Overclock.net banner
1 - 20 of 25 Posts

crazymofo

· Registered
Joined
·
252 Posts
Discussion starter · #1 ·
Hi guys, I have a quick question to ask. Currently I am using silverstone strider plus 600w power supply. I want to make custom cables for it & I checked the stock wires are labeled as 16awg. Based on what I read cable sleeving for 18awg wires are easier than 16awg.

So my question is can I use 18awg instead of 16awg that the manufacturer used? Will this cause any problems? Because I plan to extend my 24pin atx cable to about 2.5ft-3ft, and the default length was 2ft. Some say we cant use higher awg cables for longer length otherwise this will cause the cable to melt.
 
no, dont.. You will have voltage drop. if you are going to extend it 2 feet then try thicker wire maybe 15 or 14AWG.

the longer the distance the more the voltage drop, I once shifted my PSU and did some cable management and extended my 18AWG 12" PCI Express 6 pin + 8pin power cables to around 25". And after a few mins of gaming my GTX 480 would crash and the PC would reboot after a while.

use thicker wire on longer cable to eliminate voltage drop. am sure the resistance can cause the extra wire to get really hot and seriously damage your motherboard too. dont do it. maybe get a nice sleeved extension for a few buck or use thicker gauge wire.
smile.gif
 
Quote:
Originally Posted by TheBirdman74 View Post

I once shifted my PSU and did some cable management and extended my 18AWG 12" PCI Express 6 pin + 8pin power cables to around 25". And after a few mins of gaming my GTX 480 would crash and the PC would reboot after a while.
OP is talking about 3 feet total, not 25 feet.

Use a actual voltage drop calculator and you'll see that v-droop over 3 feet on AWG18 wire running 2 amps is less than 0.04 Volts. That's well within ATX spec and quite normal situation. For PCIe 8-pin cables that goes up to 4 amps, but still only 0.08V droop.
http://www.stealth316.com/2-wire-resistance.htm

Absolutely no chance of anything melting, and for this you can easily use 18AWG for all wires. Maybe above 12 feet you'll start to be able to measure significant v-droop and possibility of over heated wires, but for 3 feet it's no problem
smile.gif
 
Discussion starter · #6 ·
Quote:
Originally Posted by WiSK View Post

OP is talking about 3 feet total, not 25 feet.

Use a actual voltage drop calculator and you'll see that v-droop over 3 feet on AWG18 wire running 2 amps is less than 0.04 Volts. That's well within ATX spec and quite normal situation. For PCIe 8-pin cables that goes up to 4 amps, but still only 0.08V droop.
http://www.stealth316.com/2-wire-resistance.htm

Absolutely no chance of anything melting, and for this you can easily use 18AWG for all wires. Maybe above 12 feet you'll start to be able to measure significant v-droop and possibility of over heated wires, but for 3 feet it's no problem
smile.gif
Hi I am back again, just another quick question. What if I reduce the cable length for like molex cables? Increasing the cable length causes voltage drop so when I decrease the length I assume the voltage will increase? If I do this, it will burn my components due to the high voltage?
 
Quote:
Originally Posted by crazymofo View Post

Hi I am back again, just another quick question. What if I reduce the cable length for like molex cables? Increasing the cable length causes voltage drop so when I decrease the length I assume the voltage will increase? If I do this, it will burn my components due to the high voltage?
No. The way it works is that your PSU has +12V on the yellow wire going to the component, and +0V on the black wire coming back. If your wires are zero length then your component will get exactly 12V. As your wires get longer, some of the voltage is lost in the wire itself, due to electrical resistance. So your components will never get 'too much' voltage.
 
Discussion starter · #8 ·
Quote:
Originally Posted by WiSK View Post

No. The way it works is that your PSU has +12V on the yellow wire going to the component, and +0V on the black wire coming back. If your wires are zero length then your component will get exactly 12V. As your wires get longer, some of the voltage is lost in the wire itself, due to electrical resistance. So your components will never get 'too much' voltage.
I see, ok I get it...so only less voltage is possible if the wires are extremely long. Thanks for the reply
smile.gif
 
Quote:
Originally Posted by WiSK View Post

OP is talking about 3 feet total, not 25 feet.

