GF4 Ti4600 and CUSL2 problems - major glitches+tears!!

[raster]

Thanks

Yea, this is definately not a agp 2x/4x issue- i was able to enable 4x agp with my previous 1008 001 tweaked bios.

It is something else, very unusual as ppl with problems recorded limited success by repositioning their cards in the slot (see above links)

However, with the bios update, this "carefull positioning" becomes unneccessary. Wierd, but fact.

[JonP]

No disrespect to anyone, but I see time and time again people being told to buy huge power supplies every time they have a problem with a Geforce card. I wouldn't have an issue with this except that to go out and replace a power supply as an attempt to cure a problem is probably the most expensive attempted cure I can think of. There are lots of other things to try first that cost nothing at all. Power supplies, as we all know, are very expensive for the good ones, and even average, ubiquitous ones of around 300 - 350 watt can cost a sizeable sum.

The only time I have even found it neccessary to use even a 300 watt power supply is when I first used an Athlon XP 1800+ in conjuction with a Geforce 4. Certainly in this case the machine would not even boot with a 250 watt power supply. With my ASUS CUSL2-C and Geforce 4, I can also quite happily run this machine, with quite a few fans, 3 of the four IDE channels populated, 4 USB ports occupied, with a Geforce 4 Ti4400, all with a standard 250 watt Aopen power supply and with 100% total stability in the most intensive 3D applications.

Can I please humbly suggest that before you go out and buy that 5000 terrawatt power supply that your best mate / dog / girlfriend / grandmother / goldfish recommended, that you actually sit down and CALCULATE the ACTUAL current that your machine is actually going to draw. From that, you can calculate the required wattage for each of the power rails. Then it just a matter of checking to see what power supply fits the bill. You will actually find that in many cases the power requirements are less than you think once you have quantified them on a device by device basis.

Sorry for the rant, but if I had a dollar for all the posts where I see someone has just upgraded their power supply and it hasn't solved their problem, well I really almost could retire.

If you are not sure how to calculate your PS requirements, AMD has a lot of useful information on their website and I'm sure there is other stuff available as well.

Yes, certainly there are machines that do need very powerful supplies, but in many cases it is simply not neccessary. The trick is to look at the individual outputs of each of the rails in a power supply, not the total wattage. If you really think your power supply is the cause of problems, you can at least just try your machine with minimal components connected and perhaps even patch in two supplies to do some testing. If you find removing components or using two power supplies solves your problem, then it would be prudent to look at beefing up your power supply.

[jomiwic_abx]

Originally posted at the http://ECS PSU FAQ

OUTPUT VOLTAGES MINIMUM LOAD MAXIMUM LOAD (amps) LOAD REGULATION LINE REGULATION
........+3.3V....................0.2.....................10.6/28..................+/-5%....................1%
........+5.0v.......................0.....................40/28.5..................+/-5%....................1%
........+12.0v.....................0.........................12......................+/-5%.....................1%


Note: The total combine power of the +3.3V and +5.0V is 235W.

When +3.3V is loaded to 10.6A then +5.0V output limit is 40A

When +3.3V is loaded to 28A then +5.0V output limit is 28.5A

ATX 2.01 specs require 5% regulation on the 3.3 and 5.0 volt rails.

One important thing to note on the above table is the Total Combine Output (TCO) of the two rails, which is equal to 235W. Most quality made PSU will have this note. This is an indication a comprehensive testing was conducted in the design of the PSU. If the TCO is not written on the rating label of the PSU skip on it. Some PSU will not list the two number limits on the total amps that can be carried by each of the rails instead they will just place the lower numbers from the +3.3V and +5.5V rail outputs. So it’s not unusual to see the above example to have the 10.6A and 40A as the respective maximum output rating of the +3.3V and +5.0V rails with an included note on the total combine output rating of the two rails. But new marketing gimmick uses the high ratings of the 3.3V and 5.0V rails. You can do the math on this one and most likely the total will exceed the total combine rating of the rails.

Remember the multiple wires mentioned earlier for outputs? Well, this is the trick used to be able to harness all the power you can get from the PSU without the use of gigantic wires. Most PSU’s uses the 16AWG and 18AWG wires which are rated at 11A and 9A respectively. With the use of multiple wires you can reduce overloading a single wire thus prevent overheating and reducing power losses.

