Task TK-940TX-DF Blowing fuse

[momaka]

Re: Task TK-940TX-DF Blowing fuse

So did you try measuring the voltage across the two big caps when the PSU is plugged in and the bulb glowing?

No, but I do have a 16V 4.5A laptop adapter though.

Yes, that will be fine. Let's see what results you get above for the voltage across each cap, though. Then we will see where to go from there.

FF 48.0.2 - the last non-ESR version for socket A processors. (although i'm currently using Windows 7)
But it (as well as Basilisk, Palemoon, and even k-Meleon 76.2) returns no response when clicking on Reddit's "View entire discussion" button - for example, on this page:
https://www.reddit.com/r/Amd/comment..._cool_n_quiet/
Does it work for you, with your Opera 12.18?

No, it didn't work for me in Opera 12.18
In general, Opera 12.18 tends to not work well with new websites. But for older/simpler websites like BCN, it is a lot more efficient (can open 20-40 tabs, and RAM usage is still very small, not to mention pages are very responsive.) So I just use it alongside FF.

I tried the above link with Firefox 52.4 ESR Portable on Windows XP, and had not problems clicking on the "View Entire Discussion" button.

Some time ago i did trial Quantum, (can't recall it ever crashing or struggling), but i couldn't completely deactivate all of it's "phoning home" attempts.

Anything after FF24 is questionable to me.
For example, all of the ESR and non-ESR versions of FF after 40 have been trashing my CPU usage in random intervals. Not sure what's going on "under the hood" on those browsers, so I just simply don't trust them.

On that note, I use Quantum on PCs that don't have any personal info. I couldn't care less if the PC is tracked or even hacked. I keep an image for those and can just wipe and copy over if needed. So far, I haven't needed to do that.

Also, I built a test PC last week to test an old socket 754 board. I put an Athlon 64 3400+ and 2 GB of DDR RAM. The 3400+ (2.4 GHz version, 512 KB L2) is in the top 5 CPUs for s754 platform (if not top 3, only slower than the 3700+ and 4000+ mobile.) And while I was able to view 480p content on YouTube without stutter (with FF 52.4 ESR Portable) once the video loaded, it wasn't a smooth experience. For example, minimizing and maximizing windows (especially FF) while I had a music video playing in YT would often produce audio stutter or pause. This doesn't happen on my P4 HT CPUs. So with that said, I think single-core CPUs really are just not a viable option for online browsing anymore. Even a slow dual core CPU will do many times better. Of course, this is where the old FF architecture falls short, as it can't really take full advantage of multi-core CPUs. And that's another reasons I just had to switch to Quantum on my newer PCs. I don't like it... but there aren't that many alternatives really. Chrome? -HA, not a snowflake's chance in hell!

[socketa]

Re: Task TK-940TX-DF Blowing fuse

Took off the white resistor - Bulb still continiously Glowed.
Removed the rectifier and put it in a good PSU, and it goes well - bulb briefly lights
then dims out in about a second, in line with the primary caps charging up.
And the fan starts up when i jump-start it
So it really looks like the rectifier is not the problem

Also do you have a 18-20V laptop adapter

No, but I do have a 16V 4.5A laptop adapter though.

What version of FF? And is it ESR or Quantum.

FF 48.0.2 - the last non-ESR version for socket A processors. (although i'm currently using Windows 7)
But it (as well as Basilisk, Palemoon, and even k-Meleon 76.2) returns no response when clicking on Reddit's "View entire discussion" button - for example, on this page:
https://www.reddit.com/r/Amd/comment..._cool_n_quiet/
Does it work for you, with your Opera 12.18?

Some time ago i did trial Quantum, (can't recall it ever crashing or struggling), but i couldn't completely deactivate all of it's "phoning home" attempts.

[momaka]

Re: Task TK-940TX-DF Blowing fuse

No.
Sorry, should have mentioned that.

So i've removed those other components and got rid of the sooth
...
But the 60W bulb still comes on as seen in the photo below:

Let's keep knocking.
Try also removing the big white resistor.
If bulb is still glowing, remove bridge rectifier and test out of circuit or swap into another working PSU (that you don't mind testing on... or non-working PSU would be fine too, as long as it isn't one that blows fuses.) IDK how it would be possible for this board to burn bridge rectifiers, but there are just no other current paths.

If bulb still glows after the above, check voltage across each 200V cap while PSU is plugged in (be careful there!) and report back what you get.

Also do you have a 18-20V laptop adapter?

