3dfx Archive - AGP.5 Release.

osckhar God Member Offline www.3Dfx.es Posts: 812 SPAIN Gender:	Re: AGP.5 Release. Reply #45 - 01.12.13 at 18:58:33 Thanks! It looks amazing. I love it. I need to test if with new design motherboard detects the adapter 33 or 66MHz with jumper set. I have more ideas but we will see how it works. I would like to build a final revision but times will tell us. Regards, Oscar.
Back to top	3Dfx Site - Museo 3Dfx - 3Dfx Collection.&& IP Logged

trevormacro

God Member

Offline

Voodoo5k6 Gold SE, Legend
& Mystical Edition

Posts: 622
Gender: male

Re: AGP.5 Release.
Reply #46 - 01.12.13 at 19:15:28

osckhar wrote on 01.12.13 at 18:58:33:

Thanks!

It looks amazing. I love it.

I need to test if with new design motherboard detects the adapter 33 or 66MHz with jumper set.

I have more ideas but we will see how it works. I would like to build a final revision but times will tell us.

Regards,
Oscar.

Have a good work Oscar.
When I think it was the adaptor of an amazing man in the past and after that it was mine and with your help it become THIS adaptor.
So nice progress.
When Gold Leader said it will be in bads hands (bads guys). Now I can smile

Back to top

« Last Edit: 01.12.13 at 19:23:35 by trevormacro »

IP Logged

osckhar God Member Offline www.3Dfx.es Posts: 812 SPAIN Gender:	Re: AGP.5 Release. Reply #47 - 01.12.13 at 19:28:26 My friend we are in the dark side! LoL - Oscar.
Back to top	3Dfx Site - Museo 3Dfx - 3Dfx Collection.&& IP Logged

trevormacro God Member Offline Voodoo5k6 Gold SE, Legend & Mystical Edition Posts: 622 Gender:	Re: AGP.5 Release. Reply #48 - 01.12.13 at 20:05:29 osckhar wrote on 01.12.13 at 19:28:26: My friend we are in the dark side! LoL - Oscar. Like the Rampage very dark...
Back to top	IP Logged

osckhar

God Member

Offline

www.3Dfx.es

Posts: 812
SPAIN
Gender: male

Re: AGP.5 Release.
Reply #49 - 03.12.13 at 11:47:58

Hello,

Yesterday had some time free and assembled first prototype AGP3.0. It worked very well even with a V5-6000. Just little hot IC1 but not lock ups, very solid and stable. As well JUMPER for setting 33/66MHz did his job and worked very well.

ASSEMBLED PROCCESS:

SECOND SAMPLE READY FOR ASSEMBLING WITH NEW AGP SOCKET.

I tested it with an ASUS WS with PCI-x support.

V5-6000 AGP3.0 SET 33MHz

V5-6000 3DMARK2001SE - SCORE 33MHz

V5-6000 AGP3.0 SET 66MHz

V5-6000 3DMARK2001SE - SCORE 66MHz

I also tested using a V5-5500 AGP

V5-5500 AGP3.0 ADAPTER 66MHz SET

V5-5500 3DMARK2001SE - SCORE 33MHz

V5-5500 3DMARK2001SE - SCORE 66MHz

Regards,
Oscar.

Back to top

3Dfx Site - Museo 3Dfx - 3Dfx Collection.&&

IP Logged

trevormacro God Member Offline Voodoo5k6 Gold SE, Legend & Mystical Edition Posts: 622 Gender:	Re: AGP.5 Release. Reply #50 - 03.12.13 at 11:59:13 Look very good Congratulations!
Back to top	IP Logged

m14radu God Member Offline Take your chance ! Posts: 555 Timisoara, Romania Gender:	Re: AGP.5 Release. Reply #51 - 03.12.13 at 19:07:39 The adapter is AMAZING !!! Great Work Osckhar !
Back to top	IP Logged

Komponent Senior Member Offline 3dfx modder Posts: 282	Re: AGP.5 Release. Reply #52 - 03.12.13 at 20:41:40 Cool stuff!!! Maybe add in the next REV level shifters so the converter could be safely used in regular 5v PCI slots?
Back to top	IP Logged

osckhar

God Member

Offline

www.3Dfx.es

Posts: 812
SPAIN
Gender: male

Re: AGP.5 Release.
Reply #53 - 03.12.13 at 21:35:31

Thanks at all!

Komponent wrote on 03.12.13 at 20:41:40:

Cool stuff!!!
Maybe add in the next REV level shifters so the converter could be safely used in regular 5v PCI slots?

