Broken traces, bad capacitors, resistors, diodes, corrosion, and cracked/cold solder joints can cause that as well.

I have a link here: https://theforevernoob.wordpress.com/2012/05/ … hru-their-vrms/ that allows you to check for voltage spikes on the VRM. I'd suggest replacing the PSU as well since that must've lost its 3VDC power. They make ATX to AT w/ -5VDC and Dummy loads on eBay (not posting a link on here since it's against the rules) for hooking up modern PSUs to an AT motherboard.

I think it was last year, that I got an Luckystar 486 board. It had a dead voltage regulator upon arival. The voltage regulator was working when the seller shipped it. (I think I saw it both in pictures and on video). Though changing the voltage regulator, made the board work with an 5x86-133 again. Things can break and die down the road. Wether it is too old or the postal system does something bad.

how to properly diagnose if the regulator is indeed broken

I have M912 with a dead VR. Very easy to diagnose: CPU and transistor does not heat up.

what part number I could use to replace this potentially broken regulator

Model number is printed on it. Quite common transistor, should be easy to buy a replacement. I think 6 amper version is required.

Hi again,

thanks everyone for your feedback. I've been investigating a bit more the situation with the mobo. There seems to be two parts working together to regulate voltage: a TL431 which does the voltage regulation, and a D882 transistor which is in charge of generating enough current. When set to 3.3V neither the transistor, the regulator or the CPU gets hot, so according to @theserpentrider the regulator is pretty much dead. But what part? the TL431 or the D882 transistor? That's still not clear to me.

I tried following the guide linked by @bjwil1991 but I found it difficult to follow. Mainly because in my setup it's not clear what the VRM input and output are. The TL431 has three pins (Vref, Anode, Catode) and the D882 another three (base, emitter and collector). Which one of these can be considered the input and output pins of the VRM is not clear. I took some readings of voltages for all the aforementioned pins with the mobo setup for 3.3V and this is what I got:
TL431:
- Catode: 5V
- Anode: 0V
- Vref: 1.26V
D882:
- Base: 5V
- Collector: 5V
- Emitter: 1.72V

Hope this makes sense for anyone. If not, I guess I'll replace both parts (TL431 and D882) and do the usual praying to the gods of silicon.

Thanks in advance

You can also use a 5V to 3V adapter or Turbochip 5x86 upgrade that has voltage regulator built-in.

I think the TL431 is required for 4V output: https://www.onsemi.com/pub/Collateral/TL431-D.PDF

The TL431 is a programmable precision reference which is used in a variety of ways. It serves as a reference voltage in circuits where a non−standard reference voltage is needed.

I agree the tl431 provides the 4v output, but it should provide the 3.3V output as well, shouldn't it? In any case, I'm not getting any reading close to any of these values, only 5V 🙁

The fact that Vref reads 1.26V and the transistor base and collector read the same voltage tells me that probably the transistor is the culprit. Without the exact schematic used it's hard to tell for sure as they could be used in several different ways. It's probably worth replacing both if your soldering skills are OK as rhey're both cheap.

Quick update, it was a pity that I didn't read @stamasd message before. I indeed changed the regulator (TL431) and was able to get 3.45 volts as the output! However it didn't last for long, probably because the transistor (D882) is burnt and without it the regulator just cannot handle the load for too long. So, after some seconds (maybe one minute) I measured the voltage again and it was back to the previous measurements.

Next step is obviously to change both components and see if that makes the trick. I'll do this during the weekend, will let you know how it went.

Cheers

Hi,

I finally was able to replace both the regulator and the transistor for new parts. Soldering wasn't trivial, but not too tough either. Some experience with electronics definitely helps here.

After changing the parts, the good news is that I have a reliable 3.3v feed out of the regulator+transistor combo! My intel DX4-100 works flawlessly, however I cannot say the same thing about my 5x86-133. I have two of these and none of them POST (error code 26 on my PC Analyzer, something related to the turbo switch apparently). I'll keep investigating and let you know any progress, but at least the problem with the 3.3v mode in the mobo is solved 😁

Thanks to everybody that contributed to this thread.

By the way which transistor (manufacturer and part number) you've used ?

Yes, that's probably a good idea. Both components are easy to find nowadays, however in my local electronics store they had run out of D882s, so I had to chose an alternative option. I went for a 2SD1666 (I think it's from Sanyo, but not 100% sure), you just need to make sure you mount it in the right direction in your mobo. Regarding the regulator, I bought the TL431ACLP from Texas Instruments. It's basically the same part I was replacing.