TPU's Nostalgic Hardware Club

[MachineLearning]

Another revision I guess, it also has a slightly different PCB. It had that cooler when I bought it (and the other one) used.

Do you have shots from underneath the cooler perchance? Mainly I'd be curious to see the PCB number which I believe is this here

That's my PCB, the reference is the same string except minus the "P" at the end

 

[Ruru]

Do you have shots from underneath the cooler perchance? Mainly I'd be curious to see the PCB number which I believe is this here
View attachment 235289
That's my PCB, the reference is the same string except minus the "P" at the end

Sorry, too lazy to tear those down as it's not that long ago when I did that and repasted both.

 

[Dinnercore]

Been playing with the MSI P45 Platinum and I gotta say... I completely understand why this boards FSB world record on HWbot is only at 511 MHz. The process of overclocking this board is the worst I have ever experienced. The bios settings regarding GTLs were made by an insane person. No other way to put it. And the way this board behaves makes it a good addition to a dadaism art exhibit.

The good thing, it has a very solid power management and monitoring. All values set in bios are very close to the actual real values applied, very little droop, overshoot or undershoot. For example 2.00V DRAM results in 2.03V actual voltage. Compared to e.g. the Asus P5Q-E which has 2.07V actual voltage when setting 2.00V in bios.
This graph is from Vcore during the start of a Cinebench run. Small ~15mV droop:

Of course my DMM is not fast enough to pick up the quick transient loads, maybe with an oscilloscope I could see those.

Back to overclocking...
Let's start with the bios options (those are the same between bios versions):
Auto settings are very aggressive. RAM timings and speeds get set according to SPD information on the sticks maybe? But the board fails to apply the correct voltage for these settings, so it sometimes tries to run sticks at 5-4-4-12 - DDR2 1200 but at 1.8V... And fails every single time, not understanding that it has to use safe defaults.
Setting voltages and timings yourself and its fine, unless you want to tune your GTLs. Oh boy.

The GTL settings are very important to stabilize or even enable FSB speeds beyond 500Mhz according to my 775 experience. They are reference voltages that have to be set according to the VTT and are sometimes shown as a multiplier like x0.67 or x0.63 and sometimes as absolute voltage like 0.756V (=1.2V VTT x0.63). Asus likes the multipliers, Gigabyte uses the voltages.
Now MSI also uses voltage numbers, so you have to calculate yourself where you want to end up. BUT the options for these voltages make ZERO sense.

First of all, you can not enter the values via numbers, no you gotta use '+' and '-' or page up and down to scroll through ALL the options. Either from the max value or the min value. BEWARE it makes a difference which way you scroll, because when you use '-' you get a different range of voltages compared to when you use '+'.
For some reason it also switches the steps right in the middle, from 1.6mV steps it suddenly shifts to 0.5mV steps. Now thats all fun and games already, but lets add another complication they thought at MSI. If you scroll over to the end with '+' and roll over from max. values, they don't go to the min. value. It just starts back somewhere up top in the 0,700V range. You can only access the minimum values if you use '-', roll over from a random mid-point and suddenly find yourself at the max. value from where you NOW can use '+' to suddenly being able to enter 0.200V. Which is SO far down that it makes zero sense to ever enter values this low.
And if you use '+' to roll over, you can never get back to 'auto' settings ever again. You also can not enter 'auto', you have to reset ALL settings in bios to default in order to set the GTL field back to auto.
Oh but the fun is not over. These voltages are linked to VTT and respectivly the range changes up and down with VTT. Here is the big deal why this board is impossible to OC:

At 1.24 VTT the maximum GTL ref value you can set for CPU is 0,742V which leads to a multi of x0.598. Standard for these CPUs is x0.67 and they usually run well at x0.63 for FSB OC. The MAX multiplier value of x0.598 is not a stable setting. Not for both GTLs. You can set one usually very low for C2Ds but I had no luck with running GTL1 below x0.61 ever.
Same for the MCH GTL ref. It makes no sense, the highest value is below what all my other P45 boards run stable at. WHO made these settings? They also violate Intel spec in every way.

