Also be sure to have some airflow on the VRM to limit the risk...as it's close to the socket it may be indirectly provided by the CPU fan if the cooler permits it.

Minutemanqvs wrote on 2024-01-14, 09:44:

Also be sure to have some airflow on the VRM to limit the risk...as it's close to the socket it may be indirectly provided by the CPU fan if the cooler permits it.

Thank you so much my friend
Initially I will use a Pentium 133 type CPU, this is to verify that the board works, and do some benching, in the past I had a negative experience with a MII 300 (233 MHz), on an Intel VX and linear regulators, for two or three months it went very good, for office and TV use, with the spring season, it began to show continuous resets, the card went in for assistance, and never came back, it was replaced with a VIA VPX with switching regulators, it works but is terribly slow, never understand why???
Since then, at most on boards with linear regulators, I put up to 200 MHz, but certainly an additional fan is almost a must, at least when the weather is hot.

On the FIC PA2010+, I tried to straighten the pins of the chipsets, I just tried to distance as much as possible, the result is not the maximum, but the distance is in my opinion sufficient, they are not too close, and this ensures that they do not cross, and that the space isolates them sufficiently.

I performed a cleaning on the SIM4 slot, there was something between some pins, maybe the residue of a warranty label 🏷, attached to the RAM.

Since the regulators could be linear, but on this I will inquire to be sure, the mosfet near the capacitor is signed 9724 NDP6020P, I should look for a pdf for more information, it seems written on the MB Q2, the first pin leads to a U34 (AD 727 - 3310A - 3.3) with 4 pins on two sides.

The other two are, Q6 signed 9724 and NDP6030L, and Q9 signed 9724 and NDP6030L, basically twins ️, pin 1 of the Q6 leads directly to the S.7, this will probably be a 3.3V.

Theoretically, it would be almost ready to try the boot, the problems on the SIMM 1 and 4 remain, the SIM1 is too damaged, the SIM 4 has the hook broken , both should be replaced.

There are some points to clean up, but they are not easily accessible, and there would be the pins of the PCI1 and PCI2 to put in place, and then there is the 32-pin socket (socket) of the BIOS chip, which doesn't look nice, maybe I'll put a new one, because I don't know how to clean the clothespins.

The board thinking about it, is not yet ️ ready, because we need to repair the interrupted tracks, well I thought there were those upper side PCB, actually looking well at the back, I found others, ultimately the motherboard was in “unexperienced” hands, you can see from the broken hooks, the bent pins, in short, all things that a user, with a minimum of attention, would have avoided.

Ultimately they are damage that can be fixed, but I would have gladly avoided doing it, luckily some points are spaced enough, however it will not be easy, to fix them.

I have to correct what was written before, the pin of Q6 is not the n.1 that goes to the S.7, it is the n.2 (the one in the middle).

I move on a board on which I put my paws , as soon as possible, Soyo is the brand, the model is quite interesting , the model is SY-6IZM/3, ZX chipset and ATi integrated video.

The images show how it is, quite traditional today, but at the time it was normal not to have VGA or integrated audio, unfortunately the integrated VGA I don't think can be deactivated, and the AGP slot is missing, so if you want to change VGA you need to use a VGA PCI (to disable I think you need to disconnect the + of the VGA chip).

A little while ago I withdrew a mini batch, of hooves for BIOS chips, I need both on the 486 and possibly on other PCs, the fundamental problem, is that the oxidized plinths must be cleaned up, and even if properly cleaned, they may have not optimal connections, for this and other reasons, I prefer to replace them with the same but new ones.

Of course I could do a wash, for the oxidized ones, but it's not worth it, both for the time ⏱ it takes, and because all in all they are cheap, so I ordered others, to replace those of the MB type S.7 or 462, these I bought are 40 and 28 pins, they are for KB BIOS chips and BIOS chips (486 or earlier), I ordered the 32 pin ones, I think they can arrive within a week, or at most the next one.

I also bought a mini braid, to clean when it desolds, I will try first on a MB scrap, I usually did without, using the tin sucker, this braid should remove and clean up the stagnant contacts well, I have no idea if you need the flower, or if that serves when you have to de-sold, I think in the future (quite close) to get that too.

