@wave
I know an oscilloscope would be appropriate, but I have to set that up yet.
Is there a way to measure that AC component? Capacitor to remove DC and measure what arrives on the other side of the cap?
Sorry for stupid questions 😀
As far as I know, ripple can be measured only with a scope.
Enermax was a reputable manufacturer back then, I wouldn't connect the fan they way you did, and also try to keep the fan controller enabled because the fan's life will shorten dramatically if constantly run at full speed.
wrote:I wouldn't connect the fan they way you did
Which way would you connect it, if the fan life does not matter?
(I just change them when they do not blow well anymore. It is so hard to find out whether a fan is weak or just controlled down, much easier when you know it has full 12V.
As I have plenty of spares from servers, I do not care if they last only a few years. I do not care about noise, to me it matters way more to keep the electronics cool.)
The identical goal would be to keep electronic parts temperature as stable as possible. I would keep the fan controller and do some load/temperature tests as I have done in the past with ATX psus. If I find the fan controller to be too conservative, then I would put a resistor in parallel with thermistor (most of the times the thermistor used is NTC thermistor) and do more tests, then put a different value resistor, until I find the best value for my taste. I am not a fan at all of super silent computing but wouldn't want either a fan running too fast unless needed and bury the psu components in dust.
Btw Those AT (and early ATX) psus don't behave very well with no load. Many times they don't start at all or work but produce funny noises.
wrote:...then I would put a resistor in parallel with thermistor...
Thank you! This is a great idea!
I guess I'll do that when I got the Schaffer can, so I can do both things in one go.
(Maybe I'll just drill a hole into the psu case to mount a potentiometer, to make it easily controllable without having to open the psu)
(But I then also need to find out for how much current the fan controller is specified... 0.6A is more than common with PC fans)
Hi Retardware!
This Enermax unit is a group regulated desing, so unless the 5V rail is loaded with at least 2-3 Amps the other voltages will be acting funny. If you plan to adjust the fan speed you can connect a 4.7kohm resistor in parallel with that ttc103 thermistor. This will boost up that minimum fan speed to medium fast. You can also use a 10kohm potentiometer in parallel to adjust fan speed manualy. Be carefull about the current consumption of the fan. 0.10, 0.20 amp is ok in this circuit. A TO-92 type transistor that drives that fan can handle 400- 500mA but could become quite hot if the fan draws 0.30- 0.40 amps.
The power supply has adequate protections on the secondary so if voltages go way out of spec it shuts off.
CTC caps are prone to fail when close to heatsinks and colis. Bulging top vent or leaking electrolyte is a visual indication when they must be replaced. If they seem fine then leave them for now...
I have 2 PSUs I want to rebuild and use:
The first:
This is one of these utterly generic PSUs, however, they look to be mostly intact and not price reduced or under-built for it's wattage.
Replace the dodgy bloated capacitors, it will probably work great for years.
Lol, looking at this PSU, while it does have some input filtering, the entire thing looks like it has been on a huge diet and is looking excessively anorexic. I'm not going to bother with this one anymore.
It is an L & C PSU, so I guess that is all I can expect.
The component psu is FSP. You can tell by the IC.
FSP is a known and reputable OEM. Definitely worth recapping/repairing if needed
Hi Dave!
wrote:Be carefull about the current consumption of the fan. 0.10, 0.20 amp is ok in this circuit. A TO-92 type transistor that drives that fan can handle 400- 500mA but could become quite hot if the fan draws 0.30- 0.40 amps.
Hmm... I will copy the fan regulation schematics to paper when the Schaffer can arrived and I open the case again.
However I have some doubt about the usefulness of the regulator, as it is now, with sensor far away from heat-emitting parts.
With the schematics it will be easier to modify the regulator circuit for bigger fans (bigger transistor together with cooler).
Ideally it can be modded in a way that permits the sensor to be attached to the secondary's hottest parts respective the parts that do most damage when failing. I guess this could be the cooler of the secondary rectifier?
wrote:CTC caps are prone to fail when close to heatsinks and colis. Bulging top vent or leaking electrolyte is a visual indication when they must be replaced. If they seem fine then leave them for now...
They still look fine, no visible bulging or leaking.
But some of them are in very hot areas!
I'll proactively replace them when the filter arrived.
Thank you again!
I now feel relieved as this is apparently no PSU of the type that is likely to take out the whole computer 😀
wrote:The component psu is FSP. You can tell by the IC.
FSP is a known and reputable OEM. Definitely worth recapping/repairing if needed
Yeah, you can often tell by just having a quick look in a PSU, whether a PSU is a complete gutless wonder or not. Things like missing input filters, using 4 x diodes rather than a bridge rectifier, wafer thin heatsinks, more board than component (like the L & C board).
If you find PSUs that are well built, but otherwise have poor capacitor choice, it can be worth replacing the capacitors to extend their life. If the PSU isn't well built, it will always be an unacceptable risk to whatever you hook up to it.
wrote:Ideally it can be modded in a way that permits the sensor to be attached to the secondary's hottest parts respective the parts that do most damage when failing. I guess this could be the cooler of the secondary rectifier?
This psu is 70% efficient at best. Still running after 20+ years and there aren't any burn or scorch marks on the pcb and components hints that it wasn't running too hot. The thermistor is sandwitched between the base of the of the secondary output current stabilizer and a Pi coil. Those can get almost as hot as the heatsink with the two diode packs.
Once you reverse engineer the fan circuit we can estimate the thermal profile of the controller.
got this 200w '95 astec VL201 here, is it worth the hassle of a good clean with compressed air? caps are all nichicon which seems rare for AT PSUs of that vintage but this one seems downsized compared to pics of some other astec PSUs i've seen.
specs are: +12V 8A, +5V 20A, -5V 0.3A, -12V 0.5A
wrote:got this 200w '95 astec VL201 here, is it worth the hassle of a good clean with compressed air? caps are all nichicon which seems rare for AT PSUs of that vintage but this one seems downsized compared to pics of some other astec PSUs i've seen.
specs are: +12V 8A, +5V 20A, -5V 0.3A, -12V 0.5A
Astec units are well built, with good quality components as you can see. Definitely worth cleaning it, although I would probably replace the fan (especially if it rattles or sounds like it's about to die).
Don't worry about it being "underpowered", those ratings are more than enough for pretty much any PC from the AT era (unless we're talking about servers or workstations with many drives).
Ahoy, ahoy,
I'm wondering about this PSU I have in my portable PC. It's been pointed out to me that bad caps can cause all manner of discomfort for the system as a whole, and I have this bugger under suspicion of causing graphical issues (ripple).
Now, this PSU is a pretty specific form factor, kinda T-shaped and quite small at ca. 15x10x10 cm and with the fan facing the other way from the power switch and connector. It doesn't even say how many Watts it has on the sticker (I guess that should be ~275W), so I'm generally wondering if this is a decent quality PSU that I should keep and attempt to recap, or if I'm better off finding a replacement? Any idea what this form factor is called? I've looked around but it can't be all that common...
I imagine it could be somehow possible to make the guts of a more modern (and potentially reliable) ATX PSU fit into this chassis and use an ATX-to-AT adapter, or mod the whole PC case to accomodate a PSU with a different form factor! And is there a certain ATX form factor that's pretty small would be suitable for replacement?
The PC case in question also doesn't have a power button, so it would either have to be a PSU with a power button, or I have to install one somewhere, which would mean more alterations.
Hmmm, that PSU actually looks pretty good. I wouldn't bother recapping it as Rubycon capacitors rarely fail without serious provocation.
it might be causing your problems however, you have to test with a different PSU.
wrote:Hmmm, that PSU actually looks pretty good. I wouldn't bother recapping it as Rubycon capacitors rarely fail without serious provocation.
Okay, thank you for your input regarding the caps.
wrote:it might be causing your problems however, you have to test with a different PSU.
I have an Astec ATX PSU coming to determine just that!
This PSU is definitely of exceptional quality, as the FR4 epoxy board indicates.
On the third pic, the big cap could be bulgy. Does it feel like being under pressure when you press your finger at its top?
Hmmm, the primary side capacitors rarely fail in PSUs, and I didn't notice any significant bulge, so I'd bet money on it being perfectly fine.
wrote:On the third pic, the big cap could be bulgy. Does it feel like being under pressure when you press your finger at its top?
Hm, I tentatively pushed down on the top of the cap but there is no give if that's what you mean.