Use a actual voltage drop calculator and you'll see that v-droop over 3 feet on AWG18 wire running 2 amps is less than 0.04 Volts. That's well within ATX spec and quite normal situation. For PCIe 8-pin cables that goes up to 4 amps, but still only 0.08V droop.
http://www.stealth316.com/2-wire-resistance.htm

Absolutely no chance of anything melting, and for this you can easily use 18AWG for all wires. Maybe above 12 feet you'll start to be able to measure significant v-droop and possibility of over heated wires, but for 3 feet it's no problem
smile.gif
I know, Thats what I told him 25 inches = 25".
wth.gif

Do you know the difference between Inches " and feet ' ??
lachen.gif


OP Use Higher gauge Wire, 14 AWG or 16 AWG. take it from me, 3 Feet or 36" You are going to have voltage drop, just a spike is enough to crash, damage the components. it may cost a few dollars more or be a bit more time consuming but thicker wire is better for length.

Look, ATX specs dont matter, Your PSU Is not going to be 100% efficient or accurate but it was Made to deliver X amount of Power using X Thickness Gauge Wire up to X length properly. When you alter, extend it then there will be voltage drops.
 
Quote:
Originally Posted by TheBirdman74 View Post

I once shifted my PSU and did some cable management and extended my 18AWG 12" PCI Express 6 pin + 8pin power cables to around 25". And after a few mins of gaming my GTX 480 would crash and the PC would reboot after a while.
Quote:
Originally Posted by TheBirdman74 View Post

OP Use Higher gauge Wire, 14 AWG or 16 AWG. take it from me, 3 Feet or 36" You are going to have voltage drop, just a spike is enough to crash, damage the components. it may cost a few dollars more or be a bit more time consuming but thicker wire is better for length. Look, ATX specs dont matter, Your PSU Is not going to be 100% efficient or accurate but it was Made to deliver X amount of Power using X Thickness Gauge Wire up to X length properly. When you alter, extend it then there will be voltage drops.
If you had some problems yourself in the past making a new harness, then did you take any measurements confirming your suspicion that voltage drop was actually the problem? If so, how much voltage drop did you measure? Was it consistent across all three 12V wires?

And how did you extend the harness exactly? Did you cut the wires in the middle and solder in between? Did you check for cold solder joins?

When you changed your wires for 14 AWG, how did you crimp on the new pins? Which crimper did you use that can do 14 gauge?

Also, which sleeve would you recommend for these larger wires?

The fact is that voltage drop in a 18 AWG wire running 2A @ 12V is less than 0.04V. That means it's less than the ripple you'd expect the PSU to generate anyway.
 
Ok...im goin to share my idea here about reducing the Diameter of the cables..... The theory is that if your goin to reduce the diameter then the current passing through it will reduce also..... from the formula V=IR (V= Voltage .....I (current)Amps.....R (resistance) Ohms) ....if the current rating capacity of the wire rated for 2amps for example and your goin to reduce the wire diameter,the 2amps will reduce and base on the formula V=IR the result of the V(Voltage) will reduce also and thats the thing where the voltage drop occur....... and yes if the distance is too much long there is also voltage drop due to wire resistance....
wink.gif
 
Quote:
Originally Posted by xd9denz View Post

if the current rating capacity of the wire rated for 2amps for example and your goin to reduce the wire diameter,the 2amps will reduce and base on the formula V=IR the result of the V(Voltage) will reduce also and thats the thing where the voltage drop occur
If you reduce the wire diameter, it's not the 2 amps that reduces.
It's R, which can be calculated for copper stranded wire as follows.
R= 1.7 x length / cross sectional area (in meters and meters squared respectively)
So if going from 18 gauge (0.823 mm2) to 14 gauge (2.081 mm2) as Birdman74 suggests, then you will see a decrease in R of approx factor 2.5.
 