The first socket to be removed, on the 486 4FLUD-1.0, is that of the KB BIOS, with 40 pins, obviously the worst side is that of the battery side, in some pins you can see oxidation (green), if they then still ensure ️ the conductivity, it could also be, but there are three or four tracks below, which seem to me not good, maybe the problem of failure to start, does it depend on these?

The tracks next to it, even if oxidized, I have checked and they are not interrupted, so I would say that by removing this 40-pin socket, you will be able to understand what there is to repair (?), I think at least two tracks, and maybe it is precisely these that prevent the PC from starting.

For the moment I don't change that of the BIOS (28 pins), although changing it would give us absolute certainty, that there are no oxidation problems, and I think it doesn't change much, because visually it looks more than good.

I'll start with a couple of 40-pin socket pins, if I can free them, without too much trouble, I'll continue, otherwise (we don't get angry ), I'll stop to try something else, I mean instead of the soldering iron, could I try with the hot air gun???

Well after trying to free up some pins, I stopped, is there something wrong, or the PCB is very delicate or the soldering iron is not good.?

I do not think that by continuing they could make progress, in the sense that, on about ten pins perhaps one or two have been freed, the alternative would be to try to desolder it with the hot air, but I have never tried, and given the very delicate PCB, I would not want to cause more problems, such as damaging the back of the PCB, unfortunately for the moment I do not know how it can be removed, without damaging the PCB, after all I remove it just to fix the tracks below, and I would not want to fix others, because they have been damaged by removing the plinth.

Here unfortunately the difference could make it, a better equipment ??? Or the use of a different technique ???

For the moment it seems to me a half disaster, it is a real shame, that such a simple thing, which I then did on several cards, unsoldering hundreds of capacitors, here instead is almost impossible ???

I have some good experience with a desoldering gun, see necroware's review on that: https://www.youtube.com/watch?v=qeizk0Yquzw

On stubborn pins I found that the best sequence is:
- Add some flux/fresh solder
- Heat it up with the desoldering gun and move the desoldering gun head "around" the pin. This "moves" the pin slightly as soon as the solder melts
- While still moving around the pin, suck the molten solder

It works almost every time.

Minutemanqvs wrote on 2024-01-18, 11:37:

I have some good experience with a desoldering gun, see necroware's review on that: https://www.youtube.com/watch?v=qeizk0Yquzw […]
Show full quote

I have some good experience with a desoldering gun, see necroware's review on that: https://www.youtube.com/watch?v=qeizk0Yquzw

On stubborn pins I found that the best sequence is:
- Add some flux/fresh solder
- Heat it up with the desoldering gun and move the desoldering gun head "around" the pin. This "moves" the pin slightly as soon as the solder melts
- While still moving around the pin, suck the molten solder

It works almost every time.

Thank you so much my friend.
I saw your video, and first I want to compliment you, I really like your YouTube channel, I will follow it with interest, at the moment I have only seen the video in the link, but I have glimpsed other videos that I will watch as soon as possible.
Returning to the problem on the MB 486, it is possible that the method I usually use is not right, and that what you suggest is right.
This will make me change my system, currently I do as in the video, I use a soldering iron and a solder vacuum, even if I don't use flux, I heat the contact and when it melts I vacuum.
With the replacement of the capacitors, it works quite well, sometimes after having removed them I have to free the holes, and usually I put the soldering iron on the top side, and the solder vacuum on the back (or vice versa).
Unfortunately, when there is something with a lot of contacts, it doesn't work, because it is probably the movement that facilitates the desoldering operation, which is fundamental, I believe, and this is what makes the difference.
For this and other reasons, I will search the Internet for a soldering station (even a suction one), the same or similar to the one in the video, since it would also be useful for other jobs, such as removing RAM sockets, or AGP or PCI slots. , these are jobs that I will do on other waiting MBs, the latest being the FIC PA2010+, but I have an ASUS 462 with AGP slot to replace, then also a Fujitsu with two SDRAM banks to replace.
If you have had a chance to look at the previous pages, you will find them, I hope to update soon, with the repairs on these and other motherboards, I will try to find some time to work on them.

It's not my channel, it's necroware's channel! 😉

Minutemanqvs wrote on 2024-01-18, 19:31:

It's not my channel, it's necroware's channel! 😉

Sorry, I thought it was your channel, anyway it's very nice, I like the old PC topic, I saw a video of a CAF 286, it's externally identical to my CAF 386, it's not a famous brand but you can see that at the time, late 80's early 90's, it was still in business.