P.S. i found an app that looks like it sorts out the Firefox RAM issue - it's called FireMin "-It calls the Windows function EmptyWorkingSet over and over again in a loop to free up memory. Calling the function removes as many pages as possible from the working set of the specified process. "
It also make a DNS request, which i block with the firewall.
So far, so good - have noticed a significant improvement

Cool, thanks for sharing the info.
Currently running with FF ESR on my XP 32-bit machine with 2 GB of RAM, so haven't had issues running out of RAM anymore. I still use Opera 12.18 for the majority of browsing that allows it, though.

And, oh yeah, I always disable DNS Service from Windows Services. Otherwise, my custom HOSTS file doesn't really work. I do this both for Win XP and 7.

There's a bunch more other stuff that I disable. I always forget the specific ones until I see them, but Automatic Updates is probably the first one on that list.

For me, Firefox was stalling and crashing as soon at RAM got to around 2GB (i have 3.25GB out of 4B available, because of Windows 7 32bit RAM limitation)

Interesting.
What version of FF? And is it ESR or Quantum.
I haven't had issues with Quantum even with 2 GB of RAM. Seems to run better on dual core (or better) CPUs.
I only use ESR for Win XP, because Quantum just won't work there.

[socketa]

Re: Task TK-940TX-DF Blowing fuse

Hmm. So bulb is still glowing even with the new bridge rectifier removed?

No.
Sorry, should have mentioned that.

So i've removed those other components and got rid of the sooth (i thought that that may have being burn marks into the board from so much soldering/heat, because it wouldn't come off with isopropyl alcohol and a cotton bud - so as you suggested, i went over the board scratching it off with the test probes (perfect for the job!) - it came off like crusty flakes.
But the 60W bulb still comes on as seen in the photo below:
Well done on noticing the cracked NTC thermistor - it's part number is sck054.

P.S. i found an app that looks like it sorts out the Firefox RAM issue - it's called FireMin "-It calls the Windows function EmptyWorkingSet over and over again in a loop to free up memory. Calling the function removes as many pages as possible from the working set of the specified process. "
It also make a DNS request, which i block with the firewall.
So far, so good - have noticed a significant improvement
There is also CleanMem Pro, and Memory Fox Next plugin
For me, Firefox was stalling and crashing as soon at RAM got to around 2GB (i have 3.25GB out of 4B available, because of Windows 7 32bit RAM limitation)

[momaka]

Re: Task TK-940TX-DF Blowing fuse

I'm pretty sure that i've removed all the components that i just put back on.
Have included a photo, just in case you can spot any that i've missed, or should have removed
But the bulb is still glowing.

Hmm. So bulb is still glowing even with the new bridge rectifier removed?

Looking at the first picture, I see two more components that may be playing tricks on us: those orange "102" ceramic caps by the bridge rectifier (C43 and...?) Remove them permanently and don't put them back on. They aren't really needed.

I also see that there is a lot of black sooth under where the fuse was and the area nearby. You might want to scrape/clean that with some IPA, as the sooth can be conductive.

And while looking there, I see that one of the green disc components next to the fuse holder (looks like two NTC thermistors) has a crack in it. Check and maybe also post part number here, just to confirm what it is.

Lastly, I see a small blue ceramic cap (not the large one) next to the big white resistor in the 5VSB circuit, is still installed. If I recall, that cap is part of the snubber on the 5VSB. Remove it as well.

And there is no short on the DC side.
I thought that the 1MΩ might not be due to a resistor, because i didn't see it on a schematic that i was looking at (attached below) - but it obviously does (and shouldn't there be one of these in the schematic?)
What's the function of that resistor, as i'm picking that it's not a discharge resistor for the primary capacitors, as it looks like R1 and R2 in the schematic would do that job.

The 1 MOhm resistor is there to discharge the X2 cap(s) on the AC side. Otherwise, if you disconnect the PSU from the wall and it so happens that you do this while the AC line voltage is at its highest peak, the X2 caps will be fully charged... or at least until the big electrolytic caps on the DC side start discharging, after which point, the small charge stored in the X2 caps will flow through the bridge rectifier and discharge the X2 cap. But if the above happens, there could be a period of possibly several second where if you touch the metal prongs on the PSU plug after unplugging it, you might get a slight shock. Also, the 1 MOhm resistor is there to provide a current path in case the AC is suddenly disconnected and there is still a current flowing through the common-mode chokes. Inductors don't like sudden interruptions in current, so they always need to be provided a safe path in case of sudden disconnection. The 1 MOhm resistor provides a function for that as well.

[socketa]

Re: Task TK-940TX-DF Blowing fuse

I'm pretty sure that i've removed all the components that i just put back on.
Have included a photo, just in case you can spot any that i've missed, or should have removed
But the bulb is still glowing.

