Safe to mount a G400MAX on a i845D chipset?

[Haig]

John suggested that we shouldn't have included the 4x notch on our 2x boards.

In turn, I said if we did that, we would not work in 4x only motherboards, which we do.

There's a difference between fitting the card in the slot based on the notch(key) and what voltage is applied to the card depending on a pin that's depending on what connector type the motherboard is going to use. That's why I said these are 2 different issues.

First, the notches have absolutely nothing to do with the way Universal slots determine AGP signaling level.

Who's talking about a universal slot? This asus motherboard doesn't even have a universal slot becuase if it did, then there would be no problem since a universal slot is supposed to accept both signals and will supply the voltage accordingly depending on typedet.

They rely exclusively on TYPEDET# to determine signaling level.

Agreed, but we're not talking about a motherboard with a universal slot.

JohnK was stating that since the first-run G400 MAX cards are wired up with TYPEDET# open, they shouldn't have a 1.5v notch. He's right.

Yes he is and I never said otherwise. Not that it means anything but using the same pcb on a 1.5v ONLY motherboard that doesn't have this typedet validation works. I've listed a few of the motherboards on the thread in our forum.

Haig

[IceStorm]

drop back two steps and PUNT!

Ice do you have a clue as to how any electronics design and specs apply to reality... I honestly don't think so.

So because I can read a spec, but don't work in the industry, I'm not qualified to determine that JohnK's correct? Uh-huh...

The spec formed while the G400 was in late stages of development. It's not a "theory we're going with."

So Matrox sat on the G400 chip and board design for over a year, based on your thinking. Spec finished May, 1998. Card launched May, 1999, or thereabouts. While I'm not saying its out of the question, it would explain why Matrox's product cycle seems to be so long.

I also don't know how you deemed Matrox's implementation "violating the wrong way." I checked, there's no appendix to the spec on "How to Correctly Violate the Spec."

Spec states that two types of slots check TYPEDET#. It doesn't state that checking TYPEDET# is an option for the 3.3v. In that light, any card that's going to be a dual mode card should set TYPEDET# to ground, not open. That's how I interpret the spec. This also seems to be the way nVidia card makers interpreted it as well as later Matrox designs. There's a list of cardsthat Van's Hardware tested at the bottom of the page. While he's not the most unbiased person around, I trust he can use a multimeter.

And being smart enough to read a spec doesn't excuse a combatant, shitty, arrogant, personality, but you keep going anyway.

I'm not taking any of this personally. I'm also not going to roll over and agree that Matox isn't "wrong" in this case.

John suggested that we shouldn't have included the 4x notch on our 2x boards.

In turn, I said if we did that, we would not work in 4x only motherboards, which we do.

No, that's not what you said. You said:

Why would we take out the 4x key? This would mean that it would not work in a 4x system which it does. Don't confuse the agp notch with the voltage becuase these are 2 separate issues you are talking about.

The AGP notch MUST match TYPEDET#. Find me a place in the spec, anyone, where it says they don't have to match. As per spec you can't divorce the two from each other as you seem to be suggesting above.

The old G400 MAX card rev works with both, but it violates spec when it comes to TYPEDET#. That's possible, and you and I have already proven it's possible. But that doesn't make it right.

Who's talking about a universal slot?

You were, Haig, when you said:

Follow me here The 4x indication notch will allow the card to fit in a 4x or a universal agp slotted motherboard. Typedet has nothing to do with this.

hence my reply.

[Greebe]

What you decided this all was about is a personal vendetta against anyone whom makes any statement against yours... no proof is good enough, no matter how knowledgeable we are on the subject or in the industry.

How sad

[IceStorm]

no proof is good enough, no matter how knowledgeable we are on the subject or in the industry.

Not true. I'd like proof that somewhere in black and white Matrox wasn't wrong to do what they did when they wired up the board for 3.3v. Find me something in the specs that say it's ok to do this with a card that has a notch for 1.5v operation.

So far, all I get is people saying that Matrox tried to overcome a problem with the spec. Along with that, I get the impression that you're agreeing what Matrox did was appropriate. But what I don't see is anyone posting a thought process that would lead to this conclusion. You're saying it's due to the spec not being finalized, but that doesn't make much sense to me - you're saying it's ok to design to a draft spec and not validate it against the final spec after the spec's released and well before the product goes to market? Exactly how can that be considered acceptable?

[Greebe]

Because noone who took part in the design process of the original G400's cares to appease your morbid desire to prove it to you at this point. You are akin to a smartass punk kid whom without the forsight or sensibility to know when you've killed any chance finding why all this is with your degrading attacks against those whom at least once upon a time were attempting to help you see the forest from the trees.

[IceStorm]

You are akin to a smartass punk kid whom without the forsight or sensibility to know when you've killed any chance finding why all this is with your degrading attacks against those whom at least once upon a time were attempting to help you see the forest from the trees.

You sound like a jaded engineer who feels he doesn't have to bother explaining his rationale when confronted with the errors in his logic after he's made his first attempt at explaining the situation.

You gave a chronology. You were off by a year. I pointed this out. You've never gone back to explain that missing year. If you can't/don't want to, that's fine, but don't expect me to take that as a valid reason for Matrox having defective card designs.

[Greebe]

Oh that's right I didn't did I

OK I was off by a year... big flyin monkey poop! I knew that didn't sound right when I posted it but it was late and said what the heck and posted anyway, then crashed.

What a putz

[Greebe]

I didn't design the G400... think you need Miss Cleo for this cause all avenues from getting the answer in the specific form you want it in is now a burnt bridge with you on the wrong side.

