VDroop mod for the good old Biostar TP35D2-A7?

[Krelin]

Is there a VDroop mod for this oldie MoBo? In my case the VCore droops from ~1.42V to ~1.39V idle to stressed (E8400 C0 @ ~4GHz). :( Any help appreciated.

[ceemic]

Krelin, did You measure that Vdroop with DMM?
Do a close-up macro from that area:
http://ceemic.pri.ee/hardware/Biostar_TP35D2-A7_1.JPG
Something like that: http://ceemic.pri.ee/hardware/tpower/tpower20_mod.JPG

[Krelin]

Thanks for the reply. :) Gonna check it with my cheap-o DMM and make the macro-shot tomorrow.
Btw, interesting homepage You got there, ceemic. :yepp:
Regards from Latvia. :wave:

[ceemic]

Yeah, http://ceemic.pri.ee/hardware is public to everybody :D

[Krelin]

Ok, measured with my cheap-o DMM: Idle ~1.39 - 1.40V (Value in BIOS - 1.412V); Stressed ~1.38V
Oh, and the dusty macro shot... :rolleyes:

[ceemic]

Hi!
Sorry for my late answer, I missed Your post.
ISL6322 datasheet: http://ceemic.pri.ee/hardware/datasheets/ISL6322.pdf
Now theres one "but" with that. It's pin layout is similar to ISL6327: http://ceemic.pri.ee/hardware/datasheets/ISL6327.pdf and ISL6334: http://ceemic.pri.ee/hardware/datasheets/ISL6334.pdf
These datasheets says that "proper resistor between FB and VDIFF can set load line (Vdroop)". I have moded mobos with these ICs and Vdroop mod between FB and VDIFF worked just fine.
But ISL6322 is a little different in load line. For example DFI P45 T2RS-Plus uses that and I even did the same Vmod between FB and VDIFF. But I read that datasheet wrong that time and that mod acted strange. It helped a little in Vdroop, but when I lowered the resistance too much, mobo shut down.
So I'm not quite sure how the Vdroop mod should go to Your motherboard.
I'll ask from someone better in Vmoding ;)

[celemine1Gig]

VDroop mod usually is a resistor in parallel to Rfb, which should be located between pin#14(FB) and pin#16(VDIFF). That works, because Idroop is connected to FB.
Your problems might have originated from the fact that, like in the quote below, Idroop doesn't necessarily need to be connected to FB. Maybe that's why it didn't work.



IDROOP
The IDROOP pin is the average channel-current sense
output. Connecting this pin through a tuned parallel R-C
network to FB allows the converter to incorporate output
voltage droop proportional to the output current. If voltage
droop is not desired leave this pin unconnected.

[ceemic]

celemine1Gig, tnx!
Actually, I probably know what was the problem. The owner of that board managed to crap some SMD caps and I had to replace them. Also some traces were broken and I can't be 100% sure that I could repair it perfectly. But I'm pretty sure that FB and IDROOP were connected...
And I can't say that this Vdroop mod didn't work at all. There was noticeable lower Vdroop compared to def board. The unstability might be due to the broken components...

EDIT: note to myself. I have to read the datasheets more thoroughly next time. I read it now and it all makes sense :(

Tnx again Ingmar :up:

[celemine1Gig]

celemine1Gig, tnx!
Actually, I probably know what was the problem. The owner of that board managed to crap some SMD caps and I had to replace them. Also some traces were broken and I can't be 100% sure that I could repair it perfectly. But I'm pretty sure that FB and IDROOP were connected...
And I can't say that this Vdroop mod didn't work at all. There was noticeable lower Vdroop compared to def board. The unstability might be due to the broken components...

EDIT: note to myself. I have to read the datasheets more thoroughly next time. I read it now and it all makes sense :(

Tnx again Ingmar :up:

Your welcome.
And yes, that does make sense. If some of the replacement smd caps were not in the tolerance range of the original parts this could have caused the instability. I doubt that the trace repair caused a problem.