Learning with ATX CHIEFTEC 450W PSU

[Per Hansson]

Re: Learning with ATX CHIEFTEC 450W PSU

It is disabled, but as stated in the datasheet you can't be sure it has 0% duty cycle.
It is better to desolder the OVP/EN pin as that has the sole purpose of enabling the circuit.
Every power supply will be different, this only qualifies for a supply with a UCC3818N APFC controller.
Besides even if it works with APFC turned off you have no way of knowing that it will not blow up when you enable APFC, as you are not testing those components anyway.

[stj]

Re: Learning with ATX CHIEFTEC 450W PSU

why not just desolder the apfc mosfet?

[kaboom]

Re: Learning with ATX CHIEFTEC 450W PSU

If it hasn't been mentioned already, make sure you're using the correct common point!
If you're shunting that pin of the APFC IC, it will go to the - side of the line rectifier!
Not secondary DC common or the "third pin" of the AC line cord!
Even with the dimbulb, you're likely to cause damage by using the wrong "ground."

I would've just gone with a bigger bulb & not bothered attempting to defeat the APFC, at least til I had a "clearer picture" of what I thought was going on.
Too easy to initially get a false lead and possibly damage something...

[harp]

Re: Learning with ATX CHIEFTEC 450W PSU

Occasionally when I take soldering iron, the components start screeming, because it cannot run and hide from me - they are rooted to pcb.

Ok, I desolder all pfc board, in the name of science, and...
...and, this time bulb is not briefly flash like before, but it has immediately go to fully on! Just like that! On!
I do not have chance to connect ps-on this time.

I was thinking, that is it, for this point it is enough of "new learning" (read: stay torturing psu), because its go to way of destruction... A little break is needed for me

Why not pull out mosfet? Mosfets are screwed into heatsink, lots of legs desolder, solder back, desolder, thermalpaste, screw, solder, bridges... We are seeking for elegant solution here and now, not crude way
But if you know that this method seamlessly work in practice to cancel pfc, please confirm that. I am interesting in that info, it is still open question.

I use the ground from pin #1 to pin #13, it has been nearest to me

I am a little dissapointed that test like this are not common and well known in repair scene, and at the same time also a very gratitude to all of us.

But it is ok, I will solder everything back and do final test. I hope it will work. I learn a lot, dont mess with a pfc literaly, its a black hole

[Per Hansson]

Re: Learning with ATX CHIEFTEC 450W PSU

I am a little dissapointed that test like this are not common and well known in repair scene, and at the same time also a very gratitude to all of us.

As I explained in my previous post it will be different for each PSU design.
What works on this one will obviously not work on another with a different APFC controller and there are many of those!

It will also be different based on your line voltage, you can't just disable the APFC if you have 115VAC line voltage, it only works with 230VAC.

Here is a post by someone claiming removing the APFC board disables it on this PSU:
https://www.elektroda.pl/rtvforum/topic1394308.html

[harp]

Re: Learning with ATX CHIEFTEC 450W PSU

Ok, I am in learning mode again


So what am I do, sistematicaly:

- everything put back like to be so before
- check for short, ac input, outputs (it is low resistance 25R on 3.3 and 5v rail)
- just like last time, I put 60w bulb to compare results, and everything is absolutly the same, nothing difference, so i not damaging more with my testing:

-> when plug ac, the bulb slightly briefly flash (no output load)
-> in that step I have clear and stable 5.1v on 5vsb and 5.1v on pwron, all else is 0v
-> connect pwron, and bulb go to full light, a next second I hear litle pfc sound...
-> when is bulb on, I measure:
- orange 3.4v
- red 5.1v
- yellow 12.0v
- blue -12.0v
- black 0.00v
- grey: one moment after pwron is 5.1v, then go to 0.07v and stay there
- fan is start spinning, not maximum speed, but not lazy, it is quiet

So, I put 100W lamp, and all is same - lamp turn fully on
I put 2x60w, resulting the same - fully on x2
Then, 60w + 100w in parallel, the same - fully on... instantly, both of them.

Suggestions is wellcome

[Per Hansson]

Re: Learning with ATX CHIEFTEC 450W PSU

What voltage do you read on the primary capacitor during this test?
If your line voltage is 230VAC then it should be around 230VAC x √2 = 325VDC in standby.
And when you start it up it should be around 400VDC but exact voltage depends on the design.
All voltages you posted look fine to me so it can't really be shorted.
And this also means that the lightbulb can't be fully on: because if it were then there would not be voltage left over to run the power supply...
You can use a wattmeter to check how much the whole setup draws to know if those lamps are really fully on or not.

[harp]

Re: Learning with ATX CHIEFTEC 450W PSU

I test with 160w bulb setup (100 + 60 ||) below.
I do not have wattmeter, but I measured AC current, in standby is 0.01A, and when pwron is 0.65A steady.

In standby, main cap C1 is charged with 325V, and when prwon voltage on it is rappidly decrease (in cca 5 seconds) to 120V, and stay arround there, when bulbs go flicker, and hear pfc sounding.

I notice that voltage dropping is not present only on 5vsb rail, it stay ok, (actualy it goes from 5.14v in standby to 5.20v when all other collapse). All other 12v, 5, 3.3 after few seconds (so it is not stable like I was mention before while I measure very short time after pwron) all go drasticaly down (about 1/4 lower) and fluctate a little bit at this point. That is reason why PG very soon go down.

Problem with apfc, supply circuit or need more power?

[Per Hansson]

Re: Learning with ATX CHIEFTEC 450W PSU

I would say it is still starved for power, but you really should check the integrity of the APFC swiching mosfets (Q11/Q12 & Q1/Q3 from the schematic kaboom posted) if you do suspect that the supply is not ok.

[harp]

Re: Learning with ATX CHIEFTEC 450W PSU

I checked in circuit all mosfet and transistor, and it is ok.

So, next load I found is - no more, no less - toaster with 700w power
I try to find something arround 300w but I can not.
Serialy with it I was connected 20A multimeter to monitor behaveour.

In standby, it is 0.01A, and when ps-on connect (unloaded), it power up and draw 0.05A and a very litle fluctated last digit. All voltage are stable and correct this time, including the PG. It is repaired and tested now, some learning has occour, very happy and thanksfull for your help

[Per Hansson]

Re: Learning with ATX CHIEFTEC 450W PSU

Glad to hear it and thanks for posting back with this info, I'm sure it can help someone else