DSM-G600, DNS-3xx and NSA-220 Hack Forum

smallmeter · 2007-09-11 06:55:37

I would like the DNS-323 to power on automatically when the power is applied. I can hook the thing to a UPS, but eventually it will need to power down until the electricity comes back on. Holding down the power switch doesn't help. It seems really strange to design a SOHO server that doesn't come up automatically, but I can't find a way to do this. I've been unsuccessful tracing where the power switch goes. One lead goes to +5V on the input power supply, but I can't find where the other lead goes.

Anyone have any ideas?

fickle · 2007-09-11 08:49:29

i dont think this is possible. from what i have read here, its a nice to have of which the 323 does not support at this time.

Speijk · 2007-09-11 11:03:31

Same problem here.

I want to use the DNS-323 as a backup and planned to put it under my (concrete) floor. This is not a solution if I need to crawl under my floor to switch the unit on after a powerdip

danair · 2007-09-11 11:42:28

I have mine up in the eves of my house - not perfectly accessible but if I need to pop up to the loft to switch it back on or reset it if it crashes (not that it ever has) it is an inconvenience but perfectly possible. I certainly wouldn't put it under a concrete floor though!

fickle · 2007-09-12 01:19:43

has anyone contacted dlink to find out if auto on after power failure would be a feature in a possible new firmware release?

smallmeter · 2007-09-12 02:40:09

fickle wrote:
has anyone contacted dlink to find out if auto on after power failure would be a feature in a possible new firmware release?

I have tried, but I got no response at all.

Aaron · 2007-09-14 19:11:46

AFAIK the unit requires a hard switch power on. Meaning no firmware is going to change that.

-Aaron

w00 · 2007-09-14 22:53:27

auto power on shoulnt be impossible on the dns 323. since i come from a linksys nslu2 i would like to point you towards different solutions for the nslu2, see here:

http://www.nslu2-linux.org/wiki/HowTo/F … erAlwaysOn

there might be some useful hints on how to do it on the dns323 too. if someone finds something usefull there please let us know here!

cheers

fordem · 2007-09-15 16:51:28

Aaron wrote:
AFAIK the unit requires a hard switch power on. Meaning no firmware is going to change that.

-Aaron

I can't say that I agree with you - but it may be just a matter of interpretation.

A "hard" power switch is either on or off - like a toggle switch similar to what you probably use for the lights in your home. If you've been around long enough, you'll be familiar with the older AT & miniAT based PCs which had either a hard switch, either a toggle switch with a lever, a rocker switch or in some cases a push-on/push-off button.

What's important to recognise is that in one position it's off, in the other position it's on - there is no ambiguity - and you can leave it in the on position and turn whatever it controls on or off by connecting or disconnecting the power to the device.

A "soft" power switch is usually a momentary contact switch that signals logic somewhere to do something - newer ATX PCs have soft switches - first push turns it on, second push tells it to do something else - hibernate/standby/shutdown - and push and hold overrides the logic and forces the PC off, useful when the OS hangs.

It's easy to tell if your computer has a hard or soft switch - PCs with soft switches turn themselves off when you shut Windows down, PCs with hard switches display a message saying "It is safe to turn off" or something along those lines.

PCs with soft switches can usually be programmed to do one of three things when power is applied - always turn on, always stay off, or resume the last state whichever one it was.

The above is deliberately kept simple

Your DNS-323 has a soft switch - a momentary contact switch that sends a signal to the logic - first push turns it on, push & hold does an orderly shutdown - and you can also shutdown (to the off state) from the admin web pages.

Whether or not the hardware contains the circuitry that would be needed to sense power being applied to allow an automatic power on is another matter all together - if it does then a firmware upgrade could make auto turn on a possibility.

smallmeter · 2007-09-21 18:29:39

I had a thought. It should be possible to design a small circuit that bridges the power button with a relay. When power is applied, it delays a few seconds then momentarily activates a relay to power the thing on.. Shouldn't be too hard to design a circuit like this. I did a test and 2-3 seconds should be enough. A transistor, a few capacitors, and a relay should do it.

