Chip of the year 2008

[everell]

Re: Chip of the year 2008

If anyone has the equipment to decap one of them we might be able to figure out who the manufacturer is.

I have a good guess it's Chipgoal (lol) CG8002...

Good joke! But if it would have been a match, the part number would have been CGZ8002...............

[Th3_uN1Qu3]

Re: Chip of the year 2008

If the table lists equivalents between the various makes (which i believe it does), then that's definitely not what it is. The SG6105 and WT7522 are the same thing, and they do not support current sensing.

The SG6105 is a 494 + 2x 431 + 6 preconfigured comparators (UVP/OVP for 3.3, 5 and 12v) + 2 opamps. It really doesn't look like what we have here.

[b700029]

Re: Chip of the year 2008

If anyone has the equipment to decap one of them we might be able to figure out who the manufacturer is.

I have a good guess it's Chipgoal (lol) CG8002...

[everell]

Re: Chip of the year 2008

AND, I have added a few more things to the 2008Z pwm chip drawing, as well as making some corrections. I find it interesting that this chip has current sensing for both 5v and 12v rails.

[everell]

Re: Chip of the year 2008

At last I have found the culprit. The 3.3 volt rail with no loading whatsoever comes up at about 3.4 volts. With one HDD and one CDROM for a load, it comes up to just a tad over 4 volts, spins the fan a few times and stops. Adding a 15 ohm resistor to the 3.3 volt rail and it comes up to 3.4 volts and stays on. The 5v and 12 volt rails are not giving any problems.

What baffles me is that under no load conditions, the 3.3 volt rail goes up to just a tad over 4 volts and stops after 275 msec. With the 15 ohm resistor, the 3.3 volt rail goes up to 3.4 volts for the same 275 msec, then goes up to just a tad over 4 volts for several seconds while the HDD settles down, then decreases to 3.4 volts. I would have thought that going that high AFTER the 275 msec would still shut down the power supply??? See attached pictures!

Conclusion, I can either add another 15 ohms to the 3.3 volt rail in the power supply, or I can use a real mother board for testing and thereby have a good load on the 3.3 volt rail.

[Th3_uN1Qu3]

Re: Chip of the year 2008

Maybe you're chasing the wrong circuit. Identify the voltage reference (usually 2.5v) and the pins that do the sensing for the PWM driver. See by how much the voltage changes there when you add a load to the 5v vs 12v. Here is one area where having three multimeters (even if they are the cheap kind) and lots of alligator clips, pays off. It may not be the OCP turning them off, it may be another protection. Recall that the main protections of those PSUs are OVP and UVP - check that all caps are good and all voltages are in spec.

3.3v feeds off 5v so that isn't much of a concern, focus on the influence of 12v and 5v in various points of the chip.

[everell]

Re: Chip of the year 2008

I have been working on this Solytech power supply again. With only a CDROM for a load, it comes up working fine. If I add a 2.5 ohm load to the 5 volt rail, it works fine. If I add a 1.67 ohm (two 3.3 ohm power resistors in parallel) to the 3.3 volt rail, it works fine. If I add any load to either 12 volt rail, it stops. To balance the load, if I add 12 ohms to 12V1 and 12 ohms to 12V2, it stops. My conclusion is that the over current protection circuit for 12V1 and 12V2 is somehow the culprit.

Pins 9 and 10 on the 2008Z chip of the year are the 12volt ocp pins. I can ground the two pins, I can leave them open, still stops. Each has a 100 ohm resistor between the rail and the ocp pin on the chip. I have tried increasing this value, and decreasing this value. Still stops. Each ocp pin on the chip has a .1 uF filter capacitor to ground. I tried bridging additional capacitance across them (22 uF). Still stops. My conclusion: either the 2008Z chip is bad or I am chasing the wrong problem.

I have acquired three more......same model power supply......same problem. For the moment, I am stumped.

[Th3_uN1Qu3]

Re: Chip of the year 2008

PSU just decided to smoke the wires connected to the CD-ROM.

The 5v rail got shorted on the cable. That's what happens when 20AWG wires are used, they are unable to pass enough current to trigger short circuit protection, so they melt.

[momaka]

Re: Chip of the year 2008

im sure there are other besides deer like my allied psu...

Deer = Allied = L&C. They hide under many other names too, but those 3 are the basic ones included with cheap cases. Parent company is Solytech.

I had an L&C do the same thing as your Allied, except there wasn't anything shorted - PSU just decided to smoke the wires connected to the CD-ROM. Everything worked separately afterwards, though (CD ROM, PSU, motherboard, etc.). Chip inside that PSU was labeled "2003".

[shovenose]

Re: Chip of the year 2008

im sure there are other besides deer like my allied psu that caught a p4 mobo of ine of fire-and the pc kept working until i unplugges it. interestingly enough both mobo and psu still worked!

[Th3_uN1Qu3]

Re: Chip of the year 2008

I don't know why, but almost all protection chips set the OVP and UVP thresholds at unreasonable numbers. Most of the time the 12v OVP is somewhere around 14-15V.

I remember someone saying that a Deer will continue working even if ALL voltages are off. I had a Powerlink that would continue powering a shorted motherboard. Luckily i noticed it before it blew up.

[everell]

Re: Chip of the year 2008

One more schematic ...... the pwm drivers which are driven by the 2008Z chip.

[370forlife]

Re: Chip of the year 2008

I don't know why, but almost all protection chips set the OVP and UVP thresholds at unreasonable numbers. Most of the time the 12v OVP is somewhere around 14-15V.

[goodpsusearch]

Re: Chip of the year 2008

Yes there are also overvoltage protection circuits inside the chip... but on the 2003 and 2005 controllers (dunno about the 2008) they're set too high to make any difference.

Unfortunately, on most power supplies they are set too high..

[Th3_uN1Qu3]

Re: Chip of the year 2008

Those chips does just fine........unless they don't......

That's a good one.

Why would something do wrong with the chip? If it flips out could it keep trying to raise voltage or what could happen?

Yes, among others they can fail in this way too. Any regulated power supply is designed to keep the output constant regardless of line and load variations (well, to a point). A switchmode supply does this by adjusting the width of pulses driving the primary transistors according to the load placed on the secondary. There are other variations but PWM-driven supplies are still the most common.

If the circuitry that detects the load fails, be it external parts around the chip or a failure in the chip itself, the PWM output will go to maximum. On a low load and normal line conditions, that means bad news for whatever is hooked to that PSU. Yes there are also overvoltage protection circuits inside the chip... but on the 2003 and 2005 controllers (dunno about the 2008) they're set too high to make any difference.

[shovenose]

Re: Chip of the year 2008

Why would something do wrong with the chip? If it flips out could it keep trying to raise voltage or what could happen?

[weirdlookinguy]

Re: Chip of the year 2008

and when they blow, they splatter anywqy all over the place.

[everell]

Re: Chip of the year 2008

Those chips does just fine........unless they don't......then they blow up the power supply.....and sometimes the mother board

[shovenose]

Re: Chip of the year 2008

Btw what does thos chip do anywqy?

[everell]

Re: Chip of the year 2008

Here is my hand drawn schematic of what is in the 2008 chip. Was not able to determine what two of the pins do, but the rest should be correct. Also the 5vsb circuit for this Solytech psu.