Enough rant about the bios settings. Maybe I'll find something that works, so far I hit a wall at 475Mhz FSB and touching any GTL settings is an instant no post.

Which brings me to the strange antics of this board posting. Sometimes it just freezes on CPU initialization after changing bios settings. This is normal and not a fault of your settings, in that case you need to hold down 'RESET' (not power!) for 4 seconds and it cycles power, posting fine with your new settings.
It sometimes gets stuck on RAM initialization, again this is more often not a problem with your RAM but the board just needs a full power reset. Meaning PSU off, wait until all caps have drained. Power back on and it suddenly accepts your RAM.
You smashed 'DEL' as fast as you can, but it just opens the Windows bootmanager? Oh yeah, the time window to enter bios is less than half a second.

I kinda want to just toss it in my collection and never touch it again, but on the other hand I want to see how far I can push this thing. It's very time consuming as I have to wrestle with it after every simple crash.

 

[Ruru]

Back in the day MSI boards truly sucked. I stayed away from them until AM4 X470 Gaming Plus (and that was finally a good board).

I had a P6N SLI Platinum (LGA 775, NF650i SLI) and that was okay, but as we remember, nForce wasn't the best which came to FSB overclocking. Then I had a P67 (can't remember the exact model) and the bios was hella un-user friendly to use.

 

[stinger608]

Back in the day MSI boards truly sucked. I stayed away from them until AM4 X470 Gaming Plus (and that was finally a good board).

I had a P6N SLI Platinum (LGA 775, NF650i SLI) and that was okay, but as we remember, nForce wasn't the best which came to FSB overclocking. Then I had a P67 (can't remember the exact model) and the bios was hella un-user friendly to use.

I don't know man. I've had an MSI 1150 gaming mATX board for about 5 or 6 years with a 4790K chip in it, and it has been rock solid.

Hell, I remember back in the socket A days having many MSI boards and I think only one of them shit the bed. I've personally had great luck with MSI boards.

 

[Ruru]

I don't know man. I've had an MSI 1150 gaming mATX board for about 5 or 6 years with a 4790K chip in it, and it has been rock solid.

Hell, I remember back in the socket A days having many MSI boards and I think only one of them shit the bed. I've personally had great luck with MSI boards.

I mean by overclocker's point of view, MSI boards were pretty ass back in the day. For example that 775 board I had was a good board in overall, but just for high FSB overclocking, it sucked. :/

edit: But I have to say that MSI K8N Neo2 Platinum (S939, NF3 Ultra) was a great board, overclocked like a dream with a 3000+ Venice.

 

[MachineLearning]

NIB FirePro 2270 for <$20USD shipped, why not... It's basically a HD 5450 but called something else. I gave it a Noctua 40x10mm fan, aluminum heatsinks, MX-2 and APT2560 pads.

It doesn't support overclocking natively, so I'll have to do some VBIOS fiddling... it just won't be as easy as increasing the sliders like most consumer AMD/ATI cards of the era.

 

[Kanan]

View attachment 235363
View attachment 235364
View attachment 235365
View attachment 235366
View attachment 235367
View attachment 235368
View attachment 235369

NIB FirePro 2270 for <$20USD shipped, why not... It's basically a HD 5450 but called something else. I gave it a Noctua 40x10mm fan, aluminum heatsinks, MX-2 and APT2560 pads.

It doesn't support overclocking natively, so I'll have to do some VBIOS fiddling... it just won't be as easy as increasing the sliders like most consumer AMD/ATI cards of the era.

These red pcbs look really good.

