Moving laptop power supply inside the case: Considerations

[jester]

Hello people, first thread of mine here, glad to be around you.

I have this little project in the works and I think there are a few aspects of it that are not my expertise so to speak, so I'd appreciate any relevant input.

It involves building a small form factor computer, but with a catch, as you will see. To make things more visual, therefore clearer, allow me to show you the main components.

At the core there's the motherboard:

This is a tiny (170mm x 170mm) and also very thin mini-ITX form factor motherboard, for older generation Intel CPUs.

It differs from usual motherboards in that it's powered more like a laptop, using an external 19VDC power adapter. At the left corner you can see the standard socket for the ubiquitous HP laptop power adapter.

This is a closeup of the power adapter I got for this project. I figured that given the system I want to build, 90W would be enough:

A little spoiler here. The mains input cord for this power adapter has three prongs, but the ground prong is not connected to anything inside the power adapter. I believe this adapter belongs to a class of devices called double insulated, although it seems to lack the square marking for it (someone could perhaps verify certain markings on it are equivalent, although I don't have much uncertainty about it). So the main thing is that this adapter's input is not grounded and neither is its 19VDC output. Grounding will be a central point in this context.

Let us move on to the case:

I live in Europe, but I had to have this beautiful US made thing fly over the Atlantic. It is made of aluminum with a sand blasted finish and although very small (220mm x 220mm x 100mm), it can also accommodate normal "thick" mini-ITX boards. As mentioned, mine is thinner so this leaves some extra space empty inside the case. Which brings us to the following close up of the motherboard power socket:

Notice the white connector behind the power socket. This allows for powering the board with the exact same 19VDC, but from an internal power source.

Since the space available in the case allows for attaching the laptop power adapter, I thought why the hell not? This would yield an even more compact system that is easier to carry around (although slightly heavier).

However, this could be a major change here: the ungrounded mini-ITX case will now have to deal with two 230VAC power leads going into it. I suspect that if I am to be a good boy, I should also provide grounding, a third lead.

Now the first question arises. Let us suppose I install a three pin socket on the case that has its ground pin connected to the motherboard I/O backplate, thus connecting mains ground to multiple places on the motherboard, its ground level. Using a multimeter, one can easily discover that the DC- part of the white power connector on the board will be connected to ground. It is one of the multiple grounded places.

However, the laptop power adapter's design did not include a ground facility and its two 19VDC leads will be connected to the internal white power connector. So when grounding the motherboard's I/O backplate, I will be grounding the power adapter's DC- output lead. The question is, could this be problematic?

Thanks for reading up to now.

 

[rjenkinsgb]

Not a problem at all.
If you eg. connect a laptop to any other piece of grounded equipment, it's doing just the same thing.

Including it is a good safety precaution.

 

[Diver300]

However, this could be a major change here: the ungrounded mini-ITX case will now have to deal with two 230VAC power leads going into it. I suspect that if I am to be a good boy, I should also provide grounding, a third lead.

I assume that you mean a single power source, with three wires in it, just like any mains lead to a metal appliance.

I think that it's just terminology, but "power lead" would normally mean what is molded onto a mains plug. It might have two or three wires inside it, but it is just one lead that is plugged in and used as a single item.

Now the first question arises. Let us suppose I install a three pin socket on the case that has its ground pin connected to the motherboard I/O backplate.

Again, terminology. It would be a three pin plug, more often called a power inlet, that you install on the case, not a socket. It would be one of these:-

https://www.surpluscenter.com/produ...0+VAC+IEC+STANDARD+C14+POWER+INLET+PLUG_L.jpg

and a power lead would connect to that.

You can get panel-mount sockets, but I don't think that you want one here. A panel-mount socket was sometimes fitted to desktop computers so that a monitor could be plugged into the back of the computer, allowing the monitor to be turned on and off with the computer.

You have to use the plugs and sockets in a way that there are no live plugs when disconnected, as the pins would be easy to touch and get a shock from. There may be live sockets, but they are made so that the live bits can't be touched.

 

[jester]

Not a problem at all.
If you eg. connect a laptop to any other piece of grounded equipment, it's doing just the same thing.

Including it is a good safety precaution.

That's great news, it's good to know that the power supply won't mind having its DC- wire grounded.

