kinetix wrote on 2023-07-29, 18:53:

in my opinion, something is wrong in the jumpers settings documentation for 3.45v. I´ve been trying to find the correspondences to setting as voltage, cache type and multiplier, on both boards, the same. Unfortunately, I did not write down the configuration the board (both) had before changing it. I'm pretty sure I had it 20 years ago with that Am5x86. It would have been nice to have a "known" functional reference. and that is what I have been trying to find, the jumper configuration of those, anyone, who have one of those 3.45v processors installed on this board.

Yeah, considering that both the Intel and AMD DX4 variants support the VOLDET pin I don't understand why they would do it differently. I have another SiS471 board, the ASUS VL/I-486SV2GX4 which implements this feature in the following way: there is a jumper that allows you to connect the VOLDET pin of the CPU to the input of a 7407 buffer. The output of this 7407 is connected to the gate of the MOSFET, the drain is connected to +5V and the source is connected to the regulator output.

With the jumper cap on, if you use an Intel/AMD 3.3V CPU (which have the VOLDET pin connected to GND), the output of the 7407 will be low, so the MOSFET will be OFF and the regulator output remains at 3.3V. If you use a 5V CPU, the VOLDET pin is floating and the 7407 input is pulled to +5V through a resistor, so the gate is high and the MOSFET turns ON and the source is at +5V, effectively shorting the regulator output to provide +5V to the CPU.

Based on what I was able to test on the VI15G with my multimeter, the equivalent jumper in this board would be JP25 pin 3 (CPU VOLDET) and pin 4 (input "3A" of the 7407 at the corner of the board, near the cache tag chip). The "3Y" output of the 7407 is, as expected, connected to the MOSFET gate. So it's strange that you get 3.45V without the CPU installed, unless something else is grounding VOLDET, but there may be slight differences between your board and mine.

But if you're interested in knowing what each of the jumpers in the configuration table does, here's what I was able to find by poking around with a multimeter for Intel/AMD DX-type CPUs (again, keeping in mind my board and yours may have differences and that I can't do voltage checks on my 5V-only board):

• BLUE jumpers are common for all DX-type CPUs
• GREEN jumpers are for SL-enhanced CPUs (i.e. with power management features)
• YELLOW is for L1 WT cache, ORANGE is for L1 WB cache (obviously these two are mutually exclusive, it's either one or the other)
• RED is for voltage auto-detection, as I mentioned above
• PURPLE controls the internal multiplier for DX4-type CPUs, open sets 3x while closed sets 2x (but for 5x86, you get 4x instead of 2x!)

As for the other jumpers on the table:

• JP9 and JP10 set the "trap" to program the SiS chipset depending on the CPU type (for Intel/AMD, JP9 is always 1-2 and JP10 is 2-3 for CPUs with L1 WB cache, 1-2 for regular L1 WT)
• JP16: this one I am unable to find out. It is unpopulated on my board, but it appears to be used to set the SMOUT1 pin of the SiS chipset (#58) LOW through the 7407, not sure what this does really

kinetix wrote on 2023-07-28, 01:53:

... I measured the voltage without the CPU in the socket. I will test every jumper config and tell you the voltage in each case […]
Show full quote

...
I measured the voltage without the CPU in the socket.
I will test every jumper config and tell you the voltage in each case
....

No, that is not going to give you correct readings.
The LT1805CT needs a minimum current to actually be able to regulate its output.
Measure the voltage on the capacitor TC12 with a CPU installed.

Also use a Award BIOS if you like to lift the board from sub-par performance to one of the best performing SIS471 boards.
This also fixes the CPU identification issue.

Dorunkāku wrote on 2023-07-30, 00:55:

Also use a Award BIOS if you like to lift the board from sub-par performance to one of the best performing SIS471 boards.
This also fixes the CPU identification issue.

Nice, is that an original VI15G BIOS or is it patched from another SiS471 board?

I noticed that the AMI BIOS for SiS471 boards suck ass (both on this board and MSI MS-4132G for example). For some reason they ignore the fastest timings for DRAM and cache set in the BIOS, and other settings are messed up (local bus T2/T3 and SYNC/TRANSPARENT are inverted). I wonder why AMI fucked up these in such a way and never bothered to fix them.

