Overclocking Sandbox: Tbred B DLT3C 1700+ and Beyond

Audioaficionado said:

...
BTW what was your stable limit at stock voltages?

Click to expand...

At default Vcore 1.5 V of the TB B 1700+ DLT3C, actual Vcore measured by ASUS proble is 1.52 V, the CPU can run stable at least at 2204 MHz (200x11.5), but fails at 2300 MHz. I did not try to find out exactly where it fails between 2200 and 2300 MHz. The temperature is really nice, 34 C prime95 loaded with a fan speed of 3000 rpm, system temp 21 C.


Prime95 at 2204 MHz, at 1.5 V stock Vcore setting, 34 C loaded

It looks like your chip is a better overclocker than mine and I consider mine a pretty good one. Take care of that guy as I think it's a cherry among cherries.

I've noticed that mine doesn't go up in temp by running it up at the stock voltage. The delta was +13c at stock running <5% capacity. At full O/C while running FAH =100% I get a delta of +16c. I'm not sure how accrate the under chip temp sensor diode is. I think I'm running below the average 45-50c it indicates.

hitechjb1 said:

I agree operating the CPU like that at such high Vcore (30% above nominal), with high fan noise (for air cool) for 100-200 MHz gain is not totally justify for price-performance reason, for the long term "health" of the CPU (electro-migration, leakage current and heat stress), and enviornmental reasons (such as noise, heat), ....

It is for learning overclocking technique and getting some good bencmark and satisfaction (bragging as you said).

But actully, I learn quite a bit with this hand on experience of the low Vcore rate chip, and its high current behaviour (its rated current at 1.5V is higher than the 1.6V 1700+ by 7%). I am amazed by this chip from day one, and try to explain why and how it works so fast with smaller voltage (but it turns out that it will max out sonner than the other). I posted some explanation of it in another thread: my theory on the new dlt3c's

This is the first time I did overclocking of a CPU at this aggressive level. Normally I just oc 10-20% even just using stock HSF. I learn many things such as how clock frequency, die temperature, cooling, fan speed and Vcore vary, looking into heat sink and fan, PSU characteristics and picking, ... how heat, die temp changes w/ Vcore, ..., and many many things and from this forum.


To the end, I think it is not totally justify to stress the CPU like this for the last 100 - 200 MHz since I have to run at a Vcore 1.95 V (30% above nominal). It is better off just to run it at 1.65V (10% above nominal) at 2400 MHz, which is the best, long term operating point of this particular CPU. From 2400 MHz to the current 2550 MHz, it is very expensive in terms of price-performance, noise, voltage and heat control.

When this is done, the CPU will most likely go back to a more nature frequency and settings.

Click to expand...

Yeah, thats what I'm tying to do. This processor has a "knee" in the curve, after which it overvolting doesn't help. It's a great chip,
a ringer, but like all of us, it has it's limits.......

sanford1 said:

Yeah, thats what I'm tying to do. This processor has a "knee" in the curve, after which it overvolting doesn't help. It's a great chip,
a ringer, but like all of us, it has it's limits.......

Click to expand...

At that point, Vcore increase will produce more leakage current and heat, instead of going into active power (CV^2f) for running faster clock. The chip may not response to higher Vcore. Assume the PSU current sourcing is not the cause.

At that point, only thing that can be done is to change to a more aggressive cooling to lower the die temperature and leakage current. Apart from price performance, super cooling would be able to extend the overclocking for a few hundred more MHz.

Explanation for last two posts:

For my 1700+, I can run it at:
2.2 GHz stable using 1.5V, 34 C load, 3000 rpm fan
2.4 GHz stable using 1.65V, 38 C load, 3000 rpm fan
2.5 GHz stable using 1.85V, 48 C load, 4800 rpm fan

For reasons of CPU health, environment reason of heat and noise (air cooling), it is obvious to run the CPU at 2.4 GHz at 1.65V, 38 C load, 3000 rpm fan, or even at 2.2 GHz 1.5V, ...

The 0.2V Vcore to get the CPU from 2.4 GHz to 2.5 GHz, are mostly wasted as heat instead of useful power for computation !!!!


Relationship of clock, die temperature and Vcore
What is the active power of a CPU at frequency f and voltage V
How to estimate CPU static and active power
(page 2)

==================================================================

It can boot into Windows XP at 2600+. Not stable.

Working on it to make it stable. Probably it has reached its limit on air, unless changing to more aggressive cooling.

Any suggestions.

Current HSF: SK-7, TT SKII
PSU: Antec TP 430

WCPUID



Asus Probe

where did you order this chip/fan from? excalibur?

thanks

jhbball said:

where did you order this chip/fan from? excalibur?

thanks

Click to expand...

I got the retail version of the CPU from a local computer show. It is 0310 XPMW JIUHB, paid $64. It came with a HSF and 3 year warranty.

