a quick test, I put my newer Asus on a bench PSU and it seems to function properly.

Assholes.

[bilko]

You don't think the ID chip might have been a 1-wire temperature sensor that reduces the load on the PSU if it detects overtemp ? 

[Monkeh]

create a modded 330W dell brick for my MSI

For what purpose, exactly?

[c4757p]

You don't think the ID chip might have been a 1-wire temperature sensor that reduces the load on the PSU if it detects overtemp ? 

Could be, but I don't see much point in not just using a thermistor. You can get fancy with digital temperature monitoring, signaling to custom software in the BIOS to throttle things and let it cool. Or, you can shut down the PSU when it gets too hot with a simple thermistor and comparator, the laptop runs momentarily off the battery, and throttling happens anyway due to power management software detecting the loss of power.

[StubbornGreek]

create a modded 330W dell brick for my MSI

For what purpose, exactly?

This is my question also. Unless you're planning on 'software overclocking' (BIOS is locked down in laptops), your equipment will only pull what current it needs from the psu and nothing more. As far as software overclocking goes - never a good idea.

[baljemmett]

That's just shady. I know the ostensible reason for it - there are a lot of dangerous, crap chargers out there - but that shouldn't be the manufacturer's responsibility! Hyundai isn't responsible if I decide to put something nasty and explosive in my gas tank, why should Dell be responsible if I decide to use a Wun Hung Lo PSU?

There's another potential reason for it - telling the laptop what the power rating of the PSU is.  I remember my old (c. 2002) Inspiron had a little three-pin connector that was identical to the connector on beefier models that drew more juice; using a lower-power PSU would cause the bigger machine to either run or charge but not both simultaneously.  I believe running my own machine in a dock with the smaller PSU would do the same thing.

(And I'm jumping on the "but why?" bandwagon, too...)

[Meaker]

This is useful when you have a MSI GT60 along with a 4900MQ (maximum oc 4.2ghz on all cores) and GTX780M with full voltage, temperature and power target control.

You skyrocket up to 240W consumption pretty easily at that point (and that's taking power supply losses out).

[Monkeh]

Oh, right. Portable heaters. I thought you had an actual laptop (if you can call an MSI that).

[Meaker]

Yes you can with a 5 hour (of practical use) battery life, the 15.6" model weighs around 3kg so it's fairly easy to move around.

I'll also add that only Dell in the high performance notebook segment uses PSU ident on that third pin, Asus, MSI, Clevo/sager all will just use the two connections.

[Stonent]

That's just shady. I know the ostensible reason for it - there are a lot of dangerous, crap chargers out there - but that shouldn't be the manufacturer's responsibility! Hyundai isn't responsible if I decide to put something nasty and explosive in my gas tank, why should Dell be responsible if I decide to use a Wun Hung Lo PSU? But of course, one look at the price tag on a new brand-name charger reveals the true reason....

Thankfully I haven't run into this problem yet - and as a quick test, I put my newer Asus on a bench PSU and it seems to function properly.

Assholes.

It's not shady, there's a few reasons for it.

Dell has kept the current Latitude D/E series power plug for laptops and all in one PCs for almost 10 years now?

Prior to that, the Latitude C series used the same plug from the late 90s to early 2000s but it was proprietary in shape.

Prior to that, the Latitude XP, XPi, XPi CD all shared the same plug. In fact you could use the XP series plug on a C series laptop if you trimmed the corner off with a knife so it would fit.

However at the time of the XP series, there were 3 or 4 types of plugs in use at Dell, the XP style, LM style , ultralight style and Inspiron style.  This was because Dell was mostly developing the XP series internally, but the LM and Inspirons were clones of Winbook and Quanta made laptops. In fact Winbook used to advertise their laptops were 100% parts interchangeable (drives, cords, upgrades) with Dell Inspiron 5000 series.  The Inspiron 3000 and 3200 were twins of two HP Omnibook models.

Their Ultralight laptops like the L400 (I think that's the model) was a twin of a Gateway 2100 laptop IIRC. The L400 was listed as fixed memory but you could use the Gateway 2100 repair manual to disassemble the Dell and access the supposed non-upgradable memory module and upgrade it to 512MB.

The round D/E Series plug in current was available at 65W, 90W, 130W and now apparently 330W.

The plugs are backwards compatible which is nice because you can order new plugs and use on a 10 year range of computers.