[IceStorm]

I didn't say you didn't reply to it, just that you've not gone back to explain the missing year...

I am not assuming you designed the G400. I honestly don't know/remember who you work for.

I "burnt bridges" ? I'm defending a poster in a Tech Support forum from the ridicule of the jaded support people. And for that I've "burnt bridges" ?

[KevinST]

It seems to me that the arguments are revolving around whether Matrox released the early PCB's without testing against AGP 4x (hence TYPEDET open) and then when they did test against AGP4x they fixed the issues to make the card AGP 4x, thus closing the TYPEDET pin (I can't quote anything to back this up - it's the overall feeling I have of the direction of the discussion). This is incorrect as Matrox have 2 PCB's 906 01 and 906 02, both of which are non AGP 4x cards, however TYPEDET is open on the 906 01 card and closed on the 906 02 card.
So, with this in mind, I still maintain that Matrox intended the TYPEDET pin to be closed from day one and that the origonal PCB B]906 01[/B] had either a design mistake or a manufacturing mistake in it which was fixed in the very next revision.
I've worked in design and tech support in Security Systems, microprocessor development tools and network equipment... and I've know of this to happen in every single industry.


I appreciate what MURC has done for Martox users, and the knowledge that's contained within the minds of the forum, and I also understand your need to defend the Matrox products. However, I'm getting a little fed up with the arrogance of some of you.
In the thread I started back in December about this very issue (http://forums.murc.ws/showthread.php?s=&threadid=30491 I was flamed for even suggesting that TYPEDET should be open, that Matrox had the spec right, and Asus had it wrong. Even at the beginning of this thread we had this

>the TYPEDET pin should be connected to ground on all G400's.

You know, if you want to bring that up again, the spec hasn't changed. It still says floating. If Matrox follows the spec, and the MB mfr doesn't, then it's not Matrox's fault.

Please realise that it is possible for non "Super MURC" and "moderators" to be right.

[KevinST]

Greebe, did I say it was you who said TYPEDET should be open???? look back over this post and the other one I linked to - you'll see plenty "Super MURCers" telling me that I'm wrong.

Oh - do you really think it's mature and correct to give details of another posters work details??? I would suggest you consider editing your post to make it a little more professional - after all, I thought the idea of moderators was to be responsible???

Edit: Greebe - I'm glad you (or someone else) removed your post.

[Haig]

I'm defending a poster in a Tech Support forum from the ridicule of the jaded support people.

It's unfortunate that you feel I wanted/had (whatever tense you want to use) to ridicule our own client on our public forum. None the less, that's your opinion and you can keep it.

If John felt/feels that/this way, I'm sure he would've/would let me know. He doesn't need someone else to come to his rescue.

No, that's not what you said. You said:

quote:
--------------------------------------------------------------------------------
Why would we take out the 4x key? This would mean that it would not work in a 4x system which it does. Don't confuse the agp notch with the voltage becuase these are 2 separate issues you are talking about.
--------------------------------------------------------------------------------

Yes that is what I said and I will say it again. Why would we not include/take out (whatever tense/text you want to use) the 4x notch? The 2x board works in a system, that has a 4x key/notch.

quote:
--------------------------------------------------------------------------------
Who's talking about a universal slot?
--------------------------------------------------------------------------------

You were, Haig, when you said:
quote:
--------------------------------------------------------------------------------
Follow me here The 4x indication notch will allow the card to fit in a 4x or a universal agp slotted motherboard. Typedet has nothing to do with this.
--------------------------------------------------------------------------------

hence my reply.

You first mentioned it on our forums here on your post dated January 29th:

http://forum.matrox.com/mgaforum/Forum1/HTML/007876-2.html

Ever since then, you keep bringing it up, hence my reply.

In any case, our cards have a universal connector, the voltage is determined by typedet. Hence, the keying/notches are not the determining factor for the voltage. This is made public on Intel's site.

Regading the early G400 boards, I will give people an update on our thread sometime this week.

Haig

[Rags]

Reading the spec, it's obvious they wanted to use the type det to allow a universal connector to determine what type of card was in the slot.

Now, it had no foresight that a motherboard company would put that det in a 1.5v only design to protect from instability (that's why it's there, not because of frying a motherboard slot).

What Matrox did was make a universal card that would detect what voltage the slot was, and then switch to that mode automatically. Ironically, the G400 is the only card of its time to do this and as such is more compatible in more slots than any of the competing products.

Now to say that an early PCB with the type det open is a defect or is wrong, it's not. The original spec said that if the card could operate in 3.3, it would leave that open. Later, the spec was ammended to remove that part, and the later PCB's on the matrox cards reflect that change.

Rags

[KevinST]

1.5v only design to protect from instability (that's why it's there, not because of frying a motherboard slot).

Contradicts Asus and a few other sites that state that if a 3.3v only AGP card is placed in a AGP slot connected to a 1.5v only chipset, the motherboard is damaged.

Now to say that an early PCB with the type det open is a defect or is wrong, it's not. The original spec said that if the card could operate in 3.3, it would leave that open. Later, the spec was ammended to remove that part, and the later PCB's on the matrox cards reflect that change.

Contradicts both what Haig has said in the past (he's even advised that customers can ground TYPEDET themselves on cards where it is not grounded), ... also, do you have a copy of the original spec, and the currently available spec to show the differencees???

[Rags]

Originally posted by KevinST

Contradicts both what Haig has said in the past (he's even advised that customers can ground TYPEDET themselves on cards where it is not grounded), ... also, do you have a copy of the original spec, and the currently available spec to show the differencees???

Not only do I have a copy of the original, but I have a copy of the ammended notes to correct for uninversal cards.

Rags