 

[Greenslade]

Ha, not quite "the best I can do", no ... but back in a day it served its purpose

I had a weird tonearm resonance issue while playing LP album "The Wall", which couldn't be explained through common sense or hardware/tech components themselves. So I messaged Audio Technica representatives who asked me to make a video on the issue ... which is why & how the video was made. Oddly enough, the problem disappeared over time which makes me think that it may have been related to (then) brand new AT-95EX cartridge & its stylus suspension. At the time, it was still very rigid & apparently couldn't handle all the warps & wrinkles on the record, so the stylus assembly plowed through all of them & picked up every single vibration.

Think of it as shock absorbers on your car. Every car has suspension system so that it could drive over bumps & potholes without transferring all those vibrations onto the passengers inside. Thus, on a "normal" stylus assembly, you will have a rubber suspension system which allows the stylus to track bumps & vibrations without transferring any of them to the tonearm. However, since the suspension was too firm, it experienced resonance which made the whole tonearm shake out of control. No harm done, although in more severe situations resonance CAN destroy records and/or even the stylus itself.

So there, hopefully this answered your question(s)

AT-95EX was a good budget cartridge back in the day a lot of us had one, at one time or another Very true there.A cool looking wooden arm?.

---

 

[sam_86314]

This is hardware, right? It's an old computing device...

One of my thrift store finds from many years ago; an Omron 86R calculator circa 1974 with a lovely little vacuum-fluorescent display.

Still has the old batteries it had when I found it. They "expired" in 2015, yet they're still going strong.

Also has a fun little glitch where if you subtract a number from zero and re-add it, you'll get "-0".

 

[INSTG8R]

I’ve probably posted this before but still think it’s a beautiful card… I also realized its my only fallback card which wouldn’t be as beautiful. I have a 7970 but it”sdied of sag., didn’t turn over the last time I tire$ it…

 

[MachineLearning]

XFX 750i motherboard, as you can see... I just got this today so I haven't had time to play around with it except to make sure it works, but I already know it's gonna be an odd duck...

So judging by 1.) the sticker on the front that says "Designed by nVidia Authorized Board Partner," 2.) the fact that this breaks significantly in design from the EVGA 750i [itself being strictly reference, as far as I know], and 3.) what I remember reading, this is a board designed by XFX's team. That itself makes it interesting to me, but also it's weird that they went through the trouble of designing a dual heatpipe cooler with black fins, logos all over the place, nickel plating for said heatpipes, but then decided not to cool the fourth power stage? I seriously don't understand this decision, it isn't like newer boards where that would be IGP phase(s) and therefore more understandable. But I'm very confused by this design choice. I think they must have had their GPU team make this in their free time Lot of things make sense and/or are fine, but some stuff is just confusing.

I gave the southbridge and VRM some Arctic APT2560 0.5mm pads, the northbridge and NF200 chip both got MX-2, and the shunned power stage got 2 small aluminum heatsinks. Also, the thermal... substances on this cooler were absolutely rock solid. There is no getting these off via any mortal means. I need like industrial grade solvent, or maybe just some gumption, but I have neither and I like to live on the edge. So I did the best that I could, which wasn't much, and if the board complains then oh well, it still works and I can always try again...

XFX... why is the POST code on the rear I/O? I mean, I'd rather have this than no post code at all. And if it were in the traditional position, you may not even be able to see the POST code at all if you have multiple GPUs, so I suppose that's a plus... But how could anyone reasonably read this if it's installed in a case? If it were on a test bench it'd actually be the best position imo, but clearly that wasn't their intention, since unlike the EVGA 750i this doesn't have onboard power / reset buttons.

While researching this board, I found this laughable press release from XFX back in the day. Everyone at TPU hated it... I think that marketing guy either wasn't involved in the box design (thank god), or they axed him before he got his hands on it. I'm more than fine with that, this box looks pretty sweet in my opinion.

