VOGONS


First post, by W.x.

User metadata
Rank Member
Rank
Member

Hello, I want to know, where is approximately DTK socket 7 motherboard in quality on scale:

Where on TOP are Asus, FIC, Shuttle

In middle Gigabyte, Soyo, Soltek , Biostar, MSI , QDI, etc.

On bottom
Amptron, ECS, Matsonic
PC Chips

These are DTK motherboards
http://www.motherboard.cz/dtk.htm

Are they decent? More like 5/10? Or more like 3/10?

Reply 1 of 16, by mpe

User metadata
Rank Oldbie
Rank
Oldbie

Not sure if it could be generalised like that.

I remember when I was shopping for a socket 5 motherboard back in 1995 the Soyo was definitely a premium motherboard brand compared to Shuttle I ended up getting.

It is true that PCChips and like are usually budget offerings (some exceptions). For other brands it usually depends on the actual model

Blog|NexGen 586|S4

Reply 2 of 16, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

In middle Gigabyte, Soyo, Soltek , Biostar, MSI , QDI, etc.

Biostar in middle... ha ha.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 4 of 16, by W.x.

User metadata
Rank Member
Rank
Member
The Serpent Rider wrote on 2021-01-08, 10:59:

In middle Gigabyte, Soyo, Soltek , Biostar, MSI , QDI, etc.

Biostar in middle... ha ha.

Hmm, I have feeling, you didn't get it.
They are between PCChips, Matsonic and TOP ones. I didnt meant it to differate those in between them, of course, those middle ones have also different levels.
But definetly, I don't feel, Biostar is on PCChips level.

Reply 5 of 16, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Regardless of the brand, you would want to avoid any motherboard with RTC battery, if you lack soldering skills.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 6 of 16, by W.x.

User metadata
Rank Member
Rank
Member
The Serpent Rider wrote on 2021-01-08, 11:08:

Regardless of the brand, you would want to avoid any motherboard with RTC battery, if you lack soldering skills.

Thank for advice. I almost missed it, it has not coin battery. You're right, I won't take it then.

Reply 7 of 16, by rmay635703

User metadata
Rank Oldbie
Rank
Oldbie

OEM Proprietary DTK were rock solid through 1991, (albeit sometimes rather slow lacking cache)
Never saw a 486 DTK let alone socket 7 so can’t comment, I do know that I had issues moving a 386sx25 DTK Board into another case as nothing lined up but again can’t comment on a socket 7

Reply 8 of 16, by dionb

User metadata
Rank l33t++
Rank
l33t++

DTK was one of the bigger brands in the really early days (1980s), but quite rare by Pentium era. The old stuff was pretty robust, but can't say I've ever handled 1996-era things of theirs. Brand alone can be a very bad advisor - brand quality varies over time, and even the best brands sometimes make steaming piles of crap. Also consider that "quality" has many aspects. Are we talking build quality? Reliability? Compatibility?

I'd suggest looking up info on the specific board, and failing that (wouldn't expect to find much in this case), just 'reading' the layout. To me this one looks like a pretty generic i430VX babyAT board. 256kB of PLB cache with COAST slot for more. What looks like PS/2 header for a mouse. Potatocam pics make it tricky to be sure, but at least the multiplier & FSB settings are documented on the silkscreen.
On the downside, an old-style inefficient/hot linear voltage regulator for 2.8V, and that Dallas RTC.

As for the latter- I'd not consider it a show-stopper as you can buy them new (Dallas is called Maxim these days and still makes the DS12887) or you can get homebrew replacements (basically a little PCB with SMD DS12885 and a coin battery) or you can mod the current one with a battery. What does matter (at least in terms of work involved) is whether the RTC is socketed or soldered. Taking a good look at the last pic, I clearly see a horizontal line on the side of the RTC - it's socketed. So yes, I'd prefer a CR-2032-based board, but failing that, this one is perfectly serviceable and the RTC can be replaced without tools or effort. I pay around EUR 11 for new DS12887 modules. Look up what they cost locally from a reputable electronics supplier (not eBay, you then get old stuff with batteries already dead) and factor that in to the cost of the board.

Reply 9 of 16, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Looks soldered to me. Then again, for 50 bucks, it's probably better to look for some ATX board.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 10 of 16, by dionb

User metadata
Rank l33t++
Rank
l33t++
The Serpent Rider wrote on 2021-01-08, 12:58:

Looks soldered to me. Then again, for 50 bucks, it's probably better to look for some ATX board.

Applying best CSI-skills:

rtc.png
Filename
rtc.png
File size
55.63 KiB
Views
511 views
File license
CC-BY-4.0

Reply 11 of 16, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Yeah, I did that too. Still not convinced. Here's a better picture of the board: https://www.ebay.de/itm/174563941088

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 13 of 16, by Deksor

User metadata
Rank l33t
Rank
l33t

While making UH19 we noticed that some brands are actually the same. DTK is also Gemlight for example.

And the "bottom" range of boards can be resumed to one brand : PCChips.

Now ECS is a different story. As @dionb mentioned, quality changes over time. ECS made pretty good 486 boards in my opinion. Plus most of them didn't have any barrel batteries. Then in the late 90's they got closer and closer to PCChips until they bought them. So they pretty much became PCChips.

