Abit BM6-ZM6: it promotes for Celeron PPGA
Wise moves (and errors) of the design
Very well, let's practise the favorite sport of the country: to criticize the foreign work!
Starting on the physical part, the election of a distribution of slots 1/5/2 (1 AGP, 5 PCI and 2 ISA) presents many advantages in a modern badge, since nowadays even the sound cards and some internal modems come in format PCI, stopping to the grooves ISA only like compatibility elements.
As we comment already, there must bear in mind that to support a size of badge reasonable and compatible with the biggest possible number of casings, the fifth slot PCI shares his space with firstly ISA (slightly perfectly normal), so in practice the distribution stays like 1/5/1 ó 1/4/2.
The distribution of the components is good and it expires with the standard ATX, without hindering in the line of the cards of expansion (I had one of more than 30 cm long) and leaving the connectors for discs and disk drive in the right wing of the badge. Personally I would have placed the feeding connector in the right top end, but neither it bothers where it is...
We go, in fact it is not that it is an original design: it is the same design of the Abit BH6 but with a Socket370 instead of the Slot1. And the fact is that the BH6 is considered to be one of the best semiprofessional badges for Pentium II (especially between the fans of the overclocking), and if something works: why to change it?
One is grateful for the detail (quite common but by no means standard) of a heat spendthrift places on the chipset (in fact on North Bridge or AGPset, marked 443BX or 443ZX, one of two chips that form the chipset and whose name is taken to designate globally the chipset). He would surprise what the computers warm up nowadays, sometimes there look like more toasters than another thing...
... So something was necessary to do on this matter, that's why both badges they include 2 termistores to measure the temperature of the mike and of the interior of the casing, in addition to a connector to install a third termistor and 2 connectors for fan with aptitude to measure the number of revolutions in minute of the same ones (although unfortunately the normal fans do not provide the above mentioned information).
Otherwise, in the physical paragraph there is not any more that to say; enough anchoring points exist, the external connectors are the habitual ones and distributed of a very typical form, the general impression is of soundness and it has enough capacitadores as to assure a stable yield.
In the logical, paragraph not negative at all to be commented. The BIOS (Award) has the presentation and handling classic, with a common extra paragraph to all the badges modern Abit: the SoftMenu, to form the CPU for software. This program, in his version II, allows to select in seconds any multiplier's combination and bus speed without having to fight either with minuscule jumpers or switches.
This not only facilitates the configuration of the computer, but it turns out to be an indispensable help to do overclocking, and even allows to select the voltage of the mike or reduce the speed of the buses AGP and/or PCI to avoid problems with cards in these formats on having done overclocking. Sincerely, fondness is taken to the programita.
Otherwise, the BIOS has the whole fan of modern options as for energy or resource allocation and only one throws in absence an option to select the language in which to form it. In fact I already would not get accustomed to doing it in other that was not English, but for the newly arrived persons it would be an estimable help.
And otherwise, the cables are included for disk drive and hard disk (the only one of every, of course), a manual of more than 100 pages absolutely exhaustively, with an enviable section explaining the options of the BIOS (but it is only in English, so I do not cover danger) and finally a CD - ROM with electronic versions of the manual (in this case also in German and Chinese), drivers for Windows, a version off-line of the web site of Abit (slightly very curious) and a program to verify the temperature of the CPU, the speed of the fans and other parameters from Windows; this information also can be consulted in the BIOS.
Proving the badges and overclockeando
... And in agreement, overclockeando is not precisely a word worthy of Cervantes, but I prefer it to "raising the speed over the nominal values". We are going to see how these two badges behave in practice in real conditions... or even much worse than the normal thing.
First of all, the test configuration was the following one:
In fact, to obtain these so recent basic badges has brought with it a problem; I have thrown myself almost every February looking for shops where to obtain Celerons PPGA, without managing but answers like: "and that what is?", "oh, we still do not have it, but we are going to bring it" (3 weeks later they were saying the same), or "I feel it, but only we would sell it to you with a motherboard" (or with an entire team!). Very well, in the end it was, but it has cost; it is already known that to buy something when it is an entire innovation is not a good idea...
To prove the quality of a motherboard has never been an easy thing, since his position as heart of the system makes impossible to isolate it of the rest of the components. This way, not only we will be proving the motherboard, but a certain mike and, depending on the test, a concrete graphic card and even the hard disk, therefore to speak about yield of a badge is something quite relative.
Also, nowadays almost all the badges with the same chipset and mike give a few very similar results, therefore it turns out to be more interesting to observe other factors as the stability of the badge. If the badge works correctly for the long time in the worst conditions of possible work, and also with a reasonable speed, the badge is undoubtedly of quality.
Bearing in mind this, there has been selected the test Winstone 99 Business (v. 1.0.) of Ziff-Davis, an absolutely stressful test for the system: 250 MB of test, consistent in executing nothing more and neither more nor less than a series of macros or scripts in WordPerfect, WordPro, Word, Quattro Pro, Lotus 123, Excel and the rest of applications of 3 suites ofimáticas more important than the market, together with Netscape Navigator... during almost 20 minutes for test.
The tests recured throughout several hours, to make sure that the stability of the system was supported, in a test that turned out to be more terrible for the hard disk than for the motherboard (I went so far as to be afraid for his physical integrity, the truth).
Next and without allowing to rest to the system an informal test was realized with a few games (I love this part of the test!!); informal to the effect that there were taken measures of the number of pictures neither by second nor anything for the style, since the numbers that usually appear in Internet on these topics are always based on cards of high scale like Voodoo2 or enclosed Riva TNT, of an exceptional yield very superior to that of the economic one i740, that on the other hand is one of the cards most used at present. Anyway, I make sure you that I enjoyed to the maximum of Quake2 (to 800x600 points, of course).
Seguidamente began again the system and overclockeó (overcloqueó?) the mike up to 450 MHz, raise the speed of the system to 75 MHz, repeat the previous tests with the new speed. Also there were done other less exhaustive tests (Speedy to prove the graphic part and CPUmark 99 to prove the mike). The following results were obtained:
Difference between both badges is not done, because the results for the BM6 and the ZM6 were practically identical. The differences have never gone over 2 %, being the most probable thing that they should to an incorrect defragmentación of the disc or to minimal momentary changes in the yield. In fact, it not even seemed that the BM6 was superior to the ZM6...
As we were saying, the numbers in themselves are not the most significant thing, although it is necessary to say that they are very similar or even superior to those that Intel declares in his web for the mike to 400 MHz (30.8 in the CPUmark 99 and 18 in the Winstone 99), being frankly good thinking that they have been obtained by a graphic card of less than 10.000 pts VAT included. The most important thing is to observe that the system was hung not even only one time, not even to 450 MHz, despite 3 times repeat the tests with every speed.