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

tinybilbo · 2012-01-10 00:07:09

Hi,

My local electricity board fried every bit of equipment on one of my home circuits, that was connected to a certain phase.
About a dozen power supplies (they protected the actual equipment by taking the brunt) and a gigabit switch have died as a result.
Everything was on surge protectors - which have all survived.... go figure...

The electricity board have accepted liability, and I have begun the process of replacing everything.

However, I live on a small island, and it is'nt easy to get stuff here.
I can replace the simple power supplies from local stock, i.e 12v 1amp stuff.
However my 2 dns323's which serve content throughout the house as well as backing up my Laptops, use dual voltage (12v and 5v), (like an ATX PC power supply).
I can't find those here, and the one ebay seller I found, will only ship to the UK (where I'm from and my account is registered to).
I not due back to the UK for a few months.

So I've set about converting a PC ATX power supply by ripping out the wire and connector from the dead Jentec 323 power supplies.
I then connected them to the 12v and 5v rails (red for 5v, yellow for 12v), as well as the two grounds to gound (black) inside the ATX supply.
I have then shorted the "power on" wire (green) to the ground (black) on the ATX power supply, so that it powers on when its plugged in.
This remains shorted.

This works after a fashion.
The power supply turns on, when plugged into the wall, and delivers correct voltages (measured)...

However, when I turn on the 323, it's a fifty-fifty chance that it powers up,
sometimes it boots, but turn it off and then turn it back on, and it might or might not boot again.
It is the power supply that is failing, not the 323, because I must "reset" the power supply after a failed boot,
by unplugging it from the mains and then plugging it back in - then I'm back to a fifty-fifty chance of booting.

Googling around - I see that I must stabilize the voltages on the rails by supplying a load on the 5v rail,
by connecting a 10w, 10ohm power resistor - which I have duly done. I still have the same problem.

I have tried 2 power supplies now (I've got about a dozen old ones lying around), and they both exhibit the same behaviour.

I know there are "power ok" wires (gray - i think) and a couple of "sense" wires (orange and red) in the ATX power supply, but I have'nt done anything with these wires.

Has anyone made a similar mod and got it fully working?

Any help appreciated...

Thanks

fordem · 2012-01-10 14:43:31

I've never tried to power my DNS-323 from an ATX power supply, but I have some experience in repairing the ATX power supplies - the fact that you must "reset" the power supply after a failed boot suggests that the problem is caused by overloading it.

When you overload (pull higher currents than the supply can deliver) a PC power supply it will current limit & shutdown, and the only way to restart is to power it off & back on. It sounds strange that a low wattage unit like a DNS-323 could overload a power supply with a higher wattage rating, but those are the symptoms you're describing - it MAY be possible that you are overloading one of the power supply rails - some ATX supplies have multiple +5V rails - and another possibility is what is called the "inrush" current (the current required to start the drives spinning) is higher than the supply can handle.

The power good wires can be ignored - they are used to hold the PC in a reset state if the incoming power is out of specification, it's an outdated feature dating back to the days of the IBM PC5150, the orange & red wires are not sense wires - orange wires are a +3.3 output, and ALL red wires are +5V output.

tinybilbo · 2012-01-10 15:39:20

Hi Fordem - Thank you for your reply.

fordem wrote:
the fact that you must "reset" the power supply after a failed boot suggests that the problem is caused by overloading it....

It sounds strange that a low wattage unit like a DNS-323 could overload a power supply with a higher wattage rating, but those are the symptoms you're describing - it MAY be possible that you are overloading one of the power supply rails - some ATX supplies have multiple +5V rails - and another possibility is what is called the "inrush" current (the current required to start the drives spinning) is higher than the supply can handle.

It does appear to some kind of overload protection, The unit does briefly flash, then the PSU cuts out.
I have booted the units with and without drives, - Still the same behaviour...

I have tried 3 PSU's now - older 400w units and fairly cheap generic ones at that, The 1st I directly soldered to the board, the 2nd I just cut the wires internally and taped them together, the 3rd I did'nt even open... I just jammed the wires into a Molex connector and taped it.
I have put loads (10w 10ohm) on both the 12v and 5v rails, Individually, and combined (I did'nt try the 3.3v - maybe I will tonight - just to be sure ).
Also there appears to be only 1 rail for each voltage in the 2 PSU's I've opened.

No matter what combination I make, It's always the same fifty-fifty chance of overloading the PSU....

fordem wrote:
The power good wires can be ignored - they are used to hold the PC in a reset state if the incoming power is out of specification, it's an outdated feature dating back to the days of the IBM PC5150, the orange & red wires are not sense wires - orange wires are a +3.3 output, and ALL red wires are +5V output.

Thanks for clearing that up...

It's possible that both the DNS-323's got damaged during the power spike, however I can get them to boot, It's just hit or miss.
So my feeling is that they are not damaged.

A couple of points to note.
1) One of the leads salvaged from the Jentec PSU's has a much higher chance of booting - say seventy-thirty.
I have checked both with a meter and both leads are good.
2) I noticed that inside the leads, they are shielded with the usual plastic backed silver foil.
Also one of the "ground" wires is exposed, while the other is covered in the usual rubber/plastic.
Could the shielding be causeing some kind of short with the shield and exposed ground?, and if so, how could I eliminate that...

It's frustrating that I'm halfway there... but not quite...

I have noticed a few threads about replacing/repairing/losing the Jentec PSU's and this could be a cheap stopgap solution, as well a way of powering multiple DNS-323's from a single source. If I can get it to work - I will buy a low-noise PSU and use that to power both my DNS-323's.

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

Announcement

#1 2012-01-10 00:07:09

Power surge fried my power supplies - need help building alternatives.

#2 2012-01-10 14:43:31

Re: Power surge fried my power supplies - need help building alternatives.

#3 2012-01-10 15:39:20

Re: Power surge fried my power supplies - need help building alternatives.

fordem wrote:

fordem wrote:

Board footer