• Please review our updated Terms and Rules here

Packard Bell M415 with heat-sinks/voltage-regulators ripped off, anyone know the specs?

GearTechWolf

Experienced Member
Joined
Apr 24, 2015
Messages
264
Location
Washington
Forget when/where I got this all-in-one mobo, but it was after scrappers had ripped the heat-sinks off, taking the voltage-regulators with them.
Nothing special about this board, just want to see if it functions at all for the heck of it.
Its al;so missing the riser-card, unless I have that somewhere else in my clutter, so at most it might be useful for testing floppy/IDE drives, ram, and socket-7 CPUs.
 
Alternatively, anybody got a spec-sheet/datasheet for an 8-pin SMT chip marked National Semiconductor 79BT 2951 CMB?
Both of the voltage-regulators feed (or maybe receive from?) one of these chips, making me think that might be an avenue to knowing the voltage.
Unfortunately, my searches based on these markings have come up empty. Other data, the capacitors connected to these VRs are 16V 100uF.
How does the rule go, rate your caps at least 1.5X the voltage they'll be operating with? (or is it 2X or more?)
Reversing 1.5X gets me roughly ten volts, which doesn't make sense for a common voltage-rail AFAIK.
 
Going from this picture:

Judging from the large heatsinks and the lack of any toroidal chokes near the VRMs, I'd guess they were linear regulators. Unless they're hidden from view, it's not a very good angle.

They could be something as simple as LM317T or an equivalent. You could do a bit of measuring with a DMM between VCore and VIO pins on the CPU socket and the three pins of the VRM. VCC2 should be the core voltage, and VCC3 should be the I/O voltage, which will always be 3.3v. Depending on what CPU is installed, VCC2 (core) can vary. On older single rail CPUs, like the original Pentium, this is 3.3v. On split plane CPUs like the AMD K6, Pentium MMX or Cyrix 6x86MX, this can vary between 1.8-3.5v.

Each VRM is probably responsible for one of the VCC2 and VCC3 voltages, you'll just have to figure out which. You may be able to look at other motherboards at the time to get an idea about which linear regulator should be used.

03fig18.jpg



As for what voltage capacitors to use, anything above the working voltage. For 5 volts, you can use a 6.3v or 10v capacitor. For 12 volts, you can use a 16v or 25v capacitor. When in doubt, use the original voltage rated capacitor. Don't use a capacitor with too high of a voltage rating, because the internal plate structure changes with voltage and alters the characteristics of the capacitor, as well as the size. So you wouldn't want to use a 50, 63, 160, etc volt capacitor on a low voltage circuit because the size and ESR will preclude them from working properly.

For capacitors in hot areas, you may want to use more expensive solid polymer capacitors. These tolerate heat better if you get 105C or higher rated types. Next to smoking hot linear regulators, or being bathed in hot CPU exhaust, I wouldn't use anything but polymer caps.
 
As for the NLX card riser, you're going to have a tough time getting one, because they're all proprietary. While the NLX form factor was somewhat standardized, the riser board was not, and every manufacturer made their own. Even within manufacturers, risers were not standardized and couldn't often be swapped between boards.

You'll need to cannibalize another M415 PC to get that riser, or do some extensive heavy reverse engineering of the board and slot to figure out where all of the lines go, then lay out your own riser and have it fabbed.

It doesn't look like the riser has any active circuitry on it, just some capacitors for power filtering, and maybe some decoupling capacitors.

It also looks like the connector on the board may be a repurposed EISA slot, so a simple prototyping card could be made with an EISA end to break out all of the pins so they could be easily probed to figure out what they are.
 
Excellent, thank you! That gives me a lot to go on and will be a helpful methodology for attempting to fix future boards too!
I found a similar-era socket-7 board when checking for lingering Varta-batteries* in the PCs I had in storage, took a picture of its similar regulator.
*(found a 386 board that still had one, which had leaked of course, but fortunately the damage is very minor! Just need to replace a dip-20 socket)
 
Back
Top