I don't wish to modify the p24t in any way considering its value. What I'll probably do is replace each jumper with a three pronged transfer switch to automate the clock speed switch. So the verdict is either an underperforming capacitor, or a voltage regulator that is too weak to charge the caps and power the p24t at the same time during startup. I wonder if insufficient power delivery is the reason why many 486 motherboards were incompatible with the p24t 83 despite having the proper socket and jumper settings?

3dnow wrote on 2023-03-19, 04:06:

I don't wish to modify the p24t in any way considering its value. What I'll probably do is replace each jumper with a three pronged transfer switch to automate the clock speed switch. So the verdict is either an underperforming capacitor, or a voltage regulator that is too weak to charge the caps and power the p24t at the same time during startup. I wonder if insufficient power delivery is the reason why many 486 motherboards were incompatible with the p24t 83 despite having the proper socket and jumper settings?

Now seriously: Did you try another power supply?
A P24T is designed to work with external 5 volts.

CoffeeOne wrote on 2023-03-19, 19:37:

Now seriously: Did you try another power supply?
A P24T is designed to work with external 5 volts.

I tried three:
A 300w unit from a dell optiplex from a few years back, a 400w Insignia, and a 600w evga that I got last year.

Overdrive 83MHz has onboard regulator means you need very good 5v supply that doesn't drop during power up or change due to auto-configuration bios. Often a mosfet is not sufficent to carry current in 5V mode. Do you

You need to modify the motherboard by adding a hidden jumper on back of the motherboard's to override the mosfet to have stable 5V for the overdrive 83MHz CPU.

Cheers,

pentiumspeed wrote on 2023-03-20, 16:57:

Overdrive 83MHz has onboard regulator means you need very good 5v supply that doesn't drop during power up or change due to auto-configuration bios. Often a mosfet is not sufficent to carry current in 5V mode. Do you

You need to modify the motherboard by adding a hidden jumper on back of the motherboard's to override the mosfet to have stable 5V for the overdrive 83MHz CPU.

Cheers,

Please, provide a guide. 😀

Nexxen wrote on 2023-03-20, 19:21:

pentiumspeed wrote on 2023-03-20, 16:57:

Overdrive 83MHz has onboard regulator means you need very good 5v supply that doesn't drop during power up or change due to auto-configuration bios. Often a mosfet is not sufficent to carry current in 5V mode. Do you

You need to modify the motherboard by adding a hidden jumper on back of the motherboard's to override the mosfet to have stable 5V for the overdrive 83MHz CPU.

Cheers,

Please, provide a guide. 😀

I know the jumper setting to force 5 volts on a Asus VL/I-486SV2GX4.
The board has voltage autodetection. So forcing 5 volts means to ignore the voltdet pin of the CPU.
Re: 3 (+3 more) retro battle stations

The P24T has the VOLTDET pin on the same position than an enhanced Intel chip (like Amd Am5x86 for example).
But: I don't know how this is done on a PVI-486AP4, but I am 99.99% sure that the PVI-486AP4 also uses voltage auto detection.

3dnow wrote on 2023-03-20, 13:21:

CoffeeOne wrote on 2023-03-19, 19:37:

Now seriously: Did you try another power supply?
A P24T is designed to work with external 5 volts.

I tried three:
A 300w unit from a dell optiplex from a few years back, a 400w Insignia, and a 600w evga that I got last year.

For sure all 3 should be more than sufficient.
But those watts really say nothing, only thing that counts is amps on the 5 volts line.
Just saying, because there are some modern power supplies that deliver less current on the 5 volts rail than a super old AT supply rated with 200W.

Pentium 83 takes more watts on 5V than a 5x86 would. And is very dynamic watts when running (watts is power equals current multiplied by voltage). Using a mosfet as switch that bypasses the (basically shorting) 3.3V regulator to supply 5V direct is not ideal since silicon doped junctions incurs some voltage drop and gets more lossy (voltage drop get higher), as heat soaking get more.

Am5x86 is about 6 to 8 watts while pentium overdrive 83 is 14.7 watts, ways higher.

This is what I was saying is as part of diagnostic is to short two pins of the mosfet together using a jumper is sufficient. The mosfet has Drain and Source that you want to short, leaving gate terminal alone, (Gate pin is an control that closes the D & S to allow current flow).

I cannot read the markings very clearly, can you complete the missing markings? NDP606__ ? This is above the memory slots.

The one with LT1085 or similar is a voltage regulator to provide 3.3V if used.

Cheers,

Back to square 1: a bad cap or mosfet isn't up to the task (low quality/worn/about to die).
Buying a new one to try if it makes any difference?
Desolder and test relevant caps around the cpu?

This is just an interesting observation: There was a cold front in my area for the past few days with temperatures in the low 60's/high 50's at night. Early morning when it was still cold out the PC would post at full speed without having to manipulate the jumpers. I would then load up quake (vanilla DOS Quake yesterday and GLQuake with a voodoo 1 today) as a stress test for a few hours before I left for work. Which components do you think are most affected by the cold?

PS.
Rebooting via the reset button or the power switch would not result in a successful post.

3dnow wrote on 2023-03-21, 18:46:

..... Which components do you think are most affected by the cold? ......

The CPU 😁