TheMobRules wrote on 2023-07-29, 20:33:

Yeah, considering that both the Intel and AMD DX4 variants support the VOLDET pin I don't understand why they would do it differ […]
Show full quote

Yeah, considering that both the Intel and AMD DX4 variants support the VOLDET pin I don't understand why they would do it differently. I have another SiS471 board, the ASUS VL/I-486SV2GX4 which implements this feature in the following way: there is a jumper that allows you to connect the VOLDET pin of the CPU to the input of a 7407 buffer. The output of this 7407 is connected to the gate of the MOSFET, the drain is connected to +5V and the source is connected to the regulator output.

With the jumper cap on, if you use an Intel/AMD 3.3V CPU (which have the VOLDET pin connected to GND), the output of the 7407 will be low, so the MOSFET will be OFF and the regulator output remains at 3.3V. If you use a 5V CPU, the VOLDET pin is floating and the 7407 input is pulled to +5V through a resistor, so the gate is high and the MOSFET turns ON and the source is at +5V, effectively shorting the regulator output to provide +5V to the CPU.

Based on what I was able to test on the VI15G with my multimeter, the equivalent jumper in this board would be JP25 pin 3 (CPU VOLDET) and pin 4 (input "3A" of the 7407 at the corner of the board, near the cache tag chip). The "3Y" output of the 7407 is, as expected, connected to the MOSFET gate. So it's strange that you get 3.45V without the CPU installed, unless something else is grounding VOLDET, but there may be slight differences between your board and mine.

But if you're interested in knowing what each of the jumpers in the configuration table does, here's what I was able to find by poking around with a multimeter for Intel/AMD DX-type CPUs (again, keeping in mind my board and yours may have differences and that I can't do voltage checks on my 5V-only board):

VI15G_jumpers.jpg

• BLUE jumpers are common for all DX-type CPUs
• GREEN jumpers are for SL-enhanced CPUs (i.e. with power management features)
• YELLOW is for L1 WT cache, ORANGE is for L1 WB cache (obviously these two are mutually exclusive, it's either one or the other)
• RED is for voltage auto-detection, as I mentioned above
• PURPLE controls the internal multiplier for DX4-type CPUs, open sets 3x while closed sets 2x (but for 5x86, you get 4x instead of 2x!)

As for the other jumpers on the table:

• JP9 and JP10 set the "trap" to program the SiS chipset depending on the CPU type (for Intel/AMD, JP9 is always 1-2 and JP10 is 2-3 for CPUs with L1 WB cache, 1-2 for regular L1 WT)
• JP16: this one I am unable to find out. It is unpopulated on my board, but it appears to be used to set the SMOUT1 pin of the SiS chipset (#58) LOW through the 7407, not sure what this does really

very good info. thanks for your time

Dorunkāku wrote on 2023-07-30, 00:55:

No, that is not going to give you correct readings. The LT1805CT needs a minimum current to actually be able to regulate its ou […]
Show full quote

No, that is not going to give you correct readings.
The LT1805CT needs a minimum current to actually be able to regulate its output.
Measure the voltage on the capacitor TC12 with a CPU installed.

Also use a Award BIOS if you like to lift the board from sub-par performance to one of the best performing SIS471 boards.
This also fixes the CPU identification issue.

OK. I will do it. may I do it with a 5v cpu, just in case? (so, no VOLDET). maybe with the Intel DX-33 I have?

TheMobRules wrote on 2023-07-27, 21:00:

Just for completion as there are parts of the circuit that I'm still trying to understand: can you make out the values on the resistors R40, R75 and R76? These are not populated in my board so I have no way of knowing. I have marked the locations in red in the attached photo.

On my board:
R40=10K
R75=150
R76=270
The MOSFET on my board is a IRF Z44, the regulator is the same LT1085CT.
The regulator output is 3.55 volt.
Also my board has all the capacitor locations around the CPU socket filled, only the two SMD tantalums inside the socket are unpopulated.

The BIOS is not original, but it works well. You can still run Local bus ready set to Transparent even at 40Mhz FSB. This is something my MSI 4138 and my ASUS VL/I-486SV2GX4 cant do.