I assume that you mean a single power source, with three wires in it, just like any mains lead to a metal appliance.

Correct, it's just terminology. I mean the standard insulated mains power cable with three wires, hot, neutral, ground. Hot and neutral will connect to the power supply 230VAC input. Ground will connect to the I/O backplate - motherboard ground level. This effectively connects mains ground to the power supply DC- output wire, as mentioned earlier.

Again, terminology. It would be a three pin plug, more often called a power inlet, that you install on the case, not a socket. It would be one of these:-

https://www.surpluscenter.com/produ...0+VAC+IEC+STANDARD+C14+POWER+INLET+PLUG_L.jpg

and a power lead would connect to that.

Again, correct. You can effectively think of it as a C14 power inlet, although a C14 would be kind of bulky for this application. I am considering GX type lead connector (female, no exposed prongs) and power inlet (male, with prongs). The back panel of the case already has a bore that can accommodate such a power inlet.

You can get panel-mount sockets, but I don't think that you want one here. A panel-mount socket was sometimes fitted to desktop computers so that a monitor could be plugged into the back of the computer, allowing the monitor to be turned on and off with the computer.

You have to use the plugs and sockets in a way that there are no live plugs when disconnected, as the pins would be easy to touch and get a shock from. There may be live sockets, but they are made so that the live bits can't be touched.

I happen to own such a computer from the late 80s. Horizontal case, with a C14 input and a C13 output for the monitor. It is a safe design since the connectors that provide power are all C13. I won't be powering a monitor this way from this build, and will make sure the power lead connector is female (untouchable) with a male power inlet on the mini case.

Your mentioning of connection - disconnection is very timely. I intend to follow up on this later as it is one of my concerns.

Thank you guys for taking the time to respond.

 

[augustinetez]

I'm assuming the motherboard will be mounted to the case with metal standoff's and screws? It would be better to do that so the motherboard is earthed at all the various mounting points.

Also, the ground (earth) wire from the mains lead should be terminated directly to the aluminum case with it's own screw to meet proper safety requirements.

 

[Pommie]

Back in the 80s we bought some USA TV's for use with Nintendo NES units. All was good untill the two units had different polarity and then smoke. Glad the USA now has a ground connection. For clarity, we were using japanese NES,s so needed the 230V.

Mike.

 

[jester]

I'm assuming the motherboard will be mounted to the case with metal standoff's and screws? It would be better to do that so the motherboard is earthed at all the various mounting points.

Also, the ground (earth) wire from the mains lead should be terminated directly to the aluminum case with it's own screw to meet proper safety requirements.

Here's an exploded view of the case:

The motherboard is indeed attached to metal stand offs, and ones that are unusually long mind you, because of its fanless cpu cooling design. There is a catch however. The four side panels, top-bottom panel and also corner pillars are all aluminum, but covered with a sandblasted/anodised finish that seems to be non conductive. I tested with a multimeter and there's no conductivity between any two points of the bottom panel for example. So theoretically, the only dangerous part will be the I/O backplate, that sits in the back panel.

Please disregard the power connector candidate for now (it's a GX12 4-pin + frame for ground). This is to show the exposed metal backplate mostly. And also, since I was counting on the power inlet's frame to connect to ground through the power lead connector (which is also made of metal), this shows another problem. Fastening the power inlet on the back panel alone does not provide grounding due to the finish.

I have verified that the four metal standoffs are indeed connected to motherboard ground when the motherboard is attached. But they aren't going anywhere helpful either, just the finished inner side of the top panel.

How would you go about grounding this then?

 

[augustinetez]

Referring back to the pic of the power supply, as commented, it doesn't have the double insulated symbol anywhere I can see (all the other markings are the approval marks of various countries).

You say the ground pin is not connected to anything inside the power supply - is this from opening the supply and actually looking inside?

Reason for asking is some power supply's have the DC side earthed via a capacitor back to the ground pin.
If that is the situation, then independently earthing the DC side as you propose is a no-no as it will negate the AC-DC safety barrier separation inside the supply.

Nice case - personally, I would be keeping the mains outside it.

Earthing the backplane only could lead to a potentially dangerous situation if, somehow, the incoming mains lead gets damaged and comes in to contact with a non-earthed part of the case.

 

[jester]

You say the ground pin is not connected to anything inside the power supply - is this from opening the supply and actually looking inside?