The fan is a ThermalTake Smart Fan II, whose fan speed can be adjusted for very quiet operation (spec 17 dbA 1300 rpm 20.55 CFM) or very high CFM (spec 48 dbA 4800 rpm 75.7 CFM), or somewhere in between. I find that 2500-3500 rpm give very good operating point for low fan noise and die temperature cooling. Any popular cooling online store such as sidewinder, svc, ... should carry it, except the shipping cost is very high in term of % when just ordering a fan.

The HS is a SK7 which is around $20 now. It is economical compared to the SK800U, which should be the best one to get for air cooling. The SK800U now is around $38, and would lower the die temperature by 1-3 C more due to its lower C/W (die-ambient temp difference per thermal power), 0.23 compared to 0.25 for SK7. The thermal power for this chip is around 50W (default 1.5V, 1.5 GHz) and can be way above 130W during max oc (1.9+ V 2.5 GHz).

I use the SK7 and SF II as well. great combo I must say.

Where is this "local computer show"? I wish I would know if one pops up around here where I live lol...

FLyingHamster said:

I use the SK7 and SF II as well. great combo I must say.

Where is this "local computer show"? I wish I would know if one pops up around here where I live lol...

Click to expand...

Yes, either the SK7 or the SK800 is good. SK7 is really good for the money, without it the 1700+ will not make it to 2.4-2.5 GHz.

BTW, you have nice OC setup. I find at around 1.8V, 40C loaded and 2.4+ GHz is the best natural, operating point for the 1700+. The fan can be around 3000-3500 rpm, not excessive noise.

Can you get your FSB higher, is the mb limiting, since the TwinMOS 3200 should not be a problem.

How about the PSU, does it limit the CPU going higher?

Vcore mod for higher Vcore (using fix resistor)

I was asked about getting higher Vcore than that provided by the bios.

Without doing the Vcore mod, the highest default Vcore is 1.75 V for 1700+ (from the A7N8X-dlx). I heard for the ABIT NF7-S Rev 2, one can set up to 2.2 V in the bios ???. A7N8X-dlx is not very friendly in this regard, but it is much safer as a general purpose mb.

This is what I do:

1. Locate the VRM 5-bit dual phase controller integrated circuit (IC) L6917BD for Vcore, which is a long 28 pins IC located on the mb between the CPU socket and the pink/purple color parallel printer port for the A7N8X-dlx. I think both the ASUS and EPOX nforce2 mb use this dual phase controller, ABIT uses 3 phase controller for Vcore, so it's different.

2. I add a fix resistor between pin 9 of that IC to GND. That is, connecting a fix resistor between pin 9 and GND.

I use a small grabbler (such as SMD) to grab pin 9, and then connect the other end of the grabbler wire to a fix resistor. The other end of the fix resistor is then connected to GND. Luckily, the pin for this IC is much bigger than the pin for doing VDD chipset mod. So it is much easier to do Vcore mod.

If you are not sure how to locate GND, ask to make sure or don't do it until you are 100% sure and understand. One can use the case metal as GND. OR it can be the pin that connect to the BLACK wire of one of the fan headers on the motherboard. One can connect to GND using a crocodile clip temporarily or by soldering.

3. Where is pin 9?

The dot on the IC is pin 1, viewing the mb from the top. Locate pin 9 by counting from pin 1. That is, next to pin 1 is pin 2, next to pin 2 is pin 3, etc, etc.

14 13 12 11 10 9 8 7 6 5 4 3 2 1

4. I prefer to use a fix resistor than a variable resistor (VR) which many ppl are using for oc mod as seen on many forums. I avoid using VR because varibale resistors are more sensitive to temperature, vibration, or accidential change/mistake (in case a little brother/sister turns that variable resistor, and ends up with large Vcore and fry the CPU and mb !!!). After fixing on a resistance value, I can make the mod safe proof, i.e. any bios setting will be within safety margin.

5. What resistance is the fix resistor?

I find the "good" resistance value by experiment. In the case for the A7N8X-dlx, I find that for fix resistor of

- 20K Ohm, it adds about 0.1V to the bios Vcore setting
- 10K Ohm, it adds about 0.2V to the bios Vcore setting (getting Vcore between 1.75 - 1.95 V with bios setting between 1.5 - 1.75 V.)
- 8K Ohm, it adds about 0.25V to the bios Vcore setting (this is the one I use, getting Vcore between 1.8 - 2.0 V, with bios setting between 1.5 - 1.75 V)
- 5K Ohm, it adds about 0.4V to the bios Vcore setting (AVOID this one or use this one with EXTRA caution. If bios is at 1.75V, actual Vcore would be 2.15V, may damage CPU !!!!)

6. Higher resistance adds less to the Vcore from the bios setting. A small watt (such as 1/4 W or 1/2 W) resistor is enough, since it is not large current carrying. You can customize the value of the resistor for your own need.

I don't know whether this resistance value work unverisally and how accurate it is with other A7N8X-dlx (since I tried only one), you may have to try it out. But it should give enough guideline, and may be accurate too!!!