XFX, very conveniently and taking after the style of Foxconn, decided that publicly hosting like 1GB of downloads for their older consumer mainboards is just way too damn expensive. So it's impossible to download any BIOS updates directly from their site, they took down those pages. Well, at some point, some person in 2014 wanted to update their XFX 750i BIOS on Tomshardware but posted about how they couldn't find it. A XFX representative actually responded helpfully with what seems like an internal(?) link to all of their motherboard BIOSes. Seriously, what a great resource... I know their boards weren't popular or really all that great, but to have these organized and technically still on the internet is great. The link to that is here, BIOSes in the Motherboard folder. Hopefully anyone else who would need one of those files can come back to this. Flashing can be done either through DOS or AWDFLASH, DOS of course being recommended. I didn't upload those BIOSes but considering the context, they appear legitimate. Mine was already fully updated luckily, so I didn't have to meddle with this hidden science.

I know I'm bashing on this board a lot, but I wanna make it clear I'm still happy with it no matter what because it's funky. If the hardware is funky, it gets a pass from me. I'm gonna do some overclocking on this after I do a few other projects and see if it's as unstable as I've read online

 

[Kanan]

View attachment 235566
View attachment 235567

View attachment 235568
View attachment 235569
View attachment 235570

XFX 750i motherboard, as you can see... I just got this today so I haven't had time to play around with it except to make sure it works, but I already know it's gonna be an odd duck...

So judging by 1.) the sticker on the front that says "Designed by nVidia Authorized Board Partner," 2.) the fact that this breaks significantly in design from the EVGA 750i [itself being strictly reference, as far as I know], and 3.) what I remember reading, this is a board designed by XFX's team. That itself makes it interesting to me, but also it's weird that they went through the trouble of designing a dual heatpipe cooler with black fins, logos all over the place, nickel plating for said heatpipes, but then decided not to cool the fourth power stage? I seriously don't understand this decision, it isn't like newer boards where that would be IGP phase(s) and therefore more understandable. But I'm very confused by this design choice. I think they must have had their GPU team make this in their free time Lot of things make sense and/or are fine, but some stuff is just confusing.

I gave the southbridge and VRM some Arctic APT2560 0.5mm pads, the northbridge and NF200 chip both got MX-2, and the shunned power stage got 2 small aluminum heatsinks. Also, the thermal... substances on this cooler were absolutely rock solid. There is no getting these off via any mortal means. I need like industrial grade solvent, or maybe just some gumption, but I have neither and I like to live on the edge. So I did the best that I could, which wasn't much, and if the board complains then oh well, it still works and I can always try again...

XFX... why is the POST code on the rear I/O? I mean, I'd rather have this than no post code at all. And if it were in the traditional position, you may not even be able to see the POST code at all if you have multiple GPUs, so I suppose that's a plus... But how could anyone reasonably read this if it's installed in a case? If it were on a test bench it'd actually be the best position imo, but clearly that wasn't their intention, since unlike the EVGA 750i this doesn't have onboard power / reset buttons.

While researching this board, I found this laughable press release from XFX back in the day. Everyone at TPU hated it... I think that marketing guy either wasn't involved in the box design (thank god), or they axed him before he got his hands on it. I'm more than fine with that, this box looks pretty sweet in my opinion.

XFX, very conveniently and taking after the style of Foxconn, decided that publicly hosting like 1GB of downloads for their older consumer mainboards is just way too damn expensive. So it's impossible to download any BIOS updates directly from their site, they took down those pages. Well, at some point, some person in 2014 wanted to update their XFX 750i BIOS on Tomshardware but posted about how they couldn't find it. A XFX representative actually responded helpfully with what seems like an internal(?) link to all of their motherboard BIOSes. Seriously, what a great resource... I know their boards weren't popular or really all that great, but to have these organized and technically still on the internet is great. The link to that is here, BIOSes in the Motherboard folder. Hopefully anyone else who would need one of those files can come back to this. Flashing can be done either through DOS or AWDFLASH, DOS of course being recommended. I didn't upload those BIOSes but considering the context, they appear legitimate. Mine was already fully updated luckily, so I didn't have to meddle with this hidden science.