Reason for asking is some power supply's have the DC side earthed via a capacitor back to the ground pin.
If that is the situation, then independently earthing the DC side as you propose is a no-no as it will negate the AC-DC safety barrier separation inside the supply.

I have indeed opened the power supply (a little naptha did wonders dissolving the glue) and the ground pin was there just for added stability, definitely not connected to anything. I presume without a ground pin there is no possibility for the DC side earthing you describe, and there's no danger.

 

[augustinetez]

Pretty sure you'll be ok in that case.

 

[jester]

Pretty sure you'll be ok in that case.

I see what you did right there.

 

[jester]

Your mentioning of connection - disconnection is very timely. I intend to follow up on this later as it is one of my concerns.

So, with grounding (hopefully) out of the way, allow me to "deviate" a bit with a short story about a laptop and its user:

"Once upon a time, a certain web developer turned on his laptop using just its battery power. He saw that the battery level was a little over 30% so he thought it might be a good idea to plug the device in for charging while he worked. He reaches for the power adapter. First, he plugs the DC out of the adapter to the laptop. Then, he inserts the adapter's AC power lead to the mains socket. A spark/arc occurs, as it often happens on such occasions. But the laptop is not unaffected. In fact it does a hard reset and starts over. Thankfully, it seems there was no permanent damage done to the hardware, as it still worked without problems two years after the event. But ever since that day, and to be on the safe side, the web developer plugs in the other way: First plug in to mains socket, then connect adapter's DC out to the laptop's power socket."

The question "which one should I plug in first, the laptop's AC input or its DC output to the laptop?" comes up pretty often on various places in the internet. The most usual answer is "it really doesn't matter, you can do it any way you want as the adapter and laptop are designed to handle both ways". But for whatever reason, my earlier personal story indicates otherwise. It is possible that in a given setup, the two ways to connect to power are not equivalent.

Each way of connecting - disconnecting is sometimes associated with a theoretical problem that might occur:

- If you plug in the DC out first (as in my story), and then you plug in to mains, there's the theoretical danger that DC out fluctuates outside the laptop specifications during that initial transient due to inductive phenomena on the two coils of the SMPS (I think it's about time we started calling it by its real name). My story suggests such a scenario.

- If you plug the mains first to the SMPS, the SMPS has the time to recover from the initial transient, have its capacitors fully charged, and have a steady output voltage as per specifications by the time you connect it to the laptop. BUT, with its capacitors fully charged, as soon as it is connected to the laptop's charging circuit capacitors, a strong so called inrush current moves into the laptop. Now, laptop charging circuits are supposed to be able to handle such currents, but still there's people that advise plugging first to the laptop, then to the mains for this exact reason.

An initial question to you guys would be, do you have a preferred method of plugging in and if yes, which one? An equivalent question would be, which problem (DC out fluctuation or inrush current) would you consider more dangerous and likely to cause damage to laptop hardware? There's of course the case I am getting something wrong and seeing ghosts where there aren't any.

Of course this question is related to my small factor build which uses a laptop SMPS power adapter. And a small note, the motherboard does not have a charging circuit, it has no battery backup.

I shall further elaborate as the discussion unfolds, hopefully.

Thanks all!

 

[rjenkinsgb]

A decade or so back I repaired a couple of laptops for friends that had been damaged by plugging in an already-powered PSU.

I suspect better quality and newer machines have a soft-start type system to allow the internal PSU to charge up before fully activating and drawing high current.

I usually connect the PSU to the machine first just on principle, but I don't think it matters with any decent machine.

 

[jester]

A decade or so back I repaired a couple of laptops for friends that had been damaged by plugging in an already-powered PSU.

I suspect better quality and newer machines have a soft-start type system to allow the internal PSU to charge up before fully activating and drawing high current.

I usually connect the PSU to the machine first just on principle, but I don't think it matters with any decent machine.

For discussion's sake, is there a simple way to protect those laptop motherboards from accidentally connecting a fully charged adapter, knowing they are sensitive to inrush current?

For example, I can think of an imaginary cable that is permanently connected to the laptop, and has some kind of capacitance integrated so that when the fully charged adapter is connected to the cable, the capacitance absorbs that inrush current before it reaches the laptop battery charger. I'm just making a wild assumption here, not understanding electronics well enough.