7. Also by doing this, it solves the cold boot problem.

8. Testing: After adding the fix resistor, boot into bios and look at the Vcore. If the Vcore is way off, turn off the system and check. E.g. if adding 0.2V is expected, and bios Vcore is set to 1.6V, the bios Vcore measurement should be around 1.8V +- 25 mV.

As usual, only do this with full understanding and care, only do things that make sense. If not clear, ask. And doing this may void the warranty, and is at your own risk. It is for experiment only, to better understanding oc. I am not responsible for damage of any kind.

Wow hitechjb, I am impressed. Good work, nice to see al those pics and such, keep up the good work .

Picked up another Tbred B 1700+ DLT3C, WPMW JIUHB

It does
2.0+ GHz at rated 1.5 V
2.4 GHz at 1.725 V, 200x12
2.5 GHz at 1.875 V, 208x12, 48 C loaded (fan 4800 rpm)

Prime95, 3DMark stable

Still under testing.

Tbred B 1700+ DLT3C JIUHB WPMW 2505 MHz 1.875V

nice overclock. I jsut got some questions as to those temps.

are u using arctic silver 3 for your SK7 and SF II? how many case fans do u have? anything else that would contribute to your low temps?

I get about the same temps, but mines at 2.44ghz..

I'm using ASII on NB & CPU. According to my diode temps, I'm at 35c system and 51c CPU. My room thermometer is 24c. My air flow of four 80mm intake, two 120mm intake blowing on the CPU & PCI cards, three 80mm exhaust fans (one's a blowhole), and an 80mm PS exhaust fans is adequate. The delta for the O/C forum test of my Vantek CPU cooler on a mobo was 18c. I think my readouts are too high and several degrees cooler than indicated.

Edit: The new Abit NF7-S v2.0 beta 14 bios fixed the problem.

Temps are now amb25c/sys36c/cpu44c full load under FAH 24/7.

Air Cooling Setup

For CPU air cooling

AS3 w/ a SK7 and a TT SFII are used.

Sometime a Vantec Tornado is used for testing oc. The fan running 5700 rpm pushing 84.1 CFM can lower the die temp by as much as 5C loaded. But it cannot be used regularly due to excessive fan noise, unless a 10 Ohm VR is added to adjust its speed.

Will soon switch to a SLK-800U for further testing. The SLK-800U should lower the die temperature by 2-3 C at heavy load. The SLK-800U has a thermal resistance 0.23 C/W, whereas the SK7 being 0.25 C/W. E.g. for cooling a 100W CPU, the temp would be 2C lower.

With the SLK-800U and Vantec Tornado, I estimate it can lower the die temp by 5-8 C under heavy load, and would bring the die temp under 40 C, which is an important temperature to keep the CPU below. I find that the 1700+ DLT3C begins to increase its leakage current drastically around 40 C.

For case cooling:

Three/four case fans are used, plus the PSU fans.

1. There is a 120 mm fan about 64 CFM in the front for intake to push air to cool the motherboard and the video card GPU side, to lower the system temperature. (80/90 mm fan is OK if 120 mm does not fit).

2. There is a 80 mm fan about 32 CFM on the side panel for air intake to the CPU area.

3. One/two 80 mm fan about 40 CFM at the back of the case for outtake.

4. Also the PSU has two fans outtaking air from inside the case to exterior through the back fan.

Each case fan has noise level around 32 dBA or below.


Temperature and fan speed for daily usage

Objective is to keep CPU temp under 40C and fan around 3200-3500 CFM for continuous regular usage. This is much more enviornment friendly than trying to run and stress the last 100 MHz out of the CPU.

At this level of die temp and fan speed and fan noise, the 1700+ DLT3C should still be able to deliver 2.4+ GHz.

E.g. 2.4 GHz stable using 1.65V, 38 C load, 3000 rpm fan

I got two 0310 1700+ DLT3C, one XPMW, one WPMW, they are able to meet the objective of 2.4 GHz stable, under 40 C, 1.7 V, 3200 rpm. They both can be oc'ed to above 2.5+ GHz stable, but with higher Vcore (~1.9V), temp and fan rpm.

Latest update:
Summary for overclocking CPU and FSB (page 3)


This is summary of what I learned and used in overclocking a Tbred B 1700+ DLT3C click link. Probably most of you already know about these techniques. If you think there are something missing or incomplete or incorrect, please post. But I think this may help some new comers to get start, ...

did u write all that? or did u copy & paste it from somwhere?

FLyingHamster said:

did u write all that? or did u copy & paste it from somwhere?

Click to expand...

I wrote all the posts from scratch myself, based on my understanding of math, science and engineering. I hope it can help others and myself to understand better. If you find a paragraph or sentences identical else where, please let me know. Some ideas are original, and many are existing knowledge. I try to do it as correct and precise as possible, so from time to time, I may edit the posts for clarity and correctness. So if you find anything incorrect, incomplete, questions or comments, please post.

If I have to quote, I'll give the sources and references as done in good publications. It would violate some law and standard if I just do ctrl-C and ctrl-V without references and proper credit.

If they don't make this a sticky, I'll have to keep bumping it every so often.