Many Intel dx4-100’s are notoriously difficult to overclock.

Intel DX4-100 Overclocks?

Some folks with old motherboards were able to overclock using
a 75mhz crystal (37.5mhz FSB)

I never had any success, might be a combination of motherboard model, cache and bad cpu lottery though

Most DX4-100s from intel or AMD do not run reliably at stock voltage when overclocked to 120MHz, you need to raise the voltage to 3.6V for that, if it still does not run reliably at 3.6V, you then may need to customize the voltage regulator for 3.7-3.8V output, anything above 3.8V is harmful for the overclocked DX4 in the long run. I have two systems with DX4-100 @ 120MHz, in both cases I had to increase voltage to 3.6V and keep them cool, both CPUs are unstable w/o adequate cooling at that frequency, unfortunately.

Thank you guys, especially for the link. It seems the situation is much as I suspected it would be.

Looking @ https://www.ardent-tool.com/CPU/486_Step.html then maybe I will try and get one of the '89, '94 that support WT cache and try that. I am also wondering if these CPUs support PSE (Page Size Extension), does anyone know?

It also lists SX896 as '89, '94 with CPUID 0480, so WT cache, I tried Googling for this and wonder if it really exists. In passing I noted SX158 which seems to be an i386!

My interposer only offers 3.45 + 3.33 volts, do any exist that offer higher voltages and if so where might I get one please?

If you have the part number for the voltage regulator on the interposer, then it can be calculated what to change to increase voltage, and solder on either different resistors or regulator.

MikeSG wrote on 2025-04-16, 12:27:

the part number for the voltage regulator on the interposer

Thank you for your post I assume the VRM is the device with the small heatsink, it's a SC EZ1085CT 9517. The interposer is a PGA-DX4/2, I feel I should get another one before I start soldering, but am none too keen of the price of them. Mine came with my i486 DEC Venturis 4 motherboard along with an AMD DX2/66.

The settings are described in https://www.ardent-tool.com/CPU/486_interposer.html

It’s a waste of time and money trying to over clock a 486 CPU.
I don’t care if you over clock it to 160mhz or 200mhz.
In the end its still a 486 CPU
I can’t play MP3’s or Video very well.

You are better off with a Pentium class CPU.

Intel486dx33 wrote on 2025-04-16, 13:40:

You are better off with a Pentium class CPU.

I am never sure how to respond to posts such as this, or if I should respond at all, clearly you don't know I already have several Pentium class CPUs, I even have a Pentium class POD83 SU014 I can run in my i486 motherboard @ 100 MHz !

The main reason I choose to have an i486 system is so I can test my SIV utility on a CPU that does not have TSC support.

Before making posts such as the above I suggest you check if there is a URL in the user's profile (there is one in mine), if there is look at it and consider if your post will be helpful to them.

red-ray wrote on 2025-04-17, 07:22:

Intel486dx33 wrote on 2025-04-16, 13:40:

You are better off with a Pentium class CPU.

I am never sure how to respond to posts such as this, or if I should respond at all, clearly you don't know I already have several Pentium class CPUs, I even have a Pentium class POD83 SU014 I can run in my i486 motherboard @ 100 MHz !

Like on every troll post. Give him a stinky ><((((*>
Perhaps a DX4-100 is too fast for a user that named himself named DX33, not too slow.

red-ray wrote on 2025-04-17, 07:22:

The main reason I choose to have an i486 system is so I can test my SIV utility on a CPU that does not have TSC support.

Before making posts such as the above I suggest you check if there is a URL in the user's profile (there is one in mine), if there is look at it and consider if your post will be helpful to them.

I know you're looking for the differences between selected 486 CPUs.
Long time ago I have asked whether there is an Intel 486DX4 even without CPUID instruction, but I got no answer. With my AMD Am486DX4's I should test this. I at least know that my AMD Am486DX4 NV8T and SV8B are very different.

red-ray wrote on 2025-04-17, 07:22:

Anyone who runs NT on an i486 is unlikely to be wishing to play MP3 files on that system.

No, I don't think so. I consider playing MP3s on any 486 as ultimative test.
Playing MP3 on an AMD Am486DX4 / X5 at 100 MHz likely won't work (16 bit, 44.1 kHz, stereo, 128 kbps) but may work somewhen using a higher clock frequency (120, 133, 150, 160 MHz).
Since I don't have any Intel DX4 and their cores are better than the AMD ones it is interesting to see how they process MP3 at 100 and 120 MHz, in both &E and &EW flavour.

Intel did not like 40 MHz at all. (The only exception is an Overdrive 486 for a 486DX20 system.) And they did not like to cannibalize their early Pentium sales with a too fast 486. Intels 486 DX4-100 is almost comparable to a Pentium 75.

Contemporary overclockers used peltier coolers like Velox IceCap.
https://brassicgamer.blogspot.com/2023/01/the … 50-article.html

Edit: Modified the quotes so that they contain the respective name.

When you guys test MP3’s, are you doing it in Windows or in DOS using something like mpxplay?