Shoot, I could do this with my kid's electronics kit :-).

jboy · 2007-09-23 02:58:48

smallmeter wrote:
I would like the DNS-323 to power on automatically when the power is applied. I can hook the thing to a UPS, but eventually it will need to power down until the electricity comes back on. Holding down the power switch doesn't help. It seems really strange to design a SOHO server that doesn't come up automatically, but I can't find a way to do this. I've been unsuccessful tracing where the power switch goes. One lead goes to +5V on the input power supply, but I can't find where the other lead goes.

Anyone have any ideas?

I went with the UPS setup. I also have a UPS for my Infrant.

IMHO having the unit autopower when the power is applied is not the best idea. Around here the power usually bounces a few times before it stabilizes. In the good old days of mechanical switches I lost a few drives because of these bounces.

jcatech · 2007-09-29 05:33:30

This can be done with just a capacitor (C1) and a resistor (R1). I think the power signal is +5 volts, so the capacitor and the resistor need to be connected in serie; the end of the resistor goes to the ground and the end of the capacitor goes to to the 5v, and the node between the capacitor and the resistor goes to the power signal line (the line that connects to the power switch and goes to the chips).

When you plug the power to unit, the power supply will inject 5v to the circuit and the capacitor will start charging; during the charging time part of the currect will run through the power signal line and this will turn on the unit and then when the capacitor finished charging the currect will stop running through, so the unit stays on.

The values of R1 and C1 determine the charging time. The greater these values, the longer the time. The equation to determine charging time is as follows:

C1 (in farads) x R1 (in ohms) = Time (in seconds)

The value of C1 is in farads. Most capacitors are listed in micro-farads. To use the formula, divide the number of microfarads by 1,000,000.

So start building your own circuit and have fun.

Notes:
1Megaohm x 1uf = 1s
I found this web page that explains how this works:
http://www.techitoutuk.com/knowledge/el … capac.html

Last edited by jcatech (2007-09-29 05:48:41)

smallmeter · 2007-10-01 15:45:36

jboy wrote:
I went with the UPS setup. I also have a UPS for my Infrant.

IMHO having the unit autopower when the power is applied is not the best idea. Around here the power usually bounces a few times before it stabilizes. In the good old days of mechanical switches I lost a few drives because of these bounces.

What happens if your UPS runs out of power?

What I"m talking about is not a hard power switch. You don't want that. It's something that will press the soft power switch automatically.

smallmeter · 2007-10-01 15:52:10

jcatech wrote:
This can be done with just a capacitor (C1) and a resistor (R1). I think the power signal is +5 volts, so the capacitor and the resistor need to be connected in serie; the end of the resistor goes to the ground and the end of the capacitor goes to to the 5v, and the node between the capacitor and the resistor goes to the power signal line (the line that connects to the power switch and goes to the chips).

Unfortunately, it's not this simple. You want to wait a few seconds, then short the leads on the switch, then release them again since you don't want the switch on permanently. You don't want the circuit to draw much power. I don't know if the switch connection could take the voltage drop across a transistor, or if a relay would be required. This is not as simple as it sounds.

DSM-G600, DNS-3xx and NSA-220 Hack Forum

Announcement

#1 2007-09-11 06:55:37

Automatically power on when power is applied

#2 2007-09-11 08:49:29

Re: Automatically power on when power is applied

#3 2007-09-11 11:03:31

Re: Automatically power on when power is applied

#4 2007-09-11 11:42:28

Re: Automatically power on when power is applied

#5 2007-09-12 01:19:43

Re: Automatically power on when power is applied

#6 2007-09-12 02:40:09

Re: Automatically power on when power is applied

fickle wrote:

#7 2007-09-14 19:11:46

Re: Automatically power on when power is applied

#8 2007-09-14 22:53:27

Re: Automatically power on when power is applied

#9 2007-09-15 16:51:28

Re: Automatically power on when power is applied

Aaron wrote:

#10 2007-09-21 18:29:39

Re: Automatically power on when power is applied

#11 2007-09-23 02:58:48

Re: Automatically power on when power is applied

smallmeter wrote:

#12 2007-09-29 05:33:30

Re: Automatically power on when power is applied

#13 2007-10-01 15:45:36

Re: Automatically power on when power is applied

jboy wrote:

#14 2007-10-01 15:52:10

Re: Automatically power on when power is applied

jcatech wrote:

Board footer