Prior to the Core and Core 2 laptops the 65W worked on everything. When the Core and Core 2 laptops came out, you could still use the 65W cord to charge the laptop but if you docked it, you needed the 90W cord.

When power requirements increased more (Latitude E Series for example), you could use the 65W cord to keep the laptop on, 90W cord to charge it, and 130W on the dock.

If Dell doesn't recognize the cord, it defaults to the 65W behavior. It is somewhat handy because my wife can use the cord from her work HP laptop which has the same plug on our home Dell laptop when she's sitting on the couch. It pop's up a warning but you just click OK and keep using it.

The ID chip tells the laptop how much current it is allowed to pull. You also don't want the laptop trying to pull 130W off of a cord that wasn't designed for it.

The technology must not be too proprietary because I've not found a third party cord that listed it was Dell compatible that didn't work with a Dell when the appropriate power rating was matched.

Many of the universal cords have separate plugs for Dell and HP even though they have the same physical connection, so I suppose they embed the ID chip inside the plug.

[c4757p]

Overengineering and still borderline anticompetitive. Most laptop chargers deliver about 19V, so how about this: 19V = 65W, 19.5V = 90W, 20V = 130W, 20.5V = 330W. Surely it would be easy to throw a craptastic ADC or comparator somewhere on the motherboard that can resolve to the half volt, surely a charger can easily regulate to the half volt, and anyone can implement that without proprietary ID crap. You account for voltage lost in the cable by measuring first before switching it in as an active power source.

[Meaker]

There is naturally a drop in voltage with load though which could cause issues.

[alm]

Why would you go through the trouble and complexity of analog voltage sensing, which tightens up the Vout spec for all your supplies, if it can be easily solved with a very cheap digital EEPROM? Backwards and forwards compatibility is also much easier: the lack of EEPROM can easily be detected, and you can easily add more power levels in the future. It even leaves open the option of a power supply that changes its max output power based on eg. temperature or power source (like those dual car/mains chargers). It also makes it less likely for poorly designed third-party universal power supplies (that might deliver 20 V unloaded despite being only 65 W) to mess things up.

[c4757p]

There is naturally a drop in voltage with load though which could cause issues.

I addressed this, though...

alm, you have a point, but can you really trust the bastards not to try messing around with encryption so nobody else can make the chargers? I'd rather a bit of design inconvenience to having to pay god knows how much for an authentic charger five years down the line when they've decided it's "obsolete" and the only people making replacements are the ones already unscrupulous enough to break the crypto.

It even leaves open the option of a power supply that changes its max output power based on eg. temperature or power source (like those dual car/mains chargers).

If you're already going to the trouble of building a SMPS that can operate on either 12V or mains, surely it can't be that much harder to make the voltage vary by 0.5V depending on the power source? And what's wrong with an overtemperature cutout?

[amyk]

IBM/Lenovo use an analog sense in their recent models - a resistor to ground is connected to the middle pin, and its value identifies the wattage. In the machine it's connected to an ADC input on the EC.

Some RE on that here: http://forum.thinkpads.com/viewtopic.php?p=682644

[Stonent]

I wouldn't say it is over-engineered or anti-competitive. The Latitude line has guaranteed availability of parts and models. I remember years ago it was you could order a model for exactly 18 months from the day it hit the market. That helps businesses standardize on a platform and minimizes the cost of creating custom software images for multiple models.

By keeping the power supplies standard, they at least guarantee that part is available.  The Inspiron and XPS series laptops just benefit from sharing the cord.

I can't see what's anti-competitive about it.

I find it more anti-competitive when they change the plug each model. I can go to any electronics store, spend $40 to $50 and buy a universal power supply that will work perfect on any Dell as long as the wattage is correct.

The fact that both Dell and HP have standardized on the same plug helps with future repairs. That means the socket is more widely available since broken power sockets are very common on laptops. It has gotten significantly better now that they've gone to a larger plug with integrated pin since you can replace the power supply if the pin snaps instead of replacing the socket or board.

[rsjsouza]

That means the socket is more widely available (...)

Although I think that nothing beats a connector with two wires (V_CC and GND), the number of after market adapters do not seem to pose a problem. I agree with Stonent the popularity of these brands end up making the difference in availability in the long run - and there is always eBay to get this old stuff...

I find it more anti-competitive when they change the plug each model.

The example I recall the most is Palm: every spin of their PDAs had a different connector. Until other companies had time to pick up the pace, they were the sole supplier of accessories.

BTW, an interesting article about the identification pin. It is a Dallas 1-wire EEPROM.