Do you have a model number of the power supply you're looking to get?

300W is a good amount but you'd have to make sure it is an AT power supply most likely (and not ATX) as your motherboard probably has this connector:

It's an Athena Power AP-AT30; an industrial PSU made specifically for these older systems. It has the old P8 and P9 connectors (as they seem to be called) so it should be all good in terms of connectors.

That's a good unit and will function well. You can use a power supply no matter how much wattage it is capable of delivering. The only issue is a lack of efficiency if you go either too high or too low.

300W is perfect for a 486 build as well as early Pentium. I've seen 486s with 150W and running fine.

Alright, good to know. I'll get the supply as soon as I can. Thank you!

Welcome!

Don't worry about it. I have 850W Seasonic prime TX PSU on a socket 7 PC that uses max 60W measured at wall plug when under heavy load. I had a psu failure on it recently and that was the only ATX psu I had that was not in use.

Baoran wrote on 2023-07-24, 01:04:

Don't worry about it. I have 850W Seasonic prime TX PSU on a socket 7 PC that uses max 60W measured at wall plug when under heavy load. I had a psu failure on it recently and that was the only ATX psu I had that was not in use.

I have a Seasonic X-Series 1050W 80+ Gold PSU powering the system in my signature 🤣 I hear ya about "the only PSU I had available".

Watts don’t work that way.

The watts of a psu is how much you can draw from it. A more specific thing you need to pay attention to is how many amps on each voltage your mobo and drives will use. You need to make sure. You aren’t using more than the psu can provide.

Again with amps on the psu, it is an avability thing, that is how many the psu can provide, how many do you need? Don’t exceed it. Or the psu may fail.

So long story short, there isn’t such a thing as too many watts or amps on a psu. Only too little. It is the loads that determine what is actually used.

Majority of PSUs get very inefficient when loaded far below their spec so on that regard there is a "too many watts". A 486 will be fine with the lowest power PSU you can find, entire system will take only few tens of W at most and especially when you don't use HDD(s) anymore.

Tiido wrote on 2023-07-24, 05:28:

Majority of PSUs get very inefficient when loaded far below their spec so on that regard there is a "too many watts". A 486 will be fine with the lowest power PSU you can find, entire system will take only few tens of W at most and especially when you don't use HDD(s) anymore.

Got a kill-a-watt?

Could someone post the wattage usage between a 150w , 500w and 1000w on the same system? (Being sure that all the psus are switching, not linear)

Curious to see how much of a real world wattage difference this makes. It does make sense that a lower wattage psu would have lower idle watts leading to more efficiency with smaller loads

It’s a curious thing you suggest about the efficiency. I guess this comes from basically static idle wattage and saturating the larger inductors, driving the larger silicon, etc. A psu that is grossly oversized will still work, and work fine at that, but the efficiency angle is curious.

Could go one step farther and analyze noise levels at different loading also with a scope.

In case of a 486 I have measured the current on individual power rails and it is just few A for bare motherboard (largely depending on what CPU and RAM you use) and some more as you add more addon cards. HDD adds about 1A to the 12V rail too in normal operation and a little on 5V too. In the end it amounts to just tens of W that the PSU has to provide, which is why most of these old machines I have seen have 150W range PSUs, there just isn't a need for more .

Measurement on the wall end is always much higher, since it takes into account the PSU's own inefficiencies too and they can be quite high on a lightly loaded PSU, i.e efficiency is near 70% on low load vs 90%+ on a full load. Modern stuff does have lower losses though, but lightly loaded PSU is still pretty inefficient is most cases.

It is really all relative. Looking at efficiency curve of my Seasonic TX-850, it still shows bit over 90% efficiency at 50W compared to 96% efficiency at 50% load which means if I am getting 60W at wall plug means computer itself is using 54.55W and if I was at optimal load getting 96% efficiency with same 54.55W it should show 56.73W at wall plug which means I am losing 3.27W by not having optimal load.

And you have to consider that you really cant get low wattage atx power supplies with high efficiency. Lowest wattage psu you could buy where I live is 450W 80 plus gold one and it gets 90.6% efficiency at 50% load and it would be 84% efficiency at 54.55W load.

The differences of efficiency at low wattage are not big enough for me to worry about that. I am more worried about that I am running the psu in a way that it is probably not meant for by having most load of 5V or 3.3V rails and very little load on big 12V rails it has.

Personally, I'm not such a big fan of efficency. More than often, I think, it's just meaningless statistics without taking other relevant factors into account (like sustainability, reliability, lifetime et cetera).

It all depends on the point of view, I think.
For example, how do certain modern ATX power supplies with 80+ compare to ATX power supplies from the 90s?
Or XT and AT power supplies from the 80s and 90s?
Are they worse, better or about same?
Edit: In a 486 setup, in this case.

All in all, it's like those awkward science reports that try to connect IQ levels and BMI, I suppose.
Or like Astrology. Or like trying to associate character traits and blood groups (popular in far east).
In the end, the statistical numbers will make you see what you want to see.

Edit: To name an example: LED lamps, the number #1 liars of statistics.
On paper, LED bulbs are suuuuper efficient. And long-lived. So awesome.
In reality, they're not great, last a few months. Because their internal power supplies break (single diode as rectifier without a filter/buffer capacitor).

Whereas vintage incandescent lamps can run for years.
Their lifetime increases dramatically if the operation voltage is being reduced by 5 to 10%.
If two are wired in series, they may outlive their owners, even.
Light quality is also better (spectrum being more homogeneous). They don't hurt the eyes (blue spectrum is strong on LED lamps).
They're environmental friendly, too. They consist of metal and glass only, no toxic substances. No plastics.

Alas, statistic is being bent so far that it makes LED lamps look better in every discipline.
It's so good at creating illusions that David Copperfield can't compare with it, even. 😂

Because, the many cons are not being remembered when talking about LED lamps.
Merely their higher power efficiency (ugh!) counts if people call them eco-friendly.
Their short real-world live-spans and afterlife on the dumpsters don't count. 😔

Not sure if it is true in reality but I always thought that they have to use better quality components inside a psu to reach that higher efficiency and higher quality components would also mean that they last longer. I was never worried about efficiency itself either but I wanted a good quality psu.