Dorunkāku wrote on 2023-07-30, 08:18:

On my board: R40=10K R75=150 R76=270 The MOSFET on my board is a IRF Z44, the regulator is the same LT1085CT. The regulator outp […]
Show full quote

On my board:
R40=10K
R75=150
R76=270
The MOSFET on my board is a IRF Z44, the regulator is the same LT1085CT.
The regulator output is 3.55 volt.
Also my board has all the capacitor locations around the CPU socket filled, only the two SMD tantalums inside the socket are unpopulated.

The BIOS is not original, but it works well. You can still run Local bus ready set to Transparent even at 40Mhz FSB. This is something my MSI 4138 and my ASUS VL/I-486SV2GX4 cant do.

Very interesting, I had written off this board as a poor performer, but looks like it was just the crappy BIOS! Now I'm tempted to add the 3V regulation components to my own board! One more thing if it's not too much of a hassle: can you tell me the value of the resistor marked in red in the picture? I just noticed that one isn't populated on my board either, seems to be a pull-up to 12V but I can't quite make out the markings.

4700 Ohm

It is connected to the gate of the MOSFET. In the picture the LT1085CT does not have a heatsink, it really needs one!

Dorunkāku wrote on 2023-07-30, 10:05:

4700 Ohm

It is connected to the gate of the MOSFET. In the picture the LT1085CT does not have a heatsink, it really needs one!

Yes, I removed the heatsink to take the photo, so the board could be seen better. in the photo of the complete board you can see it

TheMobRules wrote on 2023-07-30, 09:51:

Now I'm tempted to add the 3V regulation components to my own board!

I hope these pictures may help you to do that:

kinetix wrote on 2023-07-30, 17:39:

I hope this pictures may help you to do that:

Excellent, thanks for the pictures! I'll take a look in my stash to see what parts I already have and what I need to buy.

TheMobRules wrote on 2023-07-30, 17:58:

Excellent, thanks for the pictures! I'll take a look in my stash to see what parts I already have and what I need to buy.

Well, I measured the clock signal on the ISA bus with a friend's oscilloscope. it delivered 7 point something Mhz (i forgot it), although with a certain "noise" (a small spike) at the end of the rise and fall ramps of each pulse. I also forgot to test on the CLK pin of the CPU. I also forgot to take a picture of the oscilloscope screen ( so much memory 😀 ).
I will do it again and with more care in a couple of days, today I was in a hurry.
We will also try to read the BIOSes (he also have an EPROM programmer) of both boards and hopefully update one of them. although they are UV EPROM (M27C512-15XFI) and we don't have a UV eraser, but I think we can do it using a UV emitting lamp (from a toolkit for different use) + a fluorescent lamp + lots of sun + some time (hope it should do it, 🤣!! )

I forgot to mention earlier, the board happily tries to charge the battery all the way up to 5 Volt.
Measure the voltage on the battery when board is powered on.

Dorunkāku wrote on 2023-08-05, 09:49:

I forgot to mention earlier, the board happily tries to charge the battery all the way up to 5 Volt.
Measure the voltage on the battery when board is powered on.

mmm... I get only 0.1v on the battery contacts, on both boards

Today I measured all the motherboards.
voltages are OK. and this is what the clock signals on the ISA bus and on the CPU look like. I am aware that they are dirty.
the 486#2 board was very dirty and still slightly rusty, it also had a corner (the cache corner) with a slight damaged from fire (an accident in the garage) but only had the capacitor BC4 almost unsoldered by the heat, but I measured it and it seems ok.

The 486#1 (my target) and the Pentium are pretty pristine.

How good or bad are these clock signals?
are those spikes lack of grounding or some radiointerference? (I was measuring in a wood table and also the power outlet had no ground)
If those are bad, what could be the reasons?

also, I got a dump form both 486 boards BIOS. I compared both with the 2.3 version published in the RetroWeb and are identical. while reading one an error came out, something like "Vmin validation reading error", or something like that, but it seems to me that it is not a problem as it was read OK. perhaps the surface has a very thin layer of oxidation that creates resistance. The other did not give an error but was in bad condition its pins and I sanded and scraped them before reading the chip. maybe latter I´ll do the same to the first one and try again, to make sure there are no real problems.

this one is the Pentium board (AOpen AP53). I forgot to measure the clock in the PCI bus.
need opinion too, although this board is for a latter time is the main interest for me.

please, take a look to my last two post and give me an opinion about it.
EDIT: dont worry, already got answers in a different thread
here: Help with some CLK signal quality