 

[augustinetez]

A properly designed SMPS 'shouldn't' have anything nasty on the DC output when plugged in to the mains.

Cheap and nasty SMPS things are a different story - they usually rely on the DC side being loaded to prevent anything like that happening.

I plug in our laptops SMPS either way and then turn on the mains, I never plug it in to an already on mains socket.

Of course, once these things get to be a few years old, the capacitors inside are starting to fail, that's when you get problems.

For discussion's sake, is there a simple way to protect those laptop motherboards from accidentally connecting a fully charged adapter, knowing they are sensitive to inrush current?

Don't know of any "off the shelf" device that will do that for your application, but you would be looking at something like a soft start device - however, as with a lot of computer type stuff, a slowly (relatively speaking) rising voltage generates other problems like lockups or failure of the onboard power supplies to properly start up.

As RJ says above, most if not all modern stuff should be able to cope with an already on supply being connected.

 

[jester]

A properly designed SMPS 'shouldn't' have anything nasty on the DC output when plugged in to the mains.

Cheap and nasty SMPS things are a different story - they usually rely on the DC side being loaded to prevent anything like that happening.

I plug in our laptops SMPS either way and then turn on the mains, I never plug it in to an already on mains socket.

I can't help but think of my laptop restart incident when I plugged the power cable last.

Why do you do what you do? In other words, how would plugging to an already on mains socket be any different to flicking a switch on the mains socket with the cable plugged in?

Btw, here in weird Europe, our mains sockets usually don't come with switches. So we 're stuck with the first option, at least when our device doesn't come with an AC power switch. That would be the potentially problematic case with a laptop power supply. But of course, there's the good news that in some cases you can add an AC switch when none exists (hint hint).

 

[augustinetez]

Plugging a plug in to an already on mains socket is more prone to arcing - sometimes you don't quite get the plug in the correct position and may have several goes at it, whereas most switches on mains sockets are very positive snap action and designed to minimise acing.

The arc comes about from close proximity to the contact not when it makes a proper firm mating with the contact.

There is the proviso that older mains sockets/switches can lose that snap action through aging/metal fatigue and will arc at the switch, but these should be replaced as they can become a fire risk - something I know from first hand experience.

Re the non-switched sockets, I do remember seeing at some stage a few years ago a little adapter lead that was the same plug/socket combination needed for the power supply that also had an inline on/off switch.

They came about when everyone got worried by all the electricity wastage caused by the standby current of the millions of plugpacks/power supplies and other appliances left plugged in and not being used.

Don't know if these are still available.

 

[jester]

Plugging a plug in to an already on mains socket is more prone to arcing - sometimes you don't quite get the plug in the correct position and may have several goes at it, whereas most switches on mains sockets are very positive snap action and designed to minimise acing.

The arc comes about from close proximity to the contact not when it makes a proper firm mating with the contact.

Thanks for your response. I confess I would intuitively do the same thing if I had mains outlets with switches. Always plug in, then switch the outlet on.

I wonder however, why is arcing inherently bad and to be avoided? Is it for safety (short circuit, fire) or does it also have to do with a worse initial transient of an SMPS for example, that could make it output out of spec DC voltage momentarily (especially if it's a cheap SMPS).

Let's for a moment do not consider safety-shorting-fire, assuming an arc would be a bad event for starting up an SMPS (which I don't know if it's true). I 've read somewhere about arcs that, when you 're using a switch, you are not avoiding potential problems with the initial transient of a device (always assuming arcs have a bad effect on the device). You are just hiding these potential problems (perhaps a much smaller arc is created inside the switch, which you can't see). Could this be true?

Then there's the initial effects of self induction and/or capacitance in the SMPS. Arc or no arc, these come into play. If these effects can cause damage to a laptop's charging system (or a mini motherboard without a charging system), would it make sense to isolate the device from the SMPS until it settles down? This would be the equivalent of connecting a power adapter to your laptop last, after the adapter has settled down (but is full of charge).

If anyone's suspecting I will be adding switches to the system they are right. One will be a DPDT AC switch (so I will effectively have a "mains power switch". This will interrupt both neutral and hot as power outlets are symmetrical where I live.