Efficiency of PSU’s ranges from the low 50% to the high of 75%. Efficiency is the ratio of the power taken by the PSU from the wall outlet and the power taken on its load side.

Originally posted by JonP
Sorry for the rant, but if I had a dollar for all the posts where I see someone has just upgraded their power supply and it hasn't solved their problem, well I really almost could retire.

and if I had a dollar for everytime switching to a better PSU fixed some little prob I could retire. no, it is not the solution for every problem, unless of course, it is the problem. it seems an efficient and quality made PSU is preferrable over one that is lacking in these areas but having higher wattage. trying to explain that to everyone would require a longer rant than both yours and mine combined...

power supply problems have been becoming more prevelant lately:
http://www.overclockers.com/tips00120/




back on topic raster thanks for the BIOS link, I will have to update mine from the 1008 beta 1. I don't think I will be using Asus update though.

[JonP]

Jomiwic,

I don't appreciate the tone of your post. I don't understand the basis of your insinuation that my post was off topic. Had you read the thread from beginning to end, you would have realised that power supply issues formed a sizeable body of this thread. In this particular case, the original poster actually purchased a new power supply which did not resolve his issue.

The point of my post is that rather than fork out hard earned money on a new power supply, that the proper calculations first be made to determine whether a more powerful power supply is actually required. In no way am I implying that more powerful power supplies do not resolve issues - of course they do. But it remains an undeniable fact - proven once again within this very thread - that spending money on ever more powerful power supplies is often not neccessary.

I fail to see why my recommendation that people check whether they need a new power supply in the first instance is anything but sensible and topical advice.

[jomiwic_abx]

Originally posted by JonP
Jomiwic,

I don't appreciate the tone of your post. I don't understand the basis of your insinuation that my post was off topic. Had you read the thread from beginning to end, you would have realised that power supply issues formed a sizeable body of this thread. In this particular case, the original poster actually purchased a new power supply which did not resolve his issue.

The point of my post is that rather than fork out hard earned money on a new power supply, that the proper calculations first be made to determine whether a more powerful power supply is actually required. In no way am I implying that more powerful power supplies do not resolve issues - of course they do. But it remains an undeniable fact - proven once again within this very thread - that spending money on ever more powerful power supplies is often not neccessary.

I fail to see why my recommendation that people check whether they need a new power supply in the first instance is anything but sensible and topical advice.

re-read my post and you'll find I agreed with the majority of what you posted. the comment of going off topic was in refference to myself and not a comment on your post.
anything you want explained more thoroughly feel free to contact me by clicking here:

[JonP]

Fair enough As I was recently forcibly retired, I tend to get passionate about not spending money unless it is really neccessary

[raster]

I've been away for a while - but back for a week's holiday

Thanks all for your help. Very interesting PSU information there jomiwic, and although in my case it turned out *not* to be PSU related, I always thought there was more to PSU's than just their power rating (watts).

My system is, touch wood , still running perfectly with the Ti4600 using the 1012 tweaked bios. My nice new 400W twin-fan PSU is, although not absolutely neccessary, still a welcome addition to my rig.

Right, next project - to get a Celeron Tualatin going in the trusty CUSL2, where's that old udma66 cable.......

[JonP]

Raster,

Did you see the recent article at Tom's regarding the Upgradeware adaptors? If I were wanting to run a Tualatin in a CUSL2-C I would investigate them. I just wouldn't feel comfortable modding a CPU and possibly damaging it.

http://www6.tomshardware.com/cpu/02q3/020924/index.html

These adaptors also look more elegant than the Powerleap alternatives.

[SLM crew chief]

Originally posted by raster


Right, next project - to get a Celeron Tualatin going in the trusty CUSL2, where's that old udma66 cable.......

I have a 1.3g Celeron on my cusl2. I used to be able to run it at 1587 or something like that and it would boot and be stable at 1650 but I backed down because I was afraid of frying pci cards at that speed.

Now with my Geforce 4 TI 4200 and 1014.001 bios I am stuck at 1300 as I can't OC this card at all (see thread http://www.asusboards.com/forums/sho...threadid=25131 ) That is my problems.

Anyway the Celeron is a nice way to to go faster with this board. I replaced my 800eb and almost doubled my cpu speed.

Good luck and have fun. I broke the pins off (oem cpu so who cares?) which was nerve racking and it took some time to get the wire on but I installed it and never had a problem with it.