H3nrik V! wrote on 2024-11-04, 18:54:

I'd say that anything below 7 volts could pose an issue with fans not starting ...

Depends a lot on each particular fan... though I've never ran into any 12V fans so far that don't start at 6-7V yet. I'd say 5V is more along the lines of where the less powerful ones might get a bit "too lazy" and not want to start. But for most part, I've seen very few not want to start on 5V either. Most are OK. Below 5V is definitely an issue for many, though.

And just for a fun reference...
Recently I was working on an old Bestec PSU with an 80 mm fan that I thought sounded too loud. I found there's an external fan speed adjustment signal, so I started playing with that and slowed down just a notch. Judging by the noise the fan made, I though it was running somewhere in the range of 10-11V by default and that my adjustment slowed it down to somewhere around 7-8V. When I opened the PSU to check the voltage directly at the fan, I found I was completely wrong on this. By default, the PSU was giving about 5.2V to the fan as a starting voltage. And with my mod, I was feeding it around 4.5V... which still sounded a bit on the loud side and was turning quite quickly. I found that only when I adjusted the voltage to dip below approximately 3.8V that the fan would refuse to turn. And yes, I checked, it's a 12V fan... an Adda ball bearing one, to be precise.

Anyways, long story short, YMMV with various fans about the lowest starting voltage. But in my experience, 6-8V is usually where they have the best balance between noise and air movement, and at 5V is where they are near silent but don't move much air. Below 5V, some may not start or just turn so slowly that they aren't really doing that much.

Maybe I've just been messing around with too many too worn fans, then 🤣 but at 7 volts I think that each and everyone I've tried, worked ...

You can get PWM motor controllers from china for like 1$ a piece. Why not using that?
This does not have an issue with fan startup, because the supply is always 12V. (Or 0V during the down-part of PWM)
No need to play around with voltage.

H3nrik V! wrote on 2024-11-04, 22:15:

Maybe I've just been messing around with too many too worn fans, then 🤣

Same here.
I cleaned up / refurbished around 15 fans last week alone - some from very old hardware, some newer. Most were OK and just needed a cleaning. A few are awaiting bearing servicing.
I regularly test my fans at various voltages too, just to see where they perform most optimally (in terms of airflow and noise.)

GigAHerZ wrote on 2024-11-05, 09:13:

You can get PWM motor controllers from china for like 1$ a piece. Why not using that?
This does not have an issue with fan startup, because the supply is always 12V. (Or 0V during the down-part of PWM)
No need to play around with voltage.

Not exactly how it works.
Yes, the output voltage from the PWM controller does vary only between 0V and 12V, but the *average* voltage the fan stator windings "see" will still be an analog value between 0V and 12V, mostly depending on the duty cycle of the PWM controller at that instant. So at best, it's no different than using a linear voltage control method. At worst, some fans may "sing" or "buzz" weirdly with certain PWM controllers, especially at lower speeds / duty cycles. Whether that happens or not will depend on the frequency of the PWM controller and the inductance of the fan's stator windings. Most fans will work OK, but some may not exhibit a nice linear speed control.
With that said, yes, even a PWM controller will have issues with fan startup at very low duty cycles (i.e. very low average stator winding voltage/current.)

From my experience, PWM lets the fan start reliably at much lower rpm compared to linear voltage to the fan. But yes, at some point the duty cycle will be too narrow to produce any meaningful current from inductances involved in fan windings and then you get issues with startup. These kind of problems are not difficult to solve though, it isn't hard to create short higher duty cycle or higher voltage startup that smoothly gets to lower target duty cycle or voltage. Once the fan spins it can go to quite low rpm before it'll stop and cannot start again.

@momaka, PWM supply helps the fan to properly start at lower speeds and it also stalls on lower speeds compared to just linear regulation.

Found a Sony Trinitron TV KV- 21FT1K although the cord was cut.

Likely working though, since the yoke scalpers didn't manage to get their hands on it, and I did.

I found an ASUS M2A-VM w/ Athon X2 3800 in a cool low profile minuet case, 1GB RAM 100GB HD.

And I also found MITX Intel BLKDH61DLB3 w/ an I3-2120t & 8GB ram in this cool silverstone case:

Unfortunately the powersupply seems to be gunked up with dust & second hand smoke

If I wanted to replace the power supply, is this a standard size I could replace this with?

I think there is a specialty form factor for that kind of PSU, but I can't remember what it's called.

Kahenraz wrote on 2024-11-15, 19:29:

I think there is a specialty form factor for that kind of PSU, but I can't remember what it's called.

For those small systems it's the SFF style PSU. There's so SFF-L for higher wattage units.

Just be careful not to get your search terms mixed up with the new SFF standard ITX PSU form factor. I'm pretty sure that's something new and different. Someone more knowledgeable can correct me if I'm wrong.

douglar wrote on 2024-11-15, 18:59:

If I wanted to replace the power supply, is this a standard size I could replace this with?

