liqmat wrote on 2021-07-03, 20:08:

Hey britain4, […]
Show full quote

Hey britain4,

I purchased and tested one of those AGP boards. They will only go to AGP 2x mode with the 5950 Ultra. Other cards go to 4x. It depends. I bought three and gave two Vogons members one each. One member made a video about their build based on that board.

Also, some posts I made about my experience with it as well:

Re: Bought these (retro) hardware today
Re: Bought these (retro) hardware today
Re: Windows 98SE not 4X AGP, but 2X AGP
Re: AliExpress AGP cards

Huh - looks like you guys (who picked up from Alibaba/AliExpress) all sourced the board from IIT (Shenzhen Immediate Information Technology), so that's good.
Funny that you (and others) mentioned that it's a gigabyte clone - from what I've heard ITZR used to manufacture boards in China for Gigabyte, so it could be a case of a subcontractor making a little more, changing some stuff around, and selling it as their own later on. I really don't think that board is meant for "industrial purposes", since Industrial machines typically denotes more solid builds, higher grade caps, longer service life and possibly passive cooling - I don't think any of that applies here (although brokers in that industry likes to rebadge old machines as industrial machines). It is, however, a possibly cost effective way to build a NOS system...

*sigh*. Almost every major company in mainland China likes to pretend to be something that they are not. IIT looked impressive if you saw the (rather blurry) photos of their office/manufacturing facilities on AliExpress...until you notice that they have 11-50 employees (so no way in hell those photos of the blurry buildings in their Aliexpress site would be their facilities), then you look up their website and realize that their contact address is a housing complex next to a subway station on Shenzhen Metro's line 4. Then you delve deeper and notice that their contact address doesn't match between the Chinese and the English version of the site (they are both flats/apartments in housing complexes in Shenzhen). Seems…sketchy.
As my mentor would've put it - sometimes, perception is everything, and people make inferences based on the little things.

ragefury32 wrote on 2021-07-04, 02:43:

Huh - looks like you guys all sourced the board from IIT (Shenzhen Immediate Information Technology), so that's good.

Mine came from

not Alibaba but could be same source as they deal mostly in Gigabyte plus some Asus and Asrock stuff.
Prices have gone up a lot since I bought mine.

Luckybob did a long video of surgery on one of the AGP boards. Replaced the soldered BIOS with a socket for easy replacement. (replacement with socket happens at about 49:30)

https://youtu.be/Zc9c7D5juuM

It’s pretty cool to see the board actually in action and also being a clone of a Gigabyte board. Already got an NOS S370 spare board so I really don’t need it but that’s never stopped me buying anything before…

I bought this bundle yesterday.

https://www.ebay.com.au/itm/235608712940

Pretty cheap I reckon.
I hope the listing is accurate and it comes with the CPU, Heatsink, RAM and GPU though.

What do people think of these chinese boards?

For anyone who has one of these: does it have overclocking options in the CMOS settings?

I bought one of these motherboards last year. I only gave it a quick bench test to see that it worked ok.
There are no overclocking options in the bios, but You can overclock via FSB jumpers. You can't overclock 1 mhz at a time, it's much larger steps.
It is compatible with Mendocino all the way up to Tualatin. I tried a Celeron 300A and it supported 2.0v.
I'm going to setup a system with this in the next week or so. I'll add more information when I have it.

I've come up with a simple but effective Vcore mod for this board. I've really just figured out how to implement what the manufacturers mostly decided not to.
This motherboard, as well as others that use a RC5057 or RC5058 Fairchild DC-DC converter IC, and probably others as well, can be modded for a Vcore increase by simply adding a resistor between
Voltage Feedback pin (VFB) on the converter IC, and Ground.
I'm no electronics engineer, but I suspect pulling the VFB pin closer to ground fools the IC into believing it's not outputting the required voltage set by the VID pins
which are set but whatever cpu is installed, so it bumps up the Vcore voltage in response?
Using a 51 ohm resistor and a Celeron 300A cpu results with 2.25 volts which is 0.25v above the 2.00v set by the cpu, and that's good for at least 463.5 mhz at 103x4.5.
I haven't tried higher but with only 2.1v or so and testing at 450mhz I had to reinstall all drivers as vcore was too low and I got a corrupted registry in win98se.
I've added the RC5057 datasheet as well as the frequency generator IC datasheet used on these motherboards.
I used 51 ohms as a minimum resistance then added a small board with a potentiometer for adjustment and a jumper to enable or disable the mod.
I have a dead Gigabyte 6OXM7E Intel 815 based board that has this 10% "voltage booster" jumper, so I figured out how they implemented that by poking around
with my multimeter and measuring resistances between the jumper and the RC5058 dc-dc ic that it uses.
The frequency generator on this board offers a few more options than the motherboard shows printed on the pcb. Pins 1,2,3,4 equate to FS3, FS2, FS1, FS0 .
A zero or 0 is "on" and a 1 is the same as "X" or "off". It can get a bit confusing. This is the W83194R-39A which offers more frequencies that the 39 without the A.

Some misinformation above that needs fixing. I'm not sure about any other cpus, but with a Celeron Mendocino installed in the socket there is no access to the
lower table of fsb options from the photo in the above post ,as FS3 is locked to zero by the cpu and the motherboard has a zero ohm resistor installed to ensure it stays that way.
Good thing is the resistor can be removed to get those frequencies back. I've currently got the cpu running at 105x4.5 472.2 mhz.
The Gigabyte GA-6VTXE does not have a resistor installed in this spot. Photo attached showing location of resistor to be removed.
There's a chance the auto setup by the motherboard is broken by doing this, but I don't need that feature anyway.