So i tested the AC and CD sides separately at the rectifier circuit board holes, and noticed that 1MΩ again that i previously couldn't find (so it wasn't due to my body resistance, although coincidentally that what it roughly measured while holding onto the test leads normally) it was hidden under the X capacitor. So the AC side looks normal, and i've put the capacitor and coil back on.
And there is no short on the DC side.
I thought that the 1MΩ might not be due to a resistor, because i didn't see it on a schematic that i was looking at (attached below) - but it obviously does (and shouldn't there be one of these in the schematic?)
What's the function of that resistor, as i'm picking that it's not a discharge resistor for the primary capacitors, as it looks like R1 and R2 in the schematic would do that job.

And as any short arising (which neither of my meters is showing) across the DC side, looking at the the schematic, could something be going on with those two transistors that is causing a short when the power is switched on? (assuming that my PSU is similar enough to that - hey, it's got two BJT transistors on the primary, and has an LM339, which looks like it's for the 5VSB)

Did you try removing/unwinding/cutting the the isolating tape on the transformer? A lot of traffos hide the fuse in there. Usually it's a square box fuse, kind of looking like a small X2-class safety cap.

Unfortunately, no fuse in this one - thanks for the tip though.




[URL="https://www.badcaps.net/forum

[momaka]

Re: Task TK-940TX-DF Blowing fuse

i put all of the parts back (and replaced the shorted zenner ZD1, that you can see just above the rectifier in post #27) except for that big pink resistor that i lost, and the other two large FETs - turned it on, and the light bulb instantly glows, and no 5VSB.
I'm starting to think that the original rectifier might not be the cause of the problem

Well either way, back to removing parts one by one then, as detailed in the previous pages. Cumbersome? -Yes. But you should be able to find the part responsible when following all of those steps again.

That "1MΩ discharge resistor" that i previously mentioned, doesn't exist, because when searching for it on the board, and not finding it, and after looking at a few PSU schematics, i eventually realized that it was the resistance of my body as i was holding the probes with my fingers onto the power supply wires.
Doh!

It happens.

Another good reason not to hold the probes by the metal parts is if you encounter a charged HV capacitor - could get a nice sting for sure.

How do you open them nicely? As the advice that i've read is to cut it open, or start whacking it with a hammer after freezing it.

Depends on how well the adapter is glued. Usually I try the hammer and/or clamp method first (hammer or clamp in the corners to break the case seam.) But some adapters are too stubborn for that, so I just hacksaw or dremmel them open instead. They don't look nice afterwards with the latter option, though.

Got it open by freezing it and tapping it out with a hammer and a blunt wide chisel.
Moderately heating it might have worked better because some plastic broke off inside the groove. But anyway - yep, no fuse and open circuit on the primary;

Did you try removing/unwinding/cutting the the isolating tape on the transformer? A lot of traffos hide the fuse in there. Usually it's a square box fuse, kind of looking like a small X2-class safety cap.

And i just learned why an effective short in the primary of a transformer doesn't blow the switchboard fuse - because of the winding's inductance.

Actually, shorting the secondary side pretty much "shorts" the primary side inductance too. So then the only thing left is the resistance of the windins... which is too low and allows too much current to pass through, thus burning the windings. However, the current is not high enough to blow the house breaker.

BTW, Firefox RAM usage was, a few minutes ago, up to nearly 1.5G, even after closing the three youtube/bitchute windows.
Then things started to go slow, and then the masked browser screen appears a few times with the blue ring throbber.
Then Firefox crashed.

Yeah, I just restart FF after I'm done watching YT or between different "sessions".

[socketa]

Re: Task TK-940TX-DF Blowing fuse

When these small transformers are over-loaded, the primary typically tends to open.

Got it open by freezing it and tapping it out with a hammer and a blunt wide chisel.
Moderately heating it might have worked better because some plastic broke off inside the groove. But anyway - yep, no fuse and open circuit on the primary; but it was too heavy anyway, as it kept falling out of the wall socket - maybe heavy because it needed more power for the Wi-fi.
And i just learned why an effective short in the primary of a transformer doesn't blow the switchboard fuse - because of the winding's inductance.

BTW, Firefox RAM usage was, a few minutes ago, up to nearly 1.5G, even after closing the three youtube/bitchute windows.
Then things started to go slow, and then the masked browser screen appears a few times with the blue ring throbber.
Then Firefox crashed.

[socketa]

Re: Task TK-940TX-DF Blowing fuse

i put all of the parts back (and replaced the shorted zenner ZD1, that you can see just above the rectifier in post #27) except for that big pink resistor that i lost, and the other two large FETs - turned it on, and the light bulb instantly glows, and no 5VSB.
I'm starting to think that the original rectifier might not be the cause of the problem

When i test, in circuit, across that heat-shrinked resistor, that's next to the 5VSB transformer, i get a hard short - is that a clue?

That "1MΩ discharge resistor" that i previously mentioned, doesn't exist, because when searching for it on the board, and not finding it, and after looking at a few PSU schematics, i eventually realized that it was the resistance of my body as i was holding the probes with my fingers onto the power supply wires.
Doh!

i found the adapter for the modem (i thought i threw it away, but discovered that i didn't)
How do you open them nicely? As the advice that i've read is to cut it open, or start whacking it with a hammer after freezing it.

Ah yes, that's pretty much standard not just for the latest/last Firefox, but Chrome/Chromium and IE/Edge as well. On systems with less than 2 GB of RAM, you just have to close and restart the browser after a bit of use. Heck, even with 4 GB of RAM, I've ran out after several hour-long sessions on YouTube.

Thanks for the confirmation - i spent hours searching for a way to hard limit Firefox RAM usage, to no actual avail.

[momaka]

Re: Task TK-940TX-DF Blowing fuse

Looks like we may have got it!
Took out the PBU805 bridge rectifier and replaced it with a RS405L and the bulb didn't come on.

Yay, about time, lol.
I guess I should have suspected the BR earlier to save on removing all of those components. But it's such a weird failure, as diodes usually fail short-circuit or occasionally open-circuit if enough current is forced into them. Have not seen one, at least on my bench, fail like yours, where it only breaks down under load. But I do know it is a possible failure mode... just extremely rare.

It's a downgrade to 600 Peak Reverse Voltage, is that still perfectly acceptable?

Yup, 600V is fine.
Even 400V will work, though that would be cutting it very close (could certainly fail with a power surge, as the DC voltage at the caps is nominally ~340V.)

I've put the MOV's, heatshrinked resistor, small blue cap, and 5VSB transformer back in, and the bulb remains unlit.
Should i keep replacing the parts ( in any specific order?) and doing the bulb test, after each component?

You can place back all of the 5VSB parts (including test 5VSB transistor) and then do the bulb test. If the 5VSB still has the replaced/good caps on its output, we should hopefully see the 5VSB come up now.

I've also lost that big pink resistor that you can see in post #5 - searched high and low for it - maybe it self-dissipated back into the quantum vacuum. What value would you say that it is?, as i can recall that one of the coloured bands looks a different colour in the photo than 'reality'.

R14? That one is for the main PS.
Looks like 0.1 Ohms (Brown, Black, Silver, Gold). 5th band (green) I think signifies reliability in this case, or something similar. Not to be confused with regular 5-band color-coded resistors, though.

at the start of this process i did remove the bridge rectifier and it tested good (and still does) according to this video:
https://www.youtube.com/watch?v=K-oNaQvMQco
and i also tested to both AC Legs, which over all equates to 4+4+2=10 tests in total, that all tested good, according to the diagram in the video.

Good that you tested to both AC legs for steps 1 and 3. The video doesn't tell you to do this, but you always should anyways... so you definitely tested the bridge rectifier properly. It just threw a curve ball on us, breaking down only with voltage.

That's either an earth contact or a resistance fault, often accompanied with audible line noise, on the copper phone line, or suspicious.

Probably.
The building/apartment we lived in back then was in Eastern Europe, built in the 70's during the Soviet Block days. Earth/grounding was not required by wiring code then, AFAIK, so most buildings have no earth. (The best was when someone grounded their water heater to the hot water/radiator pipes for heating, expecting them to provide a ground connection in case of fault. I remember many-a-times when going into other people's apartments and touching the radiators, I'd get a slight "tickly" feeling. On a few, it actually stung quite a bit. Many years later after I studied electronics I realized what this was.)

Also, phones had just moved to tone dialing in the cities back then in the 90's, so POTS wasn't 100% reliable quite yet. Still better than the old Pulse-dialing system in the rest of the country(side), though - on that one, you couldn't even try to use dial up.

Also, with a 56k modem, but only a few times, it'd reach speeds of close to 20kb/s when i accessed a certain FTP site to download files, or doing bit torrent. So i suspected that dial-up could go a lot faster than what they were giving us.

Eh, that still sounds pretty horribly slow, lol.
We moved to broadband/cable in 2001 and never really looked back.

Price-wise, the difference between dial-up/ADSL and cable was too small to justify sticking to such slow connection. And in the case of Eastern Europe, where my cousins continued to live for a while... dial-up was always too expensive. In the mid-2000's, small "corner-store" ISPs started cropping up and throwing ugly Ethernet wiring between buildings, as the city didn't care to enforce any communication wiring regulations. The benefit of that was that for a fraction of the price of dial-up, people were getting a fast Ethernet connection - typically uncapped 100/100... or as fast as the tubes could get ya ... because it was literally a router or two on every apