May 23, 201610 yr Author Hmmm... It seems to me that the peak ampeages would have to be considered surely?
May 23, 201610 yr Oh ok. Never quite realized that the cables didn't also have to be able to supply that current. Sent from my LG-D852 using Tapatalk
May 23, 201610 yr Oh ok. Never quite realized that the cables didn't also have to be able to supply that current. Oh, they do have to supply the current, and it will cause a momentary sag in voltage when they are called upon to power the surge of inrush current. Too much voltage sag, and the drive will not respond properly because its logic circuits no longer have enough juice to properly differentiate between ones and zeros. However... the amps per gauge per length per temperature is all about sustained heat transfer. It's a measure of how much current can be supplied over time without causing the wires to overheat. The momentary supply of power causes a small spike in temperature, but as long as it's not sustained at that level, the wire has a chance to cool back down before any damage is done. The biggest issue facing too small of a wire or bad connections is the voltage sag, not burning up the wires.
May 24, 201610 yr Oh ok. Never quite realized that the cables didn't also have to be able to supply that current. Oh, they do have to supply the current, and it will cause a momentary sag in voltage when they are called upon to power the surge of inrush current. Too much voltage sag, and the drive will not respond properly because its logic circuits no longer have enough juice to properly differentiate between ones and zeros. However... the amps per gauge per length per temperature is all about sustained heat transfer. It's a measure of how much current can be supplied over time without causing the wires to overheat. The momentary supply of power causes a small spike in temperature, but as long as it's not sustained at that level, the wire has a chance to cool back down before any damage is done. The biggest issue facing too small of a wire or bad connections is the voltage sag, not burning up the wires. was being sarcastic. My point being that why deliberately design a system that you know is going to at times place more current along the lines than they are rated to carry. The momentary heat spikes will cause the insulation to degrade over time and become brittle. If nothing else, this little debate is really bringing to the forefront the most under appreciated aspect of any PC - power delivery - that is arguably the most important aspect of any PC
May 24, 201610 yr My point being that why deliberately design a system that you know is going to at times place more current along the lines than they are rated to carry. The momentary heat spikes will cause the insulation to degrade over time and become brittle. If nothing else, this little debate is really bringing to the forefront the most under appreciated aspect of any PC - power delivery - that is arguably the most important aspect of any PC Fully agree. We spend so much time talking about data integrity and minimizing data loss when at the core of it, data is the difference between voltage being on and off, and if you compromise any part of the system's ability to differentiate between two absurdly close voltage levels by failing to deliver clean constant power it makes the whole exercise pointless. Even worse, you can physically destroy components by using too small of a supply that is poorly regulated. A properly spec'ed server system will never come close to pushing the limits of the power supply system, from supply to connectors. Data is too precious to scrimp on marginal supply and transmission gear.
May 24, 201610 yr My point being that why deliberately design a system that you know is going to at times place more current along the lines than they are rated to carry. The momentary heat spikes will cause the insulation to degrade over time and become brittle. If nothing else, this little debate is really bringing to the forefront the most under appreciated aspect of any PC - power delivery - that is arguably the most important aspect of any PC Fully agree. We spend so much time talking about data integrity and minimizing data loss when at the core of it, data is the difference between voltage being on and off, and if you compromise any part of the system's ability to differentiate between two absurdly close voltage levels by failing to deliver clean constant power it makes the whole exercise pointless. Even worse, you can physically destroy components by using too small of a supply that is poorly regulated. A properly spec'ed server system will never come close to pushing the limits of the power supply system, from supply to connectors. Data is too precious to scrimp on marginal supply and transmission gear. But the reverse is also true. By very carefully spec'ing the system you can get away with a lot in terms of the power system. EG: I've got a 20 bay 3U hotswap server sitting here doing nothing that has a 530W redundant P/S in it (35A 12v) This is absolutely 100% capable of running fully populated all 20 drives (10,000 enterprise) because the average power draw per drive is going to be less than 1A. But, it can only do that under it's spec'd out components of hardware RAID, drives always spinning, and (this is the big one) staggered spin up when turning on. Under any other circumstances, that p/s is woefully underpowered.
May 24, 201610 yr Oh ok. Never quite realized that the cables didn't also have to be able to supply that current. Sent from my LG-D852 using Tapatalk They need to be able to supply the maximum/average power of the drives, but NOT the startup current, since that is so small of a time (only a few ms). To give an example outside of the pc world, then you choose a cable for a motor, you use the current the motor will run at at maximum, and NOT the startup power. So if a motor has a current draw of 50A, but has a startup current of 200A, you use a cable that is able to supply at minimum 50A, and NOT 200A..
Archived
This topic is now archived and is closed to further replies.