T
Thickbrit
Guest
Hey gang,ever wondered what a regas to r507 is worth ?,well.... i'm about to tell you.A few weeks ago i took a stock st unit and regassed it with R507.I tuned it to best cpu temps using a p4c800ed and modified 3.2E(as posted in another thread) running about 4.4ghz.It took much trial and error,but eventually i got it as cold as i could.Then i had some family problems,and haven't been able to test further until now.Unfortunately the p4c and 3.2E have since gone to a new home,so i needed a new test setup.I used an AS8 and 660(ex D.Barker) to compare a stock GT to the R507'd ST.First i needed an accurate way to measure cpu temps,as the AS8 can't do sub-zero and i wouldn't trust software anyway,so i came up with this....
I drilled a 1mm hole through the heatspreader using my trusty dremel,then gently pushed a highly accurate probe through the hole until i felt resistance as the probe touched the core,then locked the probe in place with a blob of sealstring.Next i tested the 660 on the stock Gt running a combination of prime95 and sp2004 to achieve 100% load on both logical cpu's at all times.It managed a very respectable 274mhz x18 (4.93ghz) at 1.5v,any more mhz 1 or other prime would fail,any more volts would overload the GT.Testing criteria was 100% load for i hour minimum.With ambient at 15c the Gt hit -14c(at the probe) after 1 hour.Next i swapped the Gt for the ST and performed the same tests running 274 x18 with 1.5v.After 1 hour the ST was doing -22.4c with ambient at 15.4c.I tried to keep my testing methods as scientifically accurate as possible and it was a lot more work than i've made it sound here.
So....in conclusion :-The St regassed to R507 was 8.4c colder than the stock Gt,even though the ambient was 0.4c higher for the ST.This extra cooling power meant i was able to increase the max 100% stable o/c on the 660 from 274(4.93) at 1.5v to 283(5.09) at 1.55v.
This regassed ST will be for sale very soon for around about the £250 mark,P.M. me if you are interested.
I drilled a 1mm hole through the heatspreader using my trusty dremel,then gently pushed a highly accurate probe through the hole until i felt resistance as the probe touched the core,then locked the probe in place with a blob of sealstring.Next i tested the 660 on the stock Gt running a combination of prime95 and sp2004 to achieve 100% load on both logical cpu's at all times.It managed a very respectable 274mhz x18 (4.93ghz) at 1.5v,any more mhz 1 or other prime would fail,any more volts would overload the GT.Testing criteria was 100% load for i hour minimum.With ambient at 15c the Gt hit -14c(at the probe) after 1 hour.Next i swapped the Gt for the ST and performed the same tests running 274 x18 with 1.5v.After 1 hour the ST was doing -22.4c with ambient at 15.4c.I tried to keep my testing methods as scientifically accurate as possible and it was a lot more work than i've made it sound here.
So....in conclusion :-The St regassed to R507 was 8.4c colder than the stock Gt,even though the ambient was 0.4c higher for the ST.This extra cooling power meant i was able to increase the max 100% stable o/c on the 660 from 274(4.93) at 1.5v to 283(5.09) at 1.55v.
This regassed ST will be for sale very soon for around about the £250 mark,P.M. me if you are interested.
