news post: Efficient PSUs that don't start
Bluefront
OK if you are speaking about making tests with a common light to see if the failure disappears, but for a light to be installed permanently I think the usual loads drawn by these lights (commonly LED or light-emitting tubes) are too weak. Don't you think so ?
Skromnibog
Thank you for the informations.
j-p
OK if you are speaking about making tests with a common light to see if the failure disappears, but for a light to be installed permanently I think the usual loads drawn by these lights (commonly LED or light-emitting tubes) are too weak. Don't you think so ?
Skromnibog
Thank you for the informations.
j-p
Last edited by cavok on Tue Oct 03, 2006 11:43 am, edited 1 time in total.
Consider separate voltages ?
Thank you Mike.
Since your post on Saturday, I worked a lot to known how the different voltages are distributed in the different output lines of a PSU ; it's not so easy to find on the web.
It's now a little clearer for me.
I can connect several fans directly to the PSU for testing the hypothesis, but unfortunately I couldn't keep them because I only have one more location available for a fan of only 80 mm.
My main concern, at the moment, is to know if I have to optimize the load for every rail (I don't know if the rails come from independent lines or not, in the upstream, inside the PSU), or must I do it according to every category of voltage, or have I to consider the global load for the whole PSU ?I am inclined to believe that the second proposition (considering load for every kind of voltage) is the right one.
The minimum load requirements are 1A for 12v (X2), and also 1A for the +5v, and 0.5A for the +3.3v currents.
http://www.fsp-group.com.tw/english/1_p ... &proid=127
If the load has to be considered as a whole, I can hope that adding any device drawing a lot could fix the problem.
I understand that if the device is plugged into the motherboard the result depends on which voltage is involved.
If it is a graphic card on the PCIe port, I suppose that the current drawn would be ok because there is a specific 12v rail for it.
Am I right ?
j-p
Since your post on Saturday, I worked a lot to known how the different voltages are distributed in the different output lines of a PSU ; it's not so easy to find on the web.
It's now a little clearer for me.
I can connect several fans directly to the PSU for testing the hypothesis, but unfortunately I couldn't keep them because I only have one more location available for a fan of only 80 mm.
My main concern, at the moment, is to know if I have to optimize the load for every rail (I don't know if the rails come from independent lines or not, in the upstream, inside the PSU), or must I do it according to every category of voltage, or have I to consider the global load for the whole PSU ?I am inclined to believe that the second proposition (considering load for every kind of voltage) is the right one.
The minimum load requirements are 1A for 12v (X2), and also 1A for the +5v, and 0.5A for the +3.3v currents.
http://www.fsp-group.com.tw/english/1_p ... &proid=127
If the load has to be considered as a whole, I can hope that adding any device drawing a lot could fix the problem.
I understand that if the device is plugged into the motherboard the result depends on which voltage is involved.
If it is a graphic card on the PCIe port, I suppose that the current drawn would be ok because there is a specific 12v rail for it.
Am I right ?
j-p
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
cavok,
The problem with devices powered off the motherboard is that they may all be in the same delayed turn-on sequence and therefore not affect the initital current draw at all -- which is the main problem. Once all the components are powered up and running, almost any system would meet the minimum current draw requirements of the PSU.
And what was the result?? Just answering that would make the rest of your query unnecessary. IE, if the added fans allow the system to start consistently and run stable, then you only have to worry about the 12V rail. It does not matter whether this is because the 3.3 and 5V lines are getting enough load already or because they don't need minimum load if the 12V load is high enough.I can connect several fans directly to the PSU for testing the hypothesis, but unfortunately I couldn't keep them because I only have one more location available for a fan of only 80 mm.
The problem with devices powered off the motherboard is that they may all be in the same delayed turn-on sequence and therefore not affect the initital current draw at all -- which is the main problem. Once all the components are powered up and running, almost any system would meet the minimum current draw requirements of the PSU.
Mike,
Sorry, I wanted to say that it is not a big deal to do it ; but I have not yet got the fans.
I just keep on hand one 80 mm fan (Cooler master sleeve) but its input current is only 0.15A (IE 1.80 W).
I have hooked it up, right now, to the PSU and that starts ; but I need to do several trials especially after long power off states.
2] The devices that contribute to the initial draw are all those that are directly connected to the PSU I suppose : disks, fans, lights ...
3] What do you think about the suggestion made five posts ago to use lights ?
j-p
Sorry, I wanted to say that it is not a big deal to do it ; but I have not yet got the fans.
I just keep on hand one 80 mm fan (Cooler master sleeve) but its input current is only 0.15A (IE 1.80 W).
I have hooked it up, right now, to the PSU and that starts ; but I need to do several trials especially after long power off states.
OK, well-noted. thank you.MikeC wrote:...if the added fans allow the system to start consistently and run stable, then you only have to worry about the 12V rail. It does not matter whether this is because the 3.3 and 5V lines are getting enough load already or because they don't need minimum load if the 12V load is high enough.
1] But what about PCIe cards that are connected to the PSU via a specific 6 pins rail for extra power ?MikeC wrote:The problem with devices powered off the motherboard is that they may all be in the same delayed turn-on sequence and therefore not affect the initital current draw at all --
2] The devices that contribute to the initial draw are all those that are directly connected to the PSU I suppose : disks, fans, lights ...
3] What do you think about the suggestion made five posts ago to use lights ?
j-p
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
1. I don't know.cavok wrote:1] But what about PCIe cards that are connected to the PSU via a specific 6 pins rail for extra power ?
2] The devices that contribute to the initial draw are all those that are directly connected to the PSU I suppose : disks, fans, lights ...
3] What do you think about the suggestion made five posts ago to use lights ?
j-p
2. yes
3. doable -- if the light draws enough power.
Unfortunately, the addition of a 80 mm fan does not prevent from startup failures.
Furthermore, the failure occurs either after a long power off period or soon after a shutdown.
Anyway, the loads of my configuration seem to be over the requirements if they can be considered as a whole.
My devices are allocated as follow :
~On one rail is attached :
CD-ROM : 5V 1.7A 12V 1A (I don't know how the selection occurs)
CPU fan : 0.90A
80 mm fan : 0.15A
Temperatures display : ?
~On a second rail :
HDD Seagate Medalist : Startup : 12v (peak) 2.2A 5v (RMS) 0.6A Seek 0.5 A ...
~On a third rail (SATA) :
HDD Raptor : 12v : 12v 0.75A 5v 0.90A startup ???
The specifications chart of my PSU shows :
http://www.fsp-group.com.tw/english/1_p ... &proid=127
Output Voltage /Min. load /Max. load /Load Reg. /Ripple & Noise
+3.3V /0.5A /30A /+5% /50mV P-P
+5V /1.0A /28A /+5% /50mV P-P
+12V1DC /1.0A /14A /+5% /120mV P-P
+12V2DC /1.0A /15A /+5% /120mV P-P
-12V /0.0A /0.5A /+10% /120mV P-P
+5Vsb /0.0A /2.0A /+5% /50mV P-P
I don't know what mean +12V1DC and +12V2DC, because, if I don't make mistake, that doesn't refer to the rails, as the PSU has more than two and "Molex" seem to provide either 12V or 5V.
I don't understand how the voltages are dispatched among the rails.
Is of any hope to dispatch my devices among the rails in a better load balanced manner ?
(Difficult, for space and cables lengths considerations, but if it is helpful I will try it).
j-p
Furthermore, the failure occurs either after a long power off period or soon after a shutdown.
Anyway, the loads of my configuration seem to be over the requirements if they can be considered as a whole.
My devices are allocated as follow :
~On one rail is attached :
CD-ROM : 5V 1.7A 12V 1A (I don't know how the selection occurs)
CPU fan : 0.90A
80 mm fan : 0.15A
Temperatures display : ?
~On a second rail :
HDD Seagate Medalist : Startup : 12v (peak) 2.2A 5v (RMS) 0.6A Seek 0.5 A ...
~On a third rail (SATA) :
HDD Raptor : 12v : 12v 0.75A 5v 0.90A startup ???
The specifications chart of my PSU shows :
http://www.fsp-group.com.tw/english/1_p ... &proid=127
Output Voltage /Min. load /Max. load /Load Reg. /Ripple & Noise
+3.3V /0.5A /30A /+5% /50mV P-P
+5V /1.0A /28A /+5% /50mV P-P
+12V1DC /1.0A /14A /+5% /120mV P-P
+12V2DC /1.0A /15A /+5% /120mV P-P
-12V /0.0A /0.5A /+10% /120mV P-P
+5Vsb /0.0A /2.0A /+5% /50mV P-P
I don't know what mean +12V1DC and +12V2DC, because, if I don't make mistake, that doesn't refer to the rails, as the PSU has more than two and "Molex" seem to provide either 12V or 5V.
I don't understand how the voltages are dispatched among the rails.
Is of any hope to dispatch my devices among the rails in a better load balanced manner ?
(Difficult, for space and cables lengths considerations, but if it is helpful I will try it).
j-p
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
The total load or load balance is not the issue -- if minimum start current is the problem -- because it's the timing sequence in which various things are powered up that's the real problem. In other words, the PSU does not see enough load quickly enough. This is why additional simple loads directly on the PSU outputs can work.cavok wrote:Unfortunately, the addition of a 80 mm fan does not prevent from startup failures.
Furthermore, the failure occurs either after a long power off period or soon after a shutdown.
Anyway, the loads of my configuration seem to be over the requirements if they can be considered as a whole.
I don't understand how the voltages are dispatched among the rails.
Is of any hope to dispatch my devices among the rails in a better load balanced manner ?
I will say this one last time: add a 20ohm resistor across the 12V line. If it then starts every time you KNOW that minimum start load is the issue.
Or just return / exchange it for a different model.
-
- SPCR Reviewer
- Posts: 1850
- Joined: Sun Sep 21, 2003 11:23 am
- Location: Vancouver, BC, Canada
No, your specifications are correct, your power supply only has two +12V rails. I think you might be gettings voltage rails confused with individual cables. Just because you have more than two cables that carry +12V current doesn't mean they are electrically separate; the two cables are part of the same circuit, and are joined in the power supply itself.
I think Mike has done a good job of describing what needs to be done to try and fix the problem, so I'll just suggest taking his advice: Try placing a 20 ohm resistor across the +12V line. That means you should insert it on a wire between one of the yellow wires, and one of the black ones, preferably off of the same Molex connector. If you don't feel comfortable doing this, my advice is to replace the power supply entirely. Troubleshooting can only take you so far, especially if you're not familiar with the electronic equipment you're working with. I'm sure you can get a decent price for your Blue Storm, which should help finance your new power supply.
I think Mike has done a good job of describing what needs to be done to try and fix the problem, so I'll just suggest taking his advice: Try placing a 20 ohm resistor across the +12V line. That means you should insert it on a wire between one of the yellow wires, and one of the black ones, preferably off of the same Molex connector. If you don't feel comfortable doing this, my advice is to replace the power supply entirely. Troubleshooting can only take you so far, especially if you're not familiar with the electronic equipment you're working with. I'm sure you can get a decent price for your Blue Storm, which should help finance your new power supply.
Mike, I was talking about global load produced by the devices directly connected to the PSU only, contributing to the initial draw, as identified above : disks, fans, lights ...
I was wondering about the balance between the two rails for these devices only, because their cumulative load seems, if I don't make mistake, to be over the requirements.
But, anyway, I don't know how to identify the two different rails among the cables.
I am sorry to irritate you, but I don't want only to make my PSU start consistently, but also to fill in partially my ignorance, and I was thinking it was doable on the fora of SPCR.
One of my main concerns, for instance, is to understand why such a failure, related to an initial draw, is intermittent.
j-p
I was wondering about the balance between the two rails for these devices only, because their cumulative load seems, if I don't make mistake, to be over the requirements.
But, anyway, I don't know how to identify the two different rails among the cables.
I am sorry to irritate you, but I don't want only to make my PSU start consistently, but also to fill in partially my ignorance, and I was thinking it was doable on the fora of SPCR.
One of my main concerns, for instance, is to understand why such a failure, related to an initial draw, is intermittent.
j-p
Thank you Devonavar.
I figured out that before :
" (I don't know if the rails come from independent lines or not, in the upstream, inside the PSU), " Tue 03 Oct 2006 17:09 GMT + 1 Hour
but, after, I made confusion and made also vocabulary mix-up.
After this reading I got the resistor(s) and I did the manipulation.
I used 2 resistors of 47 Ohms and 5 W (that's what I found) that I connected in parallel to the yellow and black wires of the same "molex". If I am not in a mist, that makes 23.5 Ohms and 10 W.
But, unfortunately, the start failures still occur.
j-p
Yes I made confusion.Devonavar wrote:I think you might be gettings voltage rails confused with individual cables.
I figured out that before :
" (I don't know if the rails come from independent lines or not, in the upstream, inside the PSU), " Tue 03 Oct 2006 17:09 GMT + 1 Hour
but, after, I made confusion and made also vocabulary mix-up.
This is perfectly clear.Devonavar wrote:Try placing a 20 ohm resistor across the +12V line. That means you should insert it on a wire between one of the yellow wires, and one of the black ones, preferably off of the same Molex connector.
After this reading I got the resistor(s) and I did the manipulation.
I used 2 resistors of 47 Ohms and 5 W (that's what I found) that I connected in parallel to the yellow and black wires of the same "molex". If I am not in a mist, that makes 23.5 Ohms and 10 W.
But, unfortunately, the start failures still occur.
j-p
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Sounds like it is time to throw the PSU back to the supplier & get a new one.cavok wrote:After this reading I got the resistor(s) and I did the manipulation.
I used 2 resistors of 47 Ohms and 5 W (that's what I found) that I connected in parallel to the yellow and black wires of the same "molex". If I am not in a mist, that makes 23.5 Ohms and 10 W.
But, unfortunately, the start failures still occur.
j-p
Antec/Asus Incompatabilities
Mike, I may be wrong but I think this issue with HE PSUs first got a lot of attention when the Antec NeoHE 430 first appeared in Antec's P150 case. I happen to have an Asus P4P800-E that I'd like to house in a P150. Because of what I'd read here and in Newegg customer comments I wrote to Antec Support asking if they had a list of Asus mbs that were incompatible with the NeoHE 430, and if the P4P800-E was on the list?
Their response was that they had no list but they had made a revision to the PSU to correct the problem with Asus boards. They did not tell me what the new revision level is. They did not tell me that they had not purged existing P150 stock. They agreed that if I had a problem with my new case they would let me pay shipping to return the bad supply and wait an additional week or two while the exchange(s) took place.
I'm left with the feeling that Antec designs some pretty nice gear but maybe they'd prefer not to be in the business of selling it.
Their response was that they had no list but they had made a revision to the PSU to correct the problem with Asus boards. They did not tell me what the new revision level is. They did not tell me that they had not purged existing P150 stock. They agreed that if I had a problem with my new case they would let me pay shipping to return the bad supply and wait an additional week or two while the exchange(s) took place.
I'm left with the feeling that Antec designs some pretty nice gear but maybe they'd prefer not to be in the business of selling it.
-
- Posts: 5
- Joined: Fri Oct 27, 2006 10:02 am
- Location: Sweden
I have a problem with my new Antec P150 encased system. The mainboard is a Gigabyte GA-P965-DS3. The P150 is probably an older version, since it still has those rubber bands, as I mentioned in another thread. A sticker on the PSU says A21 (revision number?), together with a bar code and a serial number.
Tricool case fan on low setting results in case fan spinning up for a second, then stopping, trying again after a second. CPU fan (mainboard connected) doesn't move.
Tricool case fan on high setting, plus a Nexus 120mm fan, results in case fan spinning up for a few seconds, then stopping, then trying again. The CPU fan starts spinning after a few tries, but the video card remains dead.
Tricool case fan on low setting results in case fan spinning up for a second, then stopping, trying again after a second. CPU fan (mainboard connected) doesn't move.
Tricool case fan on high setting, plus a Nexus 120mm fan, results in case fan spinning up for a few seconds, then stopping, then trying again. The CPU fan starts spinning after a few tries, but the video card remains dead.
Hello.
In November, I bought and connected a Seasonic S12 PSU with sleeving on output cables, 500 W (below they seem to be old Tornado).
So far, my computer starts every time.
Thanks to this and to my trials with resistors, the conclusion is, indeed, an awry PSU (Fortron BlueStorm).
(I was absent for a few weeks and I wanted to get a long term observation because of the intermittence of the previous failure ; that's why I didn't post before).
Thank you for your advices.
j-p
In November, I bought and connected a Seasonic S12 PSU with sleeving on output cables, 500 W (below they seem to be old Tornado).
So far, my computer starts every time.
Thanks to this and to my trials with resistors, the conclusion is, indeed, an awry PSU (Fortron BlueStorm).
(I was absent for a few weeks and I wanted to get a long term observation because of the intermittence of the previous failure ; that's why I didn't post before).
Thank you for your advices.
j-p