Adapter does not need any shifter since the card has his powersupply 3.3V. I mean AGP card get 3.3V from powersupply installed in AGP2PCI. It works in both PCI and PCI-x slot 3.3V/5V. Check adapter pins and you will see how 3.3V pins are not routed from PCI_PIN to AGP_PIN.

- Oscar.

Back to top

3Dfx Site - Museo 3Dfx - 3Dfx Collection.&&

IP Logged

Komponent

Senior Member

Offline

3dfx modder

Posts: 282

Re: AGP.5 Release.
Reply #54 - 03.12.13 at 22:07:11

osckhar wrote on 03.12.13 at 21:35:31:

Thanks at all!
Adapter does not need any shifter since the card has his powersupply 3.3V. I mean AGP card get 3.3V from powersupply installed in AGP2PCI. It works in both PCI and PCI-x slot 3.3V/5V. Check adapter pins and you will see how 3.3V pins are not routed from PCI_PIN to AGP_PIN.
- Oscar.

I'm sorry, I have great respect for you, but this time you are wrong.
It's not about the power supply, it's about the signaling. The motherboard sends data in 5V logic to a 5V PCI slot. Yes, adding a power supply for VDDQ does make the video card to send data in 3.3V logic, but it will still receive 5V signals from the motherboard. That's the whole point why AGP2PCI has level shifters installed. Sure, it might work for the moment but it is not safe. Don't have to trust me, the creator of AGP5, REN, has warned about not using AGP5 in 5V PCI slots:
http://translate.google.com/translate?sl=cs&tl=en&js=n&prev=_t&hl=ro&ie=UTF-8&u=...
Quote:

What's the thread from Loeschzwerga who linkujes is bullshit. 5V mobo AGP.5 REALLY detonates card, I told you already 10x's host.

Back to top

« Last Edit: 03.12.13 at 22:08:06 by Komponent »

IP Logged

osckhar

God Member

Offline

www.3Dfx.es

Posts: 812
SPAIN
Gender: male

Re: AGP.5 Release.
Reply #55 - 03.12.13 at 22:48:59

@komponent,
You dont need apologize. My electronic knowlege is very few. Embarrassed

Question- If adapter takes the voltage from his power supply for sending data 3.3V to the VGA instead of 5V from PCI socket. How card can be damaged?

In the power rail 5V from PCI slot there are data needed for the well working of the VGA?
I understand V5 rail is only a voltage rail to power up pci device. If adapter omits it and take voltage for power up the vga from itself power supply... I dont see the problem, but I am not engineer electronic and you know a lot of more about this than me.

Thanks,
Oscar.

Back to top

3Dfx Site - Museo 3Dfx - 3Dfx Collection.&&

IP Logged

Komponent

Senior Member

Offline

3dfx modder

Posts: 282

Re: AGP.5 Release.
Reply #56 - 04.12.13 at 05:59:58

No, no, no, again, it's not about the power supply, it's about the logic level:
http://en.wikipedia.org/wiki/Logic_level
See it that way: you have AD[x], C/BE[x], and the other signal lines that are connected directly from PCI to AGP through PCB traces. If the motherboard wants to send a logic "1" to any of these lines, it will put it "high", meaning at (ideally) close to 5v (for logic "0" it will be low, at close to 0v).
That is no problem if the target is 5v input tolerant (Avenger GPUs should be OK, IIRC), but VSA100s were designed for 1.5v and 3.3v logic only. Nobody can estimate how will they tolerate being operated in 5v signaling, especially in long term.
A good hint is: Look at Voodoo4/5 PCI cards. They all have level shifters installed to convert from 5v to 3.3v logic (IDT quickswitch parts near the PCI edge). So it is safe to assume that VSA100 are NOT 5v input tolerant GPUs.
AGP2PCI also has level shifters installed. As a side note, it also has isolated power supply for 3.3v and VDDQ, like AGP5 does.