This thread is not about running MP3s on i486 CPUs! If you wish to have such a discussion would it be sensible to start a new thread rather than hijacking this thread?

I apologize.

My notes say a DX4 @ 120 has the integer performance of a Pod83, and the floating point performance of a Am5x86 (133). So it's worth it if true.

This is the schematic for the Voltage Regulator: https://www.alldatasheet.com/html-pdf/42419/S … 3/EZ1085CT.html

The two or three resistors closest to J3 appear to set voltage. To confirm, the numbers on top of the resistors need to be read or measured with a tool.

In the schematic, one is R1 and one is R2. The sum to calculate Voltage out = 1.25 * (1 + R2/R1). An R1 of 133 and R2 of 232 = 3.45 voltage out.

If you confirm those are the resistors that switch between 3.3 and 3.45 then you should be able to work out which resistor/s to change to increase voltage.

MikeSG wrote on 2025-04-17, 11:48:

If you confirm those are the resistors that switch between 3.3 and 3.45 then you should be able to work out which resistor/s to change to increase voltage.

Thank you and I can confirm this, but the values seem to be different. Rather than changing R2 I suspect I could put a resistor between pins 1 and 2, could I?

11->2  3.70  R2  436  ?
21->2  3.65  R2  436  ?
31->2  3.60  R2  436  ?
41->2  3.55  R2  436  ?
51->2  3.50  R2  436  ?
61->2  3.45  R2  436
72->3  3.30  R5  408
8            R1  255

I did the sums, but the numbers don't add up, transforming the equation it's R2 = ( volt * R1 ) / 1.25 - R1 so R1 should be 248, did I get the transformation correct?

I tried to use a table, but it does not work on this site, it does on my site, see https://siv.boards.net/thread/200/woner-bbcod … bles-work-looks

[table][tbody][tr][th]Connection[/th][th]Voltage[/th][th]Resistor[/th][th]Value[/th][th]Ext[/th][/tr]
[tr][th]1->2[/th][th]3.70[/th][th]R2[/th][th]436[/th][th]?[/th][/tr]
[tr][th]1->2[/th][th]3.65[/th][th]R2[/th][th]436[/th][th]?[/th][/tr]
[tr][th]1->2[/th][th]3.60[/th][th]R2[/th][th]436[/th][th]?[/th][/tr]
[tr][th]1->2[/th][th]3.55[/th][th]R2[/th][th]436[/th][th]?[/th][/tr]
[tr][th]1->2[/th][th]3.50[/th][th]R2[/th][th]436[/th][th]?[/th][/tr]
[tr][th]1->2[/th][th]3.45[/th][th]R2[/th][th]436[/th][/tr]
[tr][th]2->3[/th][th]3.30[/th][th]R5[/th][th]408[/th][/tr]
[tr][th][/th][th][/th][th]R1[/th][th]255[/th][/tr]
[/tbody][/table]

you could switch the resistors then measure the voltage on the socket with a multimeter without a CPU installed. I would be doing that even when confident the resistors are correct to have the peace of mind the correct voltage is there physically

red-ray wrote on 2025-04-17, 13:14:

I did the sums, but the numbers don't add up, transforming the equation it's R2 = ( volt * R1 ) / 1.25 - R1 so R1 should be 248, did I get the transformation correct?

It does look like R1 should be 248 instead of 255...

Another way to check is as Treeman said, measure the VCC on voltage regulator (middle pin, 2) while plugged in and switched on. The values you said for R5 & R1 would return 3.25v, and R2 & R1 would return 3.38v.

Changing R2 to the values in the table should increase the voltage to what you need but I'd check actual output as well.

red-ray wrote on 2025-04-17, 13:14:

I suspect I could put a resistor between pins 1 and 2, could I?

MikeSG wrote on 2025-04-18, 09:03:

I'd check actual output as well.

Thank you, I planed to check the actual voltages and also what happens in no jumper is installed. Is putting a resistor between pins 1 and 2 a sensible idea?

MikeSG wrote on 2025-04-18, 09:03:

It does look like R1 should be 248 instead of 255...

255 would work with a R2 of 448, all in all I feel I need to measure the voltages.

In overclocking the DX4 in missing one aspect: run it in DX2 mode with a 50MHz bus. Het some good 15ns cache (12 of you can find out) and 60ns ram (50 if you can find it). Watch it fly!
No more than 2 VLB cards :')

waterbeesje wrote on 2025-04-18, 11:41:

DX2 mode with a 50MHz bus.

Thank you for the idea, but my 486 motherboard only supports 40 + 33 + 25 MHz

red-ray wrote on 2025-04-18, 09:58:

Is putting a resistor between pins 1 and 2 a sensible idea?

Yes should be able to run a resistor there.

To check
Pin 1 of J3 should be ~436 to pin 1 of the Volt Regulator.
Pin 2 of J3 should be 0-ohms to pin 2 of the Volt Regulator.
Pin 3 of J3 should be ~408 to pin 1 of the Volt Regulator.