The case looks like an SG05, which accepts SFX PSUs, basically slightly smaller all round compared to ATX.

jtchip wrote on 2024-11-16, 01:41:

The case looks like an SG05, which accepts SFX PSUs, basically slightly smaller all round compared to ATX.

Thanks JT. That smells right

(Unlike the current power supply that smells like an ash tray and doesn't do nothing)

chrismeyer6 wrote on 2024-11-15, 19:33:

For those small systems it's the SFF style PSU. There's so SFF-L for higher wattage units.

Some number of years ago, a friend gave me this little Dell in which
the SFF PSU had died (enough that we didn't try to repair).

I liked it's small size enough that I decided to make it a little "server"
in my basement... Not having another SFF PSU, and having several spare ATX
supplies, I cut the power-connector with some cable from the dead supply
and used an ATX power connector recovered from a dead mainboard to make
a simple adapter to connect a standard ATX to the Dell.

As you can see the ATX didn't fit in the case, so I mounted it to the side.

This also means there is no PSU "exhaust fan" in the case now, but there
is a decent sized fan behind the front, blowing over the CPU heatsink and
exiting through the back ... internal temp is not bad at all.

This thing runs ALL the time... I built an "automatic UPS for it" which
shuts it down gracefully if power goes away for more than 3 mins, and
waits till power is restored&stable for 30mins before starting it up again.

So it runs even when I'm travelling. I use remote access software to check
my cameras, water leak sensors etc... and since it holds the repository of
"where I backup everything as I work on it" - I can get stuff from
it if I need.

It's been running for more than 5 years without a hitch.
(It outlasted the lawn-tractor battery I use for the UPS).

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

DaveDDS wrote on 2024-11-16, 03:41:

This thing runs ALL the time... I built an "automatic UPS for it" which shuts it down gracefully if power goes away for more tha […]
Show full quote

This thing runs ALL the time... I built an "automatic UPS for it" which
shuts it down gracefully if power goes away for more than 3 mins, and
waits till power is restored&stable for 30mins before starting it up again.

So it runs even when I'm travelling. I use remote access software to check
my cameras, water leak sensors etc... and since it holds the repository of
"where I backup everything as I work on it" - I can get stuff from
it if I need.

I know that a system can be configured to restore power to the "last state" (on or off) from the BIOS. But how do you get it to turn on after 30 minutes of stable power? Is this a Dell thing or a feature of the PSU?

Kahenraz wrote on 2024-11-16, 16:08:

I know that a system can be configured to restore power to the "last state" (on or off) from the BIOS. But how do you get it to turn on after 30 minutes of stable power? Is this a Dell thing or a feature of the PSU?

Neither! - It's a "feature" of the UPS I built.
It's fairly "smart" (has an STM32 microcontroller inside).

What you can't see in the picture, is that there's a new connector on the back
which connects to the front panel power button, and also the system +5 rail.

When the UPS controller detect that the mains power had failed, it immediatly
begins generating it's own 110v AC for the system, it also starts an internal
timer. If mains power returns in under 3 mins, it switches back to mains output
and everything proceeds with no other action.

If mains power remains OFF for 3 mins (or more). The UPS:
- Hits the system power button
- Waits for system to shut itself off (+5v rail drops)
- Shuts itself off.

When mains power returns the UPS does not give power to the system "right away",
it waits till power has been continuous for 30 mins. If power fails during this
time the timer "resets" when it comes back.

Once it decides power is good, the UPS applies power to the system, waits a few
seconds for everything to stabalize, then hits the system power button and waits
till the +5v rail comes up - then returns to the intial state of monitoring for
a power failure.

There are of course the usual checks and retries in these steps.

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

Have not got it yet but my nephew says he has an old tower stored in his shed he will bring me. Am hoping it is one I gave him years ago with a descent motherboard but probably not...
If something nice will post about it, if something crappy will post about it. Hey Nice and Crappy 🤣

Found a few things, the most notable being a (somewhat) working Geforce 4 Ti 4600 from VisionTek. Has some issues with Duke3D and text modes seemed really sluggish. Might need to use it in a faster system. 😀

TheMLGladiator wrote on 2024-12-15, 03:03:

Found a few things, the most notable being a (somewhat) working Geforce 4 Ti 4600 from VisionTek. Has some issues with Duke3D and text modes seemed really sluggish. Might need to use it in a faster system. 😀

Wow, nice! That's quite a thing to find these days.

If you haven't already, it's probably worth popping the heatsink off and repasting it. I wouldn't bother on a lot of cards, but the cooling on these reference design Ti 4400\4600 cards is only barely adequate, and I have seen them suffer from heat death. I baked one identical to yours in an oven once (before I knew better... it was like 12 years ago...) and it brought it back to life, but that tells me there can be some BGA issues on these. Keep'em cool! 😀

Throwing something in the metal recycling bin at work, I saw, what I recognized as a gutted-out ATX case from A PC, I knew had been smashed under an electric adjustable table ... I went directly to the electronics scrap dumpster, and there was the Intel D915PCY with a 3 GHz 1 MiB Preshott. Not a great find, but DEFINETELY not something to throw away.

IT guys have been lectured 🤣