Annoying heavy load-failure in Tecnoware Free Silent 500W

[momaka]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

No, I have no output voltage, plus I noticed that the voltage on one of the large capacitors is much lower than 200V, it signals me about 100 volts with the power supply running

Since you are in Italy and you have 230V AC line voltage, the total *DC* voltage across the two big capacitors should be about 230 x 1.41421 = 325V DC. Thus, each cap should have about 160-170V DC across it, with or without the PSU turned On.

If you do, check 5VSB output (i.e. purple wire on ATX connector to ground.)

If no 5VSB, then the rest of the PSU won't work. Need to throubleshoot the 5VSB first. As stj suggested, it's probably a good idea to start a new thread and post pictures and any information there so that people don't get confused with the problems of the PSUs in this thread.

**EDIT**
I see you started a new thread here:
https://www.badcaps.net/forum/showthread.php?t=83829
Linking for anyone else that might be interested.
Will reply there too.

[falcontech]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

with the power supply not turned on on the large capacitors I have the same voltage, theoretically it should be less or am I wrong in standby?

[falcontech]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

No, I have no output voltage, plus I noticed that the voltage on one of the large capacitors is much lower than 200V, it signals me about 100 volts with the power supply running

[stj]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

a new thread would have been good,
so is the 5v standby working?

[falcontech]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

Hello everybody
I also have a problem with this power supply, unlike omega it just won't turn on and I don't have any output voltage, I replaced some capacitor that was open, I checked the shotty diodes on the secondary and I'm ok, try do you recommend me to do?

[Behemot]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

While crapacitors are always a risk, new platforms very often die because of another reasons than caps. And with warranty starting at 5 years for some platforms, it does not matter that much - not in this price range.

momaka: Italy is usually given as an example of the "best economics" in Europe so I somehow do not believe there are no good units in there. (however it is more like fascistic hell where you pay tax when you have two chickens on your own backyard, and we should also look at how much money flows from northern countries to southern countries, which only make more debts continuously). They may be expensive because taxation in there reaches 80-90 %, but that's their own problem…after all, they kept pigs as Berlusconi in power for decades. On their defense, situation is similar in all the other socialistic countries in Europe.

[omega]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

Don't trust SpeedFan or any other software for your PSU voltages (including BIOS readings). Take a multimeter and measure the voltages of the PSU in the computer as it is running. That's by far the most accurate way.

I see. I will. But in the meantime I could compare in terminal service the voltage curves of FAL501FS12 and VS450 (by means of SpeedFan). Even if the absolute values are nonsense, the latter is far more stable, even if it is full of CapXon caps!

Check the resistors I circled in red above in one of the "good" PSU and see how they compare (resistance-wise) to the ones in the "bad" PSUs.

I am almost sure that the resistors are of the same values except the R12 which is missing.

I agree with you. But what applies to your locale may not apply to his. I've been to various places in Southeastern Europe, and I can tell you there are remote parts of the countries there that you literally have no choice - everyone is running crappy power supplies, and that's the norm. Getting parts is often cheaper than buying a reputable PSU. Not to mention gas prices - if you have to go to a city far away that sells those better PSUs, the cost of gas along is a big turn-off.

Money is undoubtedly important, but it is not everything. From this experience I learned so much about PSU, and this is also important. Now I know that even new middle-level PSU from known brands are full of crap caps which have to be recapped soon or later, therefore it has been right time to begin to master the technique.

[momaka]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

I think I succeeded in finding it (that's the good news): at present the PSUs which have worked bad after fixing all have rail voltages, as measured by the program SpeedFan (v 4.51), which exceed 5% the nominal 12V, i.e. > 12.6V (+/-5% the nominal is it the specification you meant, right?).

Don't trust SpeedFan or any other software for your PSU voltages (including BIOS readings). Take a multimeter and measure the voltages of the PSU in the computer as it is running. That's by far the most accurate way.

SpeedFan mis-reports voltages on many of my computers, and I have quite a range of different computers, so the problem is not limited to one motherboard brand or model at all.

I think that may be due to the resistor R12 (470k) missing in some of the recapped PSUs (it has been factory cut, there are still soldered rheophores), probably to increase the then low 12V rail. Indeed, the recapping and replacement of the fast recovery diode with a Schottky one, have increased the voltage of about 200-300 mV, maybe going beyond tolerance. Since I got at home a tenth PSU to recapped and in this R12 is missing too, I measured 12V rail (under 21W load) before and after putting a 470k (still no recap) and found a lowering of 600 mV. Good!

Good work.
Check the resistors I circled in red above in one of the "good" PSU and see how they compare (resistance-wise) to the ones in the "bad" PSUs.

Next week I will check at work if the correlation between ATAPI error and the lacking of R12 holds and fix the PSUs accordingly.

Yes, let us know.