A typical application of the air soldering station, would be ideal in the removal of cache chips like the one in the photo, in the specific case, it is an L2 cache chip, of a Soyo SY-5BT (which I have), I don't know the reason but one pin has been removed, another next to it maybe is not optimally connected, so the alternatives would be, either leave everything like this and disable the cache from BIOS, or use an AMD CPU with built-in L2 cache, or replace the chip with another working one.

However, there would be another possibility, to repair the missing pin, and I think I'll try, milling the top of the chip (plastic), until I discover the broken contact, so that we can solder a thread to it, and make a bridge between the chip and the PCB, in this way, if the chip works, the 512 KB cache should come back to working.

Also about the Soyo SY-5BT or 5BT5, I accidentally saw that there is a very similar model, which is SY-5XB5, this has some very interesting jumper settings, in addition to the usual 2.8V or 2.2V there are also those for the 2.7V and 2.1V, let's say that they are not normally needed, but they could be very useful to put particular CPUs, or to do overvolt to stabilize, in short, if you want to try you could even install the K6+ with a slight overvolt (+0.1V).

For the 5BT5 I think there is no modified BIOS, for the support of these CPUs, if you want you could ask if there is the possibility, to have a suitable one as soon as available, I think you can make the necessary changes, only if you are experienced enough.

I put the photos of the pages of the manual, but I know that there is a version 1.0 and a 2.0, version 2.0 has an additional jumper for selecting the 2.2V.

I tried to figure out if the two motherboards use the same scheme, for the selection of the VCORE, actually not having the other board, but only the SY-5BT Rev.2. X, I can't verify the correspondence of those settings, and if they use the same SMD (resistors etc...), I just realized that the 5BT5 always uses two jumpers, while the other doesn't, but I'm pretty sure there is some undocumented secret setting, to get other voltages, always using two jumpers, first of all we notice that 3.52V and 2.8V use the same position as the jumpers, here I think some pins of the Socket 7 intervene, a kind of automatic recognition, like if the circuit is open then it will be dual voltage (or vice versa), so in my opinion, the possible combinations would be these:
Jumper JP30 5BT5

1-2 3-4 5-6 7-8 9-10 11-12 Rev.2.X
1-2 3-4 5-6 7-8 9-10 Rev.1.X

Rev.2.X 3-4 9-10 ON
Rev.1.X 1-2 7-8 ON
= 3.3V>2.6V?

Rev.2X 3-4 7-8 ON
Rev.1.X 1-2 5-6 ON
= 3.2V>2.5V?

Rev.2.X 3-4 5-6 ON
Rev.1.X 1-2 3-4 ON
= 2.9V>2.7V?
Or the last one instead of 2.7V? Would it actually be 2.2V?, because the others have a lower voltage of 0.7V.

Certainly there will be a difference, and maybe you could try with a measurement of the VCore on the mosfets, usually the cards designed for the 2.8V can go down to 2.4-2.5V, removing the jumper from all the pins, in this card I don't know if it is enough to remove a jumper of the two, to get the same result, maybe using a single jumper you can get a lower voltage, like something around 1.25V, and for me it would be insufficient to boot the PC with CPU type MMX or K6 (without 3DNow! And with ?).

Here, looking at the circuit of the VCORE jumpers, of the 5BT5 Ver.2. X, a voltage of 3.3V is currently selected, because I use a Pentium 133 CPU, if I switch to a dual voltage CPU, I could get a lower voltage, maybe 2.6V?

Maybe yes, or maybe not, let's see if it's possible to understand something?

There are two methods to determine the exact voltage, one is mathematical, just make a calculation and you get the voltage, I am not an electronic technician, so I am not able to do this calculation, theoretically you should start from 3.3V and by resistance, you should get the value of the VCORE, which is definitely lower (I imagine yes ), the value of the resistance can be read from the photo (see SMD), but there could be a second resistance connected in some way, which could change the final result.

The other method, let's call it experimental, would be in practice this, before trying a dual voltage CPU, you can calculate from the +3.3V of the power supply, what resistance is needed to get to 2.6V, after several attempts you will find a value in OHm 🕉, but it could be completely useless, because in reality the voltage is not reduced directly by the +3.3V, but it comes from a Mosfet that has a VRef and a VOut (as well as a VIn), and therefore only at a certain VRef you get a VOut of 2.6V, so if you are capable, just do some simple calculations, otherwise (could we get angry?), we risk The CPU on duty, and we measure the voltages with the tester.