Back to top

IP Logged

osckhar

God Member

Offline

www.3Dfx.es

Posts: 812
SPAIN
Gender: male

Re: AGP.5 Release.
Reply #57 - 04.12.13 at 09:41:43

Komponent wrote on 04.12.13 at 05:59:58:

No, no, no, again, it's not about the power supply, it's about the logic level:
http://en.wikipedia.org/wiki/Logic_level
See it that way: you have AD[x], C/BE[x], and the other signal lines that are connected directly from PCI to AGP through PCB traces. If the motherboard wants to send a logic "1" to any of these lines, it will put it "high", meaning at (ideally) close to 5v (for logic "0" it will be low, at close to 0v).
That is no problem if the target is 5v input tolerant (Avenger GPUs should be OK, IIRC), but VSA100s were designed for 1.5v and 3.3v logic only. Nobody can estimate how will they tolerate being operated in 5v signaling, especially in long term.
A good hint is: Look at Voodoo4/5 PCI cards. They all have level shifters installed to convert from 5v to 3.3v logic (IDT quickswitch parts near the PCI edge). So it is safe to assume that VSA100 are NOT 5v input tolerant GPUs.
AGP2PCI also has level shifters installed. As a side note, it also has isolated power supply for 3.3v and VDDQ, like AGP5 does.

Now understand what you mean. I was thinking the problem was about to give the right voltage to the card 3.3V/5V but the problem is about signals data between PCI socket & VGA card. In a PCI-x slot signals are sending 3.3V but in a normal PCI slot signals data are sending 5V.

Just not sure if is a worth to add shifter as used on original adapter or in VSA-100 pci family since the idea with this adapter is take advantage from 66MHz bus instead of 33MHz.

I will check it but I guess I will need to change all the traces or design for adding shifter.

Thanks,
Oscar.

Back to top

3Dfx Site - Museo 3Dfx - 3Dfx Collection.&&

IP Logged

Loeschzwerg

Senior Member

Offline

AAlchemy PSU maker :-)

Posts: 396
Germany -> Bavaria -> Augsburg
Gender: male

Re: AGP.5 Release.
Reply #58 - 04.12.13 at 09:56:53

Komponent wrote on 03.12.13 at 22:07:11:

I'm sorry, I have great respect for you, but this time you are wrong.
It's not about the power supply, it's about the signaling. The motherboard sends data in 5V logic to a 5V PCI slot. Yes, adding a power supply for VDDQ does make the video card to send data in 3.3V logic, but it will still receive 5V signals from the motherboard. That's the whole point why AGP2PCI has level shifters installed. Sure, it might work for the moment but it is not safe. Don't have to trust me, the creator of AGP5, REN, has warned about not using AGP5 in 5V PCI slots:
http://translate.google.com/translate?sl=cs&tl=en&js=n&prev=_t&hl=ro&ie=UTF-8&u=...
Quote:

What's the thread from Loeschzwerga who linkujes is bullshit. 5V mobo AGP.5 REALLY detonates card, I told you already 10x's host.

Well, REN is right about his comment about my post here:
Loeschzwerg wrote on 14.03.12 at 23:27:33:

The adapter will work in PCI 5V and 3.3V slots Wink

Why? Simply because the I/O voltage pins aren't used within Ren's adapter design Wink

Take a look at the pinout:
http://pinouts.ru/Slots/PCI_pinout.shtml (difference between 3.3V and 5V PCI)

And now take a look at the eagle layout of Ren's adapter. The adapter does have it's own powersupply and it generates the 3.3V (needed for 3dfx cards) out of 5V provided either over a molex or via the PCI bus. No matter in what slot type you put that adapter in, the adapter always will support the AGP slot with it's own power supply.

So no problem from this side, the adapter won't kill your card Wink

...but because I wasn't precise enough. He is right when he says that a PCi 5V only board will fry the card (or it simply doens't work/boot).

BUT! Most (if not all) more modern mainboards, that are PCI2.1/2.2 compliant will work with the selfmade adapter and don't harm the card. Why? Simply because the southbridges are designed to work with 3.3v cards even when they are keyed for 5V.

Want an example?
http://uk.ts.fujitsu.com/rl/servicesupport/techsupport/boards/Motherboards/Micro...

Quote:

Supports 3.3V/5V PCI bus Interface

Now, why does the original AGP2PCI adapter and the V4/V5 cards to have those signal shifters? Simply to be compatible to all older boards and configurations.

Back to top

« Last Edit: 04.12.13 at 10:03:15 by Loeschzwerg »

IP Logged

goriath God Member Offline Ruggedized Retro Gamer Posts: 599 Italy/Slovenija Gender:	Re: AGP.5 Release. Reply #59 - 04.12.13 at 14:37:48 Does this imply that the power supply circuitry on the AGP.5 adapter is redundant?
Back to top	« Last Edit: 04.12.13 at 14:38:07 by goriath » OLD SCHOOL: Because it was done right the first time IP Logged

	Welcome, Guest. Please Login