Guidelines for PC Upgrades

[Mercutio]

Here's a fun idea. Just got off the phone with a guy who wants to upgrade his PC. I told him there's probably no point, he should just buy a new one.

Nope, he wants an upgrade. He has a 1.4GHz P4. A six-year-old PC. It would need a new PSU (200W), a new motherboard, new RAM (his uses RIMMs). Oh, and he'd like one of those DVD burners, too. Oh, and what is Windows XP like?

Which means his "upgrade" would pretty much be everything but his hard disk, floppy and case.

With that in mind, this is more or less what I told the guy.

1. Do not upgrade PCs that shipped with an older version of Windows to any newer version of Windows
Exception: Windows 2000 from Windows ME is probably a good idea. XP probably isn't.
2. An upgrade of CPU in the same socket type is probably not worthwhile; too little performance gain for the money.
3. RAM is a safe upgrade. However, if RAM for your PC costs twice as much as the same amount of RAM for a new PC, don't bother.
4. A motherboard upgrade nearly always means a CPU and RAM change. Which might as well be a whole PC.
5. Video hardware upgrades are only worthwhile if you're a gamer or an engineer.

Can anyone think of any others?

 

[Buck]

Depending on the type of tasks an OEM system is used for, many times an upgrade isn't necessary at all. I'm finding many older VAIO systems with P4 1.8 GHz processors operating at a ruinously slow speed because they come glutted with obese software. Just sell a new copy of Windows XP, install a good browser, mail client, anti-x ware and the necessary business applications. Suddenly, these systems pick up in speed and can easily handle another three years of service.

 

[ddrueding]

I agree with Buck, but when a faster system is needed or wanted:

1. If MB (therefore CPU and RAM) and PSU need to be swapped, an upgrade is silly.
2. If OS is to be updated, an upgrade is probably silly.
3. Don't keep the hard drive from an old system, it will cripple your upgrade.
4. Don't upgrade and don't plan for upgrades...it's easier that way.

 

[Bozo]

I agree with everyone here, but:

You could install a Celeron D 340 (2.93GHz) for $57.00.
Do clean install of Win2K.
Add the DVD burner. ~$30.00

Bozo :joker:

 

[Stereodude]

I agree with everyone here, but:

You could install a Celeron D 340 (2.93GHz) for $57.00.
Do clean install of Win2K.
Add the DVD burner. ~$30.00

Bozo :joker:

It probably doesn't support 533MHz FSB, so that might not be an option.

 

[Tannin]

It always depends: on the machine and on the customer. Often, Merc's approach is the right one. Equally often, it isn't. For example: you have a customer with a Duron 800, 40GB, 256MB of SDRAM, some crappy 8MB video card. Six months ago they put in a new optical drive. Last month, the kids put XP onto the system and now they wonderwhy it's such a slug.

New box? Cost them around $AU800

Replace the board, CPU, video card and RAM with a Gigabyte VIA all-in-one board, 512MB of DDR and a Sempron 3000. Back up the data and reinstall from scratch. For what they need (surfing, email, some office stuff, digital camera) it's plenty. Costs them around half as much.

It's a pretty common scenario, as are variations on that general theme.

The bottom line there is no blanket One Right Answer, you have to think it through every time, play every ball on its merits.

 

[Mercutio]

Intel dropped socket 423 in 2002, if I recall correctly.

 

[Tannin]

Huh?

 

[Mercutio]

That was a reply to the suggestion that the gentleman seeking an upgrade could buy a high-end Celeron.
Which he couldn't, 'cause his motherboard uses socket 423, which Intel hasn't used in ages.
And you posted that while I was replying.

I *do* believe that there are "just scrap the whole damned thing" limits. I'm thinking there are sensible rules of thumb that can be applied generally, rather than on Tannin's case-by-case argument.

 

[Tannin]

Yeah, well, what I do is apply sensible general rules of thumb on a case by case basis. Errr .... or something like that.

Also, there is the psychological aspect to consider. I regularly meet people who won't actually come out and say so, and may not even realise it themselves, but who have already decided that they are going to buy a new unit or that they are not going to buy a new unit.

So we go through the routine where I recommend this and suggest that, and that's all fine, but they wind up saying:

"OK, but couldn't you just put in a new motherboard, CPU, RAM, video card, power supply, DVD burner, a Windows XP, and a flash card reader?"

"Yep", I say, "but that will cost $950, where a whoile new system will only cost $990".

"No", they reply, "we can't afford to buy a whole new system when there is nothing wrong with the old one apart from the motherboard, CPU, RAM, video card, power supply, lack of a DVD burner, the old Windows 98, and not having a flash card reader. Do the upgrade please."

OK, I've exaggarated a bit, but less than you might think (you have met these guys too, I think Merc). And then there is the other sort. Perfectly good system, just needs a small repair, but no matter what the logic is, they want a shiny new one. I had one of those today. The old 17 inch CRT monitor blew up, so when it's all said and done, it will "be more sensible" to buy an entire new system - keyboard, system unit, the lot. There is no real need for that at all, in this particular case, just like others similar, but once the mind is made up, it's made up, so I'm building a new Athlon 64 tonight for them.

These days, I don't fuss too much about it: I make my recommendation and then I do whatever it is they have already decided to do anyway. If they want to follow my advice, that's great - and 9 times out of 10 it's the right advice - but if they want to do an upgrade when it makes more sense to buy new or they want to buy new when it makes more sense to upgrade, then I've given them the best advice I have and if they don't want to follow it, then so be it.

The one thing I won't do is spend hours and hours trying to breathe life into a system that I believe to be so old and crappy that it will break again before the ink of the credit card statement is dry. You wind up spending ridiculous amounts of time on it, not charging anywhere near enough to cover yourself, and they usually bring it back later and complain anyway.