Enermax EG651P-VE FMA (again)

[JEWilson]

Have a recurrent problem with this PSU.
Recently recapped all with Pana FC and Samxon.

However, the +5.0VSB rail on my go/no-go tester consistently
reads 5.4V and is out of regulation.
I have taken the +5.0VSB cct to pieces and cannot find
any faults with parts.

I have measured the voltage with a meter at the ATX plug.
It reads 5.3V, so is still out of regulation.
Also, when the PSU switches off, it make a tick sound and
it is only one tick only.

Any suggestions from the forum members?

[kc8adu]

Re: Enermax EG651P-VE FMA (again)

may be ok.
try it with a load and see if it is on spec.

[JEWilson]

Re: Enermax EG651P-VE FMA (again)

Interesting...

The rails for this PSU detail;
Rail O/P as found on PSU label
+12.0V : 4.0 A min., 24.0 A max.
+ 5.0V : 11.0 A min., 46.0 A max.
+ 3.3V : 0.0 A min., 40.0 A max.
- 5.0V : 0.0 A min., 1.0 A max.
-12.0V : 0.0 A min., 1.0 A max.
+ 5.0V SB : 0.0 A min., 1.8 A max.

+5.0/3.3V Combined - 300 W max.
All Combined - 551 W max.

Used only 4 HDDs as a load for starters.
At start-up, the go/no-go tester beeps with +5.0VSB flashing
at 5.4V. This reads 5.3V with the meter.
Also, as this PSU has a 46.A 5.0V rail typically at startup
it will record 5.3V on the tester but with a load such as
above, this will drop to 5.2V on the tester readout.

Curiously, the flashing and beeping stops after abount 20
seconds and the meter reads 5.28V for 5.0VSB.

Is this a matter of loading?
Would it be the case, in service with the appopriate load,
a mobo, optical drive and HDDs present might solve this
issue?

I did not want to risk using the PSU on the basis it was faulty.
Also, another PSU (same model) post recap did not
exhibit the +5.0VSB issue other than reading 5.3V for the +5.0V rail.
With a minimal 4HDD load, this drops to 5.2V.

Hence, I was suspicisous it was something I did.

It may be now, the way forward is to test it with a full
system and get readings with the meter.

Next, I tried 7 HDDs and this had the same effect, after
10-15 secs, the alarm for +5.0VSB stops, reads 5.3V
and 5.28/2.59 on the meter.
This is still out of regulation even so.

Should I consider anything else?
Is this a common issue with recapped PSUs?

Just a thought - might it be one or both of the new caps on
the +5.0VSB rail are bad?

[linuxguru]

Re: Enermax EG651P-VE FMA (again)

Check the primary-side electrolytics on the aux flyback circuit - there may be one in the range 22-100 uF, 50 V which is charged from one of the windings of the aux flyback transformer, through a rectifier.
Replace it with a slightly larger value, long-life series.

[JEWilson]

Re: Enermax EG651P-VE FMA (again)

This PSU does not appear to have that feature.

However, at the time of recap, there was a 22uF 50V PCE-TUR
which was tied to pin 7 Vcc of the UC3842BN controller.
I replaced this with a 33uF 50V Pana FC.

There is further, a 2.2uF 50V PCE-TUR which is tied to
the Vfb pin, pin 2 of the controller.

I have looked at other possibilities and it has been suggested
the out of regulation condition as described may refer to a weak opto-
isolator, a Q817B.
This tested ok out of cct., so I will replace and determine if 5.0VSB
comes back within regulation.

I also replaced the TL431C shunt regulator as well with a new one
as I suspected this may be at fault.

Will report back...
Thx for the help

[linuxguru]

Re: Enermax EG651P-VE FMA (again)

> there was a 22uF 50V PCE-TUR
which was tied to pin 7 Vcc of the UC3842BN controller.
I replaced this with a 33uF 50V Pana FC.

That's probably the right one, and your replacement is fine. Check the voltage at Vcc also - it should not be below 10-12V, usually 15V. If it is 8V or below, there's most probably a fault in the UC3842.

Be very careful when probing here - this is on the live primary side.

[Gianni]

Re: Enermax EG651P-VE FMA (again)

Hi JEWilson

I don't know much about ATX PS circuit because I have never worked on them, but I know UC3842 and flyback PS.
If the output voltage is obtained by UC3842 and secondary feedback with TL431 and optocoupler, the output voltage must be in regulation.

I would check the resistor divider connected to TL431, they should be connected to the output and they should have the same value to obtain 5V.
So the only reason I can think about wrong regulation is a problem with divider, maybe one resistor has been damaged and its value has changed. I don't know if somewhere in the circuit there are ceramic caps (100nF - 1uF) but it happen to me that some faulty ceramic caps caused bad regulation.
The resistor used to drive the optocoupler trough TL431 should be checked too.

On the reference pin of TL431 there must be 2.5V.
Check the 2.2uF connected to Pin 2 and caps connected to Pin 1, they are in the control loop and can cause problem.

Check also tracks and connection.

Good luck
Gianni

[JEWilson]

Re: Enermax EG651P-VE FMA (again)

The voltage divider is fine.
I have taken the resistors out and physically
checked these re - colour codes and have checked
ratings with a MMeter - these are all ok.
There is an orange cap at the TL431 between
the ref and the cathode. I suspect this is a ceramic
disc cap. I have not checked this yet.

Also, there are a number of caps at the controller
end, I will check these in turn. These appear
to be tantalum bead caps other than the lytics
referred to. These are marked, so should be able to check ok with MMeter.

The opto-couplers arrive tomorrow.
Will change this first.

Thx.

[arneson]

Re: Enermax EG651P-VE FMA (again)

Optoisolators are a pretty hardy device as failures go.
I've seen a bad one when one ten was sent thru a DC circuit.
They're very easy to test with simple meter and are generally a go no go thing.
http://www.electronicrepairguide.com/optocoupler.html

[JEWilson]

Re: Enermax EG651P-VE FMA (again)

I've tested the opto-isolator with my MMeter.
This checks out ok (BTW - don't have an analog meter).

I've have checked all of the tant. caps on the monitoring
side of the PWM controller and these are ok.
The remaing 2.2uF 50V lytic checks out ok and has an ESR 1.9.
Nothing unusual there.
There is also what appears to be a mylar polyester cap 2.2nF
100V as well; this appears to be ok also.

I have checked the tant. cap between the ref and anode of
the TL431 on the 5.0VSB, this is a 1uF part and is ok.

As to the matter of opto-isolators, I have had these go
with a shorted primary side chopper but have read, where
the device checks out ok but may not work to spec in cct.,
this can account for out of regulation conditions such as with
the fault condition I am experiencing.

Remedy is to replace the part and test again.
This I will do later Friday and will test the relevant pins on
the PWM controller as suggested by linuxguru.

Thx.

[Wizard]

Re: Enermax EG651P-VE FMA (again)

I had a UC3842 go weird. Replace it, and change the output capacitor for that standby supply.

Sometimes I take out the capacitor and esr this removed cap again to be SURE.

Cheers, Wizard

[JEWilson]

Re: Enermax EG651P-VE FMA (again)

As to the +5.0VSB pi filter
This has a 10V 1000uF and 100uF 16V Pana FC.

As I understand it, top switches don't like anything more
than 1000uF. I believe a UC3842BN can operate with
higher capacitances than 1000uF.

Would this be correct?
In the first instance, I had 2 of 1000uF in the pi filter
for the +5.0VSB but changed this back to the above
as the supply was experiencing start up issues.

Might this be a root cause of the PSU subsequently
presenting +5.0VBSB reg issues?

[Gianni]

Re: Enermax EG651P-VE FMA (again)

As I understand it, top switches don't like anything more
than 1000uF. I believe a UC3842BN can operate with
higher capacitances than 1000uF.

Would this be correct?

AFAIK it depends on the regulation loop and current limit threshold. If the feedback is connected on the output caps after output inductance and you change the caps, from 100uF to 1000uF, it is possible that you change too much the parameters on the feedback loop and you introduce some instability so the the control circuit is not able to start correctly.

For the regulation problem: you have checked the output divider, are the resistor of the same value? If they have the same value, on the reference pin of TL431 you must have a voltage that is out of spec. (TL431C MIN 2440mV TYP 2495mV MAX 2550mV) at 2650mV.

If you find this voltage, than you can check if TL431C is drawing max current from optocoupler, the TL431C cathode pin should be at very low voltage, I think around 1V.
In this case there's something wrong on the primary side or the optocoupler is not working correctly.


Ciao
Gianni