Quote:
Originally Posted by WiSK View Post

If you reduce the wire diameter, it's not the 2 amps that reduces.
It's R, which can be calculated for copper stranded wire as follows.
R= 1.7 x length / cross sectional area (in meters and meters squared respectively)
So if going from 18 gauge (0.823 mm2) to 14 gauge (2.081 mm2) as Birdman74 suggests, then you will see a decrease in R of approx factor 2.5.
it supports only my explanation......why? if the resistance decrease the more that the voltage drop is eliminated,because if the resistance is increasing then the more it resist the flow of current(amperes).....
thumb.gif
 
WISK is right on the 16ga (lower "ga" numbers are heavier and have less resistance, hence lower voltage drops). At PC cable lengths and power, 16ga vs 18ga is a non-issue.

In TheBirdman74's the more likely culprit is connectors. The cheap tin-lead connectors used on most cable connectors have quite a high resistance. They are also VERY poor contacts with the split pins loose when new and they quickly get looser with use and age. I've had motherboards overheat, blacken, and melt the connector for the 4 pin motherboard connector. Only reason, loooooose fitting pins!!!

(Example: ever notice how Hot the A/C plug gets on something like your vacuum cleaner, but ONLY when the A/C wall outlet is old and loose? Same issue, loose contact equals poor contact equals high resistance and thus higher heat and bigger voltage drops.
 
Quote:
Originally Posted by WiSK View Post

If you had some problems yourself in the past making a new harness, then did you take any measurements confirming your suspicion that voltage drop was actually the problem? If so, how much voltage drop did you measure? Was it consistent across all three 12V wires?

And how did you extend the harness exactly? Did you cut the wires in the middle and solder in between? Did you check for cold solder joins?

When you changed your wires for 14 AWG, how did you crimp on the new pins? Which crimper did you use that can do 14 gauge?

Also, which sleeve would you recommend for these larger wires?

The fact is that voltage drop in a 18 AWG wire running 2A @ 12V is less than 0.04V. That means it's less than the ripple you'd expect the PSU to generate anyway.
1. Did you originally see you mistook my Inches " for Feet ' ? 25" is not 25'
biggrin.gif


2. Yes I used the volt meter and it was consistent and all fine untill it would start heavy loading on the GTX480, Which you know can consume upwards of 300W + when Overclocked.

3. So What would happen is after 5 Mina of heavy load on a game or say Furmark the Computer would crash.

4. I extended it using soldering, First I cut them in between both the PSU and the connector arund 50% and added the extra length of wire and extended it. There were no cold soldering joints.

5. Replacing the Cable to stock length or replacing the PSU solver the problem.

6. For Sleaving first you heat shrink the joints and then slide paracord 450 or 550 over it, If the cable is thicker then 550 paracord you use shoe laces which look even more nice..You can get great leangths of shoelaces off ebay for a few bucks in various colors along with 100 foot of para cord for a few bucks shipped. cant beat paracord
thumb.gif
 
Quote:
Originally Posted by RedModRed View Post

WISK is right on the 16ga (lower "ga" numbers are heavier and have less resistance, hence lower voltage drops). At PC cable lengths and power, 16ga vs 18ga is a non-issue.

In TheBirdman74's the more likely culprit is connectors. The cheap tin-lead connectors used on most cable connectors have quite a high resistance. They are also VERY poor contacts with the split pins loose when new and they quickly get looser with use and age. I've had motherboards overheat, blacken, and melt the connector for the 4 pin motherboard connector. Only reason, loooooose fitting pins!!!

(Example: ever notice how Hot the A/C plug gets on something like your vacuum cleaner, but ONLY when the A/C wall outlet is old and loose? Same issue, loose contact equals poor contact equals high resistance and thus higher heat and bigger voltage drops.
I agree...Lose + Resistance + Thin Wire = Disaster waiting to happen..
Once I melted a motherboard 24pin extender connector with a molex next to it, it went black...I didin't know till I one day touched the right side of the casing (where the connectors and wires were hidden behind) and it was hot as hell. So I removed and found it melted and also part of a molex connector next to it was charcoal. It was a cheap home made extender, that was a few years ago.
tongue.gif
 
Quote:
Originally Posted by TheBirdman74 View Post

Yes I used the volt meter and it was consistent and all fine untill it would start heavy loading on the GTX480, ... the Computer would crash
So you're saying now that there was not any voltage drop at idle? That is odd, because the 480 draws over 50W at idle. For a crash to be caused by voltage drop at full current, it should be measureable at 50W. I still haven't seen any numbers from you. And that's why I find it difficult to believe the story. I mean, I believe your computer crashed, but not that it was caused by voltage drop. Expected ripple over the PCIe wires would be more significant than internal resistance of a 2 foot cable.
Quote:
Originally Posted by TheBirdman74 View Post

Replacing the Cable to stock length or replacing the PSU solver the problem.
So your advice was to the OP to increase wire size, but now it seems you never tried this for yourself, and actually it was the removal of your DIY soldering job which solved your crashes. Right?

Why I keep replying to you, is that for sleeving AWG14 brings more trouble and no benefits.

It's hard to crimp onto MiniFitJr pins, and if you solder it into the middle of an AWG 18 wire then you're actually increasing resistance not decreasing. As soon as you put sleeve around it then it will hardly fit into the connectors - biggest I've tried with sleeve is 2.5mm diameter wire. So you'd have to use wire with very thin insulation to get it and the sleeve into the connector. You say it fits in paracord 550, but does that then fit into a connector? Or if not heatshrinkless, how chunky is it with heatshrink around it as well? I'd like to see some pics before I believe it's a good idea.
 
Quote:
Originally Posted by WiSK View Post

So you're saying now that there was not any voltage drop at idle? That is odd, because the 480 draws over 50W at idle. For a crash to be caused by voltage drop at full current, it should be measureable at 50W. I still haven't seen any numbers from you. And that's why I find it difficult to believe the story. I mean, I believe your computer crashed, but not that it was caused by voltage drop. Expected ripple over the PCIe wires would be more significant than internal resistance of a 2 foot cable.
So your advice was to the OP to increase wire size, but now it seems you never tried this for yourself, and actually it was the removal of your DIY soldering job which solved your crashes. Right?

Why I keep replying to you, is that for sleeving AWG14 brings more trouble and no benefits.

It's hard to crimp onto MiniFitJr pins, and if you solder it into the middle of an AWG 18 wire then you're actually increasing resistance not decreasing. As soon as you put sleeve around it then it will hardly fit into the connectors - biggest I've tried with sleeve is 2.5mm diameter wire. So you'd have to use wire with very thin insulation to get it and the sleeve into the connector. You say it fits in paracord 550, but does that then fit into a connector? Or if not heatshrinkless, how chunky is it with heatshrink around it as well? I'd like to see some pics before I believe it's a good idea.
So now its a story, I come here and I make up a PSU story. got it.
rolleyes.gif


I could send you pictures you know, I probably still have the original piece of cable that I soldered.

Removing my wiring and Soldering it back to the stock wires length fixed the issue of the GTX480's crashing. It would be fine idle, you could run the computer for as long as you wanted but as soon as I started say Metro or Mafia 2, under say 5 - 10 min's it would usually crash. But I have run it as far as 30 mins and then it would crash, which leads me to think a volt drop at a time. But I dont know, I just ended up cutting the extra length of wire and re soldered back to stock lengths. I dont think it was my soldering, since if that was the cause then the new soldered wire would also fail. but yes surely there was resistance when you solder something in between, .

Paracord is the best stuff, everybody uses it, Look it up on OCN or on HeadFi. Its Cheap Looks Great, and easy to work with, does't usually need heatshrink but I would imagine if you were to use it you could just cut it about a cm before the connector and then add the heatshrink. I dont know. Just my opinion. You want Op to go with 18AWG to extend something 3 foot then do it. I just suggest he used thicker wire. I personally would not even consider anything below 16AWG but that just my opinion. do what you want.
 
1 - 20 of 25 Posts