I know I'm bashing on this board a lot, but I wanna make it clear I'm still happy with it no matter what because it's funky. If the hardware is funky, it gets a pass from me. I'm gonna do some overclocking on this after I do a few other projects and see if it's as unstable as I've read online

That's a great find. Well I'm not surprised you're bashing it a lot, they simply had not enough experience designing mainboards and its way harder than doing GPUs. XFX is a funny brand, not in a negative sense, just when you observe their history, it's a up and down story. At the moment they're doing pretty well and I like most their new designs a lot.

 

[INSTG8R]

Geez I completely forgot about these. Totally complete minus the cables. I was really lost when Abit disappeared I basically basically always used them from zP4 up to C2D

 

[Kanan]

Geez I completely forgot about these. Totally complete minus the cables. I was really lost when Abit disappeared I basically basically always used them from zP4 up to C2D
View attachment 235605
View attachment 235606

What happened to ABIT anyway?

 

[INSTG8R]

What happened to ABIT anyway?

They basically disappeared without a teak reason. I was for quite awhile hosting their entire BIOS collection via Torrent.
All that’s left…

Universal abit > Motherboard, Digital Speakers, iDome, AirPace, Multimedia

Universal abit > Motherboard, Digital Speakers, iDome, AirPace, Multimedia

abit.ws

 

[Ruru]

I miss Abit the most of the manufacturers that isn't in the game anymore. In general they made great boards with usually great overclocking capabilities.

 

[Kanan]

I miss Abit the most of the manufacturers that isn't in the game anymore. In general they made great boards with usually great overclocking capabilities.

same with DFI and their LAN PARTY boards

 

[Ruru]

same with DFI and their LAN PARTY boards

Yeah, I remember those, especially the NF4 Lanparty boards were pretty common back in the day.

 

[INSTG8R]

I miss Abit the most of the manufacturers that isn't in the game anymore. In general they made great boards with usually great overclocking capabilities.

That AMD Board had 2 16x slots way back then! Built my first “crazy rig” with it…

TechPowerUp

www.techpowerup.com

 

[Ruru]

That AMD Board had 2 16x slots way back then! Built my first “crazy rig” with it…

TechPowerUp

www.techpowerup.com

Gotta love that old-school custom loop.

 

[Kanan]

That AMD Board had 2 16x slots way back then! Built my first “crazy rig” with it…

TechPowerUp

www.techpowerup.com

This is "old school cool". Who knows it appreciates it

 

[INSTG8R]

This is "old school cool". Who knows it appreciates it

That External Unit was actually Asetek. Back when they actually made stuff and weren’t just patient trolling. It had six crazy/server fans I replaced with all Silen-X another company that’s gone
And I still have the X1900XTX Master Card
Actually I couldn rebuild that whole rig again on air.

 

[Kanan]

That External Unit was actually Asetek. Back when they actually made stuff and weren’t just patient trolling. It had six crazy/server fans I replaced with all Silen-X another company that’s gone
And I still have the X1900XTX Master Card

I had a 92mm fan by Silen-X for my old Chieftec case, quality and not expensive.

And I still have the X1900XTX Master Card

I missed all gens between Geforce 3 and Geforce 7 (and Radeon equivalent) because I didn't game anymore for a while, so I had a GF 3 Ti 200 for a long time and then switched to GF 7800 GT... it's like you're reminding me of a whole in my head

 

[Greenslade]

Can anyone give me some advice i have this Thin client

i want to upgrade the flash dom.Do i go for the 8GB one NMB one i have never heard of that make before, that would be £16.84.Or do i go with the more common Apacer 4GB which is £13.30 .The extra storerage would be nice for just £3.54 for the extra 4gb.?I will be using Windows 98 .Do any of you use a thin Client for Retro gaming?Both are from the same seller. The unit is cheap but the storage isn'tI have just looked up NMB they made memory chips for Intel until they terminated there agreement with them in 1994.