Is an L&C power supply suitable for a Celeron 2.2GHz PC?

[Newbie2]

I added the coil, I think I can keep the four diodes there since this PSU really isn't going to be used in a more demanding system than a Celeron 2.2GHz computer.

[momaka]

You work faster than I can type . That L&C looks much better now.
----
Those diodes limit you to about 165 watts (1.5A x 110v)... if they even make it that far without heating up and derating like crazy. You already have the power supply opened, just replace them anyways.

[Newbie2]

Originally Posted by momaka

You work faster than I can type . That L&C looks much better now.
----
Those diodes limit you to about 165 watts (1.5A x 110v)... if they even make it that far without heating up and derating like crazy. You already have the power supply opened, just replace them anyways.

I do have a nice bridge rectifier sitting in the power supply, when I get the time I'll replace the diodes.

[370forlife]

a bridge rectifier will make it a lot better, but still needs some more beef for the 12v.

Did you add that MOV and two y caps?

[Newbie2]

Where does an MOV go on this power supply?

My father is an experienced electrical enginner with 10+ years of experience and he said that the additional power inductors on the secondary side and the bridge rectifier on the input side is not necessary for the application I will use it in. He said replacing the two 100V 3A FR302 diodes on the 12V rail and adding the filtering components to the input side were the only things necessary.

[370forlife]

Unless you are not going to draw more than about 150W, it isn't. But those diodes are only good for a little bit, and as they heat up that rating goes down.

They are close to the pcb and they will build up heat quicker.

The inductors on the secondary wouldn't be necessary if it is used in a application where 65mv (the max for the atx spec is 120mv) of ripple is not good.



The MOV is that green disk next to the X cap.

[Newbie2]

Well this Celeron 2.2GHz PC isn't a very demanding computer in terms of power consumption so I'd guess the four diodes will suffice, even if they will generate more heat.

The MOV and the two blue Y caps were already there, all that had to be added was the coil and X cap.

With such an old and obsolete computer, Celeron 2.2GHZ, ripple isn't an issue here with the L&C. If the components in it start failing, I wouldn't mind as my main desktop is the only important PC I own, and it's powered by a good Antec EarthWatts 380W power supply (a Seasonic rebranded PSU).

This is my "junker" computer, it's slow and thus doesn't deserve a great power supply.

[momaka]

Originally Posted by 370forlife

The MOV is that green disk next to the X cap

No that's a NTC thermistor. It limits the inrush current when you flip the power switch so the fuse doesn't blow.
MOVs are only for extra protection against power surges, but they're not really necessary if there's enough coils and X/Y caps for filtering.
----
Newbied2: You have the power supply opened. The diodes on the primary side should be a quick job anyways.

[Newbie2]

Okay tomorrow when I get the time the diodes will be replaced.

[Newbie2]

I did put in a 10A bridge rectifier today, but when I powered up the power supply it was making unusual sounds (spzzz-spzzz-spzzz...) I think around the transformer section while it was running. I assumed that wasn't good, and I put back the four 2A diodes.

Rectifier:


The diodes placed back:


After putting back the diodes, the power supply no longer made the unusual sounds.

I'll conclude that I'm done with the primary input side, when I get the parts I will do the secondary side.

[Newbie2]

Originally Posted by momaka

Those diodes limit you to about 165 watts (1.5A x 110v)... if they even make it that far without heating up and derating like crazy. You already have the power supply opened, just replace them anyways.

Since there are four diodes and they're rated at 2A each, wouldn't there be 8A total?

[momaka]

Originally Posted by Newbie2

Since there are four diodes and they're rated at 2A each, wouldn't there be 8A total?

No, because with each half-cycle of the AC current, only 2 diodes are doing the rectification at a time, one for the negative rail and the other for the positive. Therefore the total amount of current is limited by each diode's rating. Since you have the same 2A diodes throughout, the highest current that can pass at any given time is 2A.
That gives you 220 watts to work with on the primary side. If the power supply is 75% efficient [at best], the output will be limited to 165watts.
That actually might be just enough for your system. Would have been nice to have that bridge rectifier, but the buzzing is kind of worrying me too. Not sure why it does that. Did you double-check your solder joints when you put the rectifier?