The risk that everything goes up in smoke, could be very low, already why risk the CPU, when you can do this measurement without the CPU on board, how do you do it ???

That's why I have to do some research, they could be fast, like not, as soon as I have news I'll update you.

Here is this image, it shows how to do the measurement of the VCore, without CPU, three resistors are enough, one I think is needed to have the double voltage, the other for the I/O, and finally there is that of the VCORE.

Personally I've never tried, but I think it can be very useful, and I guess it works, the alternative would be to use a K6, maybe the most suitable is a 3.2V 233.

I have between my legs, a Soyo motherboard, I think it is not a fairly common model, but it is undoubtedly particular, because it is a kind of ante i810, I know that that chipset integrated the VGA and this one did not, but it is present on board a VGA ATi with maybe 4 MB (or 8 MB?), plus the audio also integrated, and to complete the equipment, there are included two PCI modem and LAN cards, and an ISA TV card, I do not think the latter can receive any broadcaster, however it was included and I am not sorry at all.

With the card, the 40PIN IDE and 34PIN floppy cables have also arrived, the second serial port, a speaker and the leds and switches, for boot and reset, there is then a CD that I believe contains the drivers and utilities of this card.

I wanted to try it already tonight, but I noticed that there are two swollen electrolytic capacitors, they absolutely have to be replaced, maybe together with the other neighbors, it's a job that can be done, but I don't think soon, apart from that it looks like new, I would say in very good condition, the CPU is a PII 266 MHz, and there are three RAM, I think 64 MB each, at least for the moment that's all.

Returning to the 5BT5, I was observing on other MB Soyo always S.7, which use only four jumpers, to select the voltage, it usually works like this, one adds +0.8V the second +0.4V the third +0.2V and the fourth +0.1V, but not using any jumpers you usually get 2.0V, which would be the minimum voltage of the VCore.

In the MB Ver.2. X I have, there are six jumpers, so at least two could be, a kind of duplicate that would change the way you get the voltage , that is, if normally 2.6V you get with two jumpers on +0.4 and +0.2, here there could be a jumper or two that select +0.6V or +0.7V for example.

To be more precise, I would have to compare a circuit with four jumpers, just to see if it has similar resistances, and to see if the connections (tracks) are made the way.

While I was looking for, I found an interesting jumper, always on Soyo S.7, it is called JP16, and it serves to make the PCI work at 32 MHz, any FSB is selected, this is very interesting, because you can make it work with FSB 83.3 MHz without any problem on the PCI.

I don't think there is also on the SY-5BT, but who knows, there is for example a JP10 of which nothing is known, maybe it is useful for something else, but I remember that on the frequency generator, there are a couple of settings, which had FSB 75 and 83.3 with PCI of about 33 MHz (asynchronous), unfortunately with the P133 it does not work, I have not tried with other CPUs, maybe it can work ???

The next MB I will put my legs on, will be a particular S.7, maybe I'm the one looking for particular cards, which then seen with today's eyes, would be a very normal ATX card, but in its time, it was quite strange both in format and equipment.

Unfortunately I was going to get that SOYO with VPX chipset, but someone bought it, so I chose this one.

For those who ask what it is, it is a Siemens MB ATX, with Intel HX chipset and integrated video, as an option there would also be integrated audio, but I don't think it is present, the model is D969, I think there is no L2 cache, to have it you have to use a special cache module, to be inserted in the cache slot, and it can be a maximum of 512 KB.

From what I understand, it does not have dual voltage, in the pages of the manual, there is nothing written about the CPU voltages, but as an alternative to the Pentium it is possible to use the Overdrive, or the AMD K5, but among the settings there is even a 233 MHz, maybe Intel thought to get to that frequency with the Pentium (not MMX), but there is an interesting AMD K6 233 released from the factory with VCORE 3.3V, unfortunately I do not have it, but easily the 3.2V one also works at 3.3V, and it can safely tolerate a voltage of just 0.1V more, there is only to See the controllers what type they are, I guess linear, and so to have some peace of mind and a particular system, I would opt for a K5 CPU, for DOS it is quite efficient, but not so much.