A second switch could be used to isolate the motherboard from the SMPS till it settles down. This would be a single pole switch as I will be interrupting just DC+. DC- is always connected to ground. This would mimic the DC connectors of laptops that are designed so that DC- (ground or "ground") connects first, then the tip connects afterwards and delivers DC+. Again, I have read that these DC inlets are designed this way in order to properly handle potentially high initial inrush current from a fully charged SMPS. Supposedly ground will help in that case, especially if it's a true ground.

I suspect the AC switch is a must. After all, every desktop PC power supply has one. And US home power outlets have them too. The DC switch I don't know. I wonder if the switches will be of help or useless or potentially damaging, depending on how I start the device.

My primary objective is to protect the motherboard because a lot of other hardware depends on it, and I 'd have to throw everything away if it gets damaged. I don't mind having two switches or a more complex startup procedure if that will protect the motherboard.

Apologies for the longer post, but I think it was about time I stopped being too cryptic about a relatively simple practical issue. And as you can probably tell, I am the curious kind, but with not enough knowledge.

 

[augustinetez]

I wonder however, why is arcing inherently bad and to be avoided?

Because an arc is a series of undefined voltage pulses, some very short duration and generating harmonic frequencies that can be imposed on the supply in to the SMPS.

These harmonics might beat with the units own natural running frequency and cause weird effects, some harmless, others not so - think making the SMPS run at the incorrect frequency and generating wrong or no voltages with possible consequential damage to the electronics.

In some cases these harmonics/pulses can make their way to the output side of the SMPS and do things such as you have seen.

Then there's the initial effects of self induction and/or capacitance in the SMPS.

Well, keeping it simple, without these, the SMPS just isn't going to work. An SMPS relies on inductance and capacitance to maintain oscillation at the correct operating frequency and if you are talking spurious inductance/capacitance, it is taken in to account in the design (usually).

would it make sense to isolate the device from the SMPS until it settles down?

Why?

Think about what is going to happen in either case - plugged in and fired straight up or wait for the SMPS supply to 'settle'.

First case, SMPS is going to supply voltage and current to the mainboard power circuits basically in a smoothish ramp up and relatively fast. There may be a slight overshoot in the voltage, but the on board power system should cope with that.

Second case, you fire up the SMPS and wait for it to stabilise. Then you throw the switch and send a torrent of voltage and current cascading in to the on board power supplies, because of the sudden heavy load on the SMPS, it may dip in voltage slightly then overshoot a bit as it recovers.

Bottom line - both scenarios produce roughly the same effect (although the second one is slightly more violent).

My primary objective is to protect the motherboard

If the motherboard is new, of relatively recent manufacture, it should cope with having the SMPS turned on at the mains and let it run, if it doesn't cope then it was either faulty or just plain rubbish to start with.

 

[jester]

Because an arc is a series of undefined voltage pulses, some very short duration and generating harmonic frequencies that can be imposed on the supply in to the SMPS.

These harmonics might beat with the units own natural running frequency and cause weird effects, some harmless, others not so - think making the SMPS run at the incorrect frequency and generating wrong or no voltages with possible consequential damage to the electronics.

In some cases these harmonics/pulses can make their way to the output side of the SMPS and do things such as you have seen.


Well, keeping it simple, without these, the SMPS just isn't going to work. An SMPS relies on inductance and capacitance to maintain oscillation at the correct operating frequency and if you are talking spurious inductance/capacitance, it is taken in to account in the design (usually).


Why?

Think about what is going to happen in either case - plugged in and fired straight up or wait for the SMPS supply to 'settle'.

First case, SMPS is going to supply voltage and current to the mainboard power circuits basically in a smoothish ramp up and relatively fast. There may be a slight overshoot in the voltage, but the on board power system should cope with that.

Second case, you fire up the SMPS and wait for it to stabilise. Then you throw the switch and send a torrent of voltage and current cascading in to the on board power supplies, because of the sudden heavy load on the SMPS, it may dip in voltage slightly then overshoot a bit as it recovers.

Bottom line - both scenarios produce roughly the same effect (although the second one is slightly more violent).


If the motherboard is new, of relatively recent manufacture, it should cope with having the SMPS turned on at the mains and let it run, if it doesn't cope then it was either faulty or just plain rubbish to start with.

Great response augustinetez, clears things up in one simple stroke. Many thanks.