I told ya there may be voltage problems and that it may be better to use two superfasts rather than single schottky.

The problems (i.e. HDDs getting errors) were there in the first place even before he installed the schottky rectifiers.

Anyway, I still somehow don't understand how putting SO MUCH effort, time (=money), components (=money) and running such inefficient craps (=electricity=money) is better over buying new 350W 80+ (80+ Bronze) cheap units with 3year warranty. There are already shitloads of them in Europe and under competition pressure the warranties are getting longer even throughout mainstream and low-end.

I agree with you. But what applies to your locale may not apply to his. I've been to various places in Southeastern Europe, and I can tell you there are remote parts of the countries there that you literally have no choice - everyone is running crappy power supplies, and that's the norm. Getting parts is often cheaper than buying a reputable PSU. Not to mention gas prices - if you have to go to a city far away that sells those better PSUs, the cost of gas along is a big turn-off.

[Behemot]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

I told ya there may be voltage problems and that it may be better to use two superfasts rather than single schottky.

Anyway, I still somehow don't understand how putting SO MUCH effort, time (=money), components (=money) and running such inefficient craps (=electricity=money) is better over buying new 350W 80+ (80+ Bronze) cheap units with 3year warranty. There are already shitloads of them in Europe and under competition pressure the warranties are getting longer even throughout mainstream and low-end. So I bet even in lower mainstream, by the end of 2016 there will be units with 5y warranty.

[omega]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

This to post bad and good news together
So far I recapped nine Technoware FAL501FS12 (just the secondary section), with replacement of the 12V rail diode as momaka suggested me, and had also time to test them "in the field" under stress load (100% CPU 24/7), as usual in my field of research.
The bad news is that some work, some not (namely I get the infamous ATAPI error 9 and 11). After checking the soldering of every "bad" PSU, have been trying for hours to give myself an explanation through a correlation and today I think I succeeded in finding it (that's the good news): at present the PSUs which have worked bad after fixing all have rail voltages, as measured by the program SpeedFan (v 4.51), which exceed 5% the nominal 12V, i.e. > 12.6V (+/-5% the nominal is it the specification you meant, right?). I think that may be due to the resistor R12 (470k) missing in some of the recapped PSUs (it has been factory cut, there are still soldered rheophores), probably to increase the then low 12V rail. Indeed, the recapping and replacement of the fast recovery diode with a Schottky one, have increased the voltage of about 200-300 mV, maybe going beyond tolerance. Since I got at home a tenth PSU to recapped and in this R12 is missing too, I measured 12V rail (under 21W load) before and after putting a 470k (still no recap) and found a lowering of 600 mV. Good!
Next week I will check at work if the correlation between ATAPI error and the lacking of R12 holds and fix the PSUs accordingly.
The bottom line: electronics repair may be either frustrating or funny, however all is best after one finds good teachers. Thanks to the group members and in particular to Momaka!

[momaka]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

I can also see that a place on the left of the red square you sketched is EMPTY. If I am remember, it is NOT missing, but this pic has been taken from one of the two PSUs with high voltage. I will check it soon. Anyway they appear to be 1% precision (namely hard-to-find) resistors.

Yeah, I noticed that as well. Looks like there was a resistor (R12?) in there originally and it was cut off at the factory perhaps? Looks like it's part of the compensation network, but I don't know what exactly it does or what to put in there.

One easy thing you can do, however, is take one of the PSUs that you fixed that has normal voltages and compare the resistors it has to the PSUs that have the abnormal high voltage condition. Let me know what you find out.

Also, don't worry that the resistors are 1% precision. You could use 5% too if the 5% resistor happens to measure within 1% accuracy. The manufacturer only used 1% so that they don't have to take time tweaking the PSU outputs if the outputs weren't exactly spot-on right.

I am talking about measuring output power with meter on input. That is guessing at best as you have no idea what's the efficiency of that thing.

You're right, I am guessing indeed . But for an old half-bridge design, you can assume the efficiency to be around 70% +/-5% most of the time and your results won't be too far off. If anything, you should get a pretty decent *estimate* of the power consumption - usually within 10% of what it is. I know that is not precise at all, but it does give you an idea more or less how much power your PC is using.

[Behemot]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

I've already done that with a few random SMPSes of mine, including some ATX PSUs, and got reasonable results as I did above. Anyways, we've already had a thread on this stuff, so I won't get into this discussion here.

I have a strange feeling you are talking about something completelly different.

I am talking about measuring output power with meter on input. That is guessing at best as you have no idea what's the efficiency of that thing. Compared to that, even cheap clamp meter with like +-3 % accuracy is still much more precise.

What the wall meter actually shows is even another, completelly different thing (which also affects the result, however).

[omega]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

Momaka, thanks you so much for the steady support!

Speaking of which, what do you mean by, "nominal before recap, 21W after, 21W after, no load"?

sorry a comma has been dropped out!
nominal is the value which has been printed n PCB
before (after) recap is before (after) the cure, with 21W load (or no load).
It's a pity that I cannot attach a Table

I think you *may* be able to fix that, if you're talking about the Technoware PSU revision in post #14.

Yes, they all are FAL501FS2

With some tracing on your PSU, I believe the resistors circled in red in the picture below is what sets the output voltage.

I can also see that a place on the left of the red square you sketched is EMPTY. If I am remember, it is NOT missing, but this pic has been taken from one of the two PSUs with high voltage. I will check it soon. Anyway they appear to be 1% precision (namely hard-to-find) resistors.

[RJARRRPCGP]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

Notice it's total system draw in the Anandtech graphs, not CPU draw alone.

65nm Kentsfield Q6600 are 95~105W TDP. 45nm Yorkfield Q9505 are about the same, 95W TDP.

I believe the motherboard and the GPU only add a tad more, unless doing 3D rendering with the GPU...

It does seem to be exponentially higher, if not exactly that and without exaggeration , it's easily about 230W alone, IIRC, when I saw one for x264, which was a CPU-only test...

3.0 OTOH seems to use exponentially less and that's why I use that for the system with my Halo CE server...

I think the 45 nm version is easier on the system...

[TELVM]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

... I'm talking about a 65 nm Core 2 Quad. Way before i-series!

http://www.anandtech.com/show/2303/4

Notice it's total system draw in the Anandtech graphs, not CPU draw alone.

65nm Kentsfield Q6600 are 95~105W TDP. 45nm Yorkfield Q9505 are about the same, 95W TDP.

[momaka]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

Most of the recapped follow the voltage schema depicted in post #56, however two PSU get following similar voltages:
nominal before recap, 21W after, 21W after, no load
12 V 12.4 V 12.7 V 12.9 V
5 V 5.5 V 5.4 V 5.4 V
5 VSB 5.1 V 5.1 V 5.1 V
3.3 V 3.4 V 3.3 V 3.4 V
–12 V –11.9 V –12.0 V –12.0 V
PG 5.4 V 5.3 V 5.3 V
I think the voltages are high with respect the average, am I wrong? any clue how comes and how to fix?

Yes, they are quite high (the 12V and 5V rails, that is). In fact both are out of spec on the last two columns. Speaking of which, what do you mean by, "nominal before recap, 21W after, 21W after, no load"?

I think you *may* be able to fix that, if you're talking about the Technoware PSU revision in post #14. That one uses an SDC2921 PWM controller. This is its datasheet:
https://cdn.badcaps-static.com/pdfs/...1d9ec648b2.pdf

According to the typical application on page 9, it appears that pin 16 (VIN+) sets the voltage output through variable resistor "R2". With some tracing on your PSU, I believe the resistors circled in red in the picture below is what sets the output voltage.
https://www.badcaps.net/forum/attach...1&d=1454552789
All 3 of them appear to be in parallel, and that is what makes up "R2" on the SDC2921 datasheet. Can you please post the resistance values of each resistor? You will need to remove them out of the PSU if you want to measure them.

[RJARRRPCGP]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

^ Nope you're misinformed RJARRRPCGP, it'll only draw ~135W @ 3.9 GHz. See here.


Even an i7-950 @ 4.2 GHz is ~50W short of drawing 250W:

That's a 45 nm Core 2 on the right, which uses a lot less power.

I'm talking about a 65 nm Core 2 Quad. Way before i-series!

http://www.anandtech.com/show/2303/4

Mine required more Vcore than theirs for 3.3...

[TELVM]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

^ Nope you're misinformed RJARRRPCGP, it'll only draw ~135W @ 3.9 GHz. See here.


Even an i7-950 @ 4.2 GHz is ~50W short of drawing 250W:

[RJARRRPCGP]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

Also, the "500W" rating is absolutely fake! At most, this power supply will give you 200 Watts raw power.

QFT! And if you OC a Q6600 to 3.3 with 1.4V of Vcore, the CPU alone will use 250W easily!

[omega]

Re: Annoying heavy load-failure in Tecnoware Free Silent 500W

Back to the main subject of this thread, now that I 've recapped nine PSUs FAL501FS12 I have some statistics of the voltages (most ATX are already under stress, I will let you know about the results). Most of the recapped follow the voltage schema depicted in post #56, however two PSU get following similar voltages:
nominal before recap, 21W after, 21W after, no load
12 V 12.4 V 12.7 V 12.9 V
5 V 5.5 V 5.4 V 5.4 V
5 VSB 5.1 V 5.1 V 5.1 V
3.3 V 3.4 V 3.3 V 3.4 V
–12 V –11.9 V –12.0 V –12.0 V
PG 5.4 V 5.3 V 5.3 V
I think the voltages are high with respect the average, am I wrong? any clue how comes and how to fix?