[Newbie2]

I did triple-check and redo the solder joints three times with no improvement. The buzzing was fixed when the diodes were put back. What would the buzzing mean?

[lucky13]

If the bridge rectifier is a known good piece then perhaps the layout on the PCB is not the usual format (aka diode directions) so the way you install the bridge is actually wrong?

Buzzing noise could be a result of some clipped since wave causing coils to "sing"?

[everell]

WOW! you are really putting a lot into this power supply. To go this far, you should also consider some improvements to the 5vsb circuit. The existing circuit looks a lot like the Bestec mobo killer. You could try installing a PWM chip such as the DM311.

Looking at the schematic of the original circuit, R30 is only ten ohms. The application notes for the optoisolators typically show a value of about 240 ohms. That means your circuit is pulling a wide range of current for regulation, approaching about 30 ma on the extreme edge. So I drew a simple circuit you can breadboard and experiment with to improve your L&C power supply. Might as well make a Cadillac out of it!

[Newbie2]

Originally Posted by lucky13

If the bridge rectifier is a known good piece then perhaps the layout on the PCB is not the usual format (aka diode directions) so the way you install the bridge is actually wrong?

On the PCB of the L&C power supply, there are four circles in the bridge rectifier/four diodes area, which I assumed was for the bridge rectifier. Also, the PCB shows the orientation of the bridge rectifier.

Originally Posted by everell

So I drew a simple circuit you can breadboard and experiment with to improve your L&C power supply. Might as well make a Cadillac out of it!

I'll go ask my father (who's an experienced electrical engineer) to see if your design is good to him and then I'll consider it. It's my "experimental" power supply anyway.

Right now I'm aiming for Toyota if you say Cadillac, simple, reliable, and easy to maintain and repair.

[Newbie2]

Two major upgrades!

One, I decided to put back the 10A bridge rectifier that I had. When I powered it on, the weird sounds that my power supply made with it earlier did not happen. I powered it up multiple times to see. I realized that without screwing the PCB to the case and putting the top cover on, the noise was present. When I screwed the PCB to the case and put the top cover on, the noise is gone.



Also, I was able to get two PI coils from an old power supply and put them in.



I'll get the two diodes replaced when I get the 20A 200V power rectifier, and replace the secondary and the two primary caps with caps of higher value, get the fan replaced, and then my upgrades will be complete!

[370forlife]

Very nice, input side is pretty much finished.

Fix up the secondary and you will have a half way decent psu now.

[Newbie2]

OMG woah at the L&C Technology building that Deer has (photos from Deer website www.deer-group.com):





They must of used the money saved from not adding actual high-quality components in their power supplies to finance their big building and nice fountain!

[everell]

Hello again - I don't want to seem rude, but want to say that even with all those nice improvements to your L&C power supply, if the 5vsb circuit gives trouble, it takes down the whole power supply. The importance of a GOOD 5vsb circuit cannot be overemphasized.

I have a Raptor model LC-B400ATX very much like yours. The critical capacitor in the 5vsb circuit in both your and my power supply is C8, 22 Mfd at 50 volts. If nothing else, change this capacitor due to age and replace it with a low esr capacitor.

I ran some tests on three different brands of power supplies - Bestec, SeaSonic, and the Raptor L&C. For the Raptor, I used two different values of capacitor in the test, 22 Mfd and 47 Mfd (to see if increasing the capacitance would help). The problem is that when the capacitor C8 goes bad (esr increases) the 5vsb voltage goes up. Here is a table to show my results:

ESR series resistance 22 Mfd 47Mfd
6 ohms 5.07 volts 5.07 volts
12.5 ohms 5.08 volts 5.08 volts
25 ohms 5.57 volts 5.58 volts
50 ohms 6.69 volts 6.70 volts
100 ohms 8.65 volts 8.66 volts
150 ohms 9.48 volts 9.52 volts
200 ohms 10.63 volts 10.72 volts

It is interesting that when you go into BIOS and look at your "computer health" voltage readings, the 5vsb voltage is seldom included. But IF you lose 5vsb, nothing works! When the 5vsb goes high, the mother board can also go up in smoke.

I made some oscilloscope pictures of what the signal looks like across capacitor C8 in series with these resistors. I can post them if you are interested.