As for PCChips, in my opinion build quality have always been between "meh" to terrible and they've always done shady things like hiding the chipset's brand, soldering the bios, having 10 brands for one board, etc.

However despite all this many of their board still work to this day (might be the survival bias tho ?). Maybe that's also related to the fact they dropped the varta batteries quite early which makes them much more convenient these days. Moreover they often put features that other brands didn't put in their boards and they also managed to make their boards quite fast.

It's almost too bad that they were trying to scam people. If they didn't do that but kept their qualities they could have been a pretty good and respected business nowadays.

This is why some people consider the PCChips m919 as one of the "Ultimate 486 board" I think.

Trying to identify old hardware ? Visit The retro web - Project's thread The Retro Web project - a stason.org/TH99 alternative

Reply 14 of 16, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Yeah, it boils down to specific boards. ASUS also had quite meh boards in that time period. I was surprised to discover that Zida Tomato made decent PPro motherboard.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 15 of 16, by W.x.

User metadata
Rank Member
Rank
Member
Deksor wrote on 2021-01-08, 15:45:

And the "bottom" range of boards can be resumed to one brand : PCChips.

Now ECS is a different story. As @dionb mentioned, quality changes over time. ECS made pretty good 486 boards in my opinion. Plus most of them didn't have any barrel batteries. Then in the late 90's they got closer and closer to PCChips until they bought them. So they pretty much became PCChips.

As for PCChips, in my opinion build quality have always been between "meh" to terrible and they've always done shady things like hiding the chipset's brand, soldering the bios, having 10 brands for one board, etc.

Well, in this Toms reviews of VIA socket 370 boards, from mid-2000, ECS ended "quite well". (like, not bad, I mean. For the price)
https://cdn.mos.cms.futurecdn.net/YisY5MyXstpzCR6RPxaeuH.gif

Who would expect for example, Abit would end up about same as QDI? Of course, it's good always to test direct board (and even BIOS version, and revision)

It's Comprehensive VIA motherboard roundup June 2000, ECS description is here in slightly older review , although he didnt recommended the board. Of course... but I think, price was quite good. Still , not so bad, when you expect them on bottom, like PC Chips. Definetely didnt ended on last place thought.
https://www.tomshardware.com/reviews/12-socke … ards,196-7.html

So I would say, overall, they've made slightly better boards, than PC Chips. And from time to time, they suprises (possitively - for the price, that was low)

The purpouse of the hierarchy was, when you don't know about board, and you need to take it fast, you are risking, but overall, when you take Abit or Asus, you have less chance, when you take random board, that it will be bad. When you take ECS, you have much higher chance, that it will be eighter bad, or below average. Also with worse capacitors, thinner PCB, and durability. It was not like, no board from ECS can be good, and all boards from Asus will be always on top. But , when you take statistic when you blindly take board, highest chance will be on Asus and Abit, it will be good take. On PC Chips, you almost always will have terrible, or below -average board. That was purpuse of hierarchy. I wanted to know about DTK, where it "overall" is.

Reply 16 of 16, by dionb

User metadata
Rank l33t++
Rank
l33t++
W.x. wrote on 2021-01-16, 12:16:
[...] […]
Show full quote

[...]

Well, in this Toms reviews of VIA socket 370 boards, from mid-2000, ECS ended "quite well". (like, not bad, I mean. For the price)
https://cdn.mos.cms.futurecdn.net/YisY5MyXstpzCR6RPxaeuH.gif

Who would expect for example, Abit would end up about same as QDI? Of course, it's good always to test direct board (and even BIOS version, and revision)

It's Comprehensive VIA motherboard roundup June 2000, ECS description is here in slightly older review , although he didnt recommended the board. Of course... but I think, price was quite good. Still , not so bad, when you expect them on bottom, like PC Chips. Definetely didnt ended on last place thought.
https://www.tomshardware.com/reviews/12-socke … ards,196-7.html

So I would say, overall, they've made slightly better boards, than PC Chips. And from time to time, they suprises (possitively - for the price, that was low)

The purpouse of the hierarchy was, when you don't know about board, and you need to take it fast, you are risking, but overall, when you take Abit or Asus, you have less chance, when you take random board, that it will be bad. When you take ECS, you have much higher chance, that it will be eighter bad, or below average. Also with worse capacitors, thinner PCB, and durability. It was not like, no board from ECS can be good, and all boards from Asus will be always on top. But , when you take statistic when you blindly take board, highest chance will be on Asus and Abit, it will be good take. On PC Chips, you almost always will have terrible, or below -average board. That was purpuse of hierarchy. I wanted to know about DTK, where it "overall" is.

Tbh, Asus boards certainly won't always be on top. Yes, if you completely randomly buy a board based on brand only, you have better chances with Asus than with ECS, but that won't help you one little bit with say the P3C-D (I literally needed to acquire 10 of those boards from 2002-2020 until I found one that actually worked last spring), P4P800 (great boards, but the VRMs had the alarming tendency to utterly fail without warning) or the A7N8X-range (fast and stable - but memory compatibility was horrendous, DIMMs that had no problems on Gigabyte or Abit boards with same chipset refused to boot or gave massive errors).

'Hierarchy' stuff is just a crutch for insufficient information. A far better idea is to solve the information problem and to research the specific board. If it's great, it doesn't matter if brand reputation in general is awful, if it's shite, no amount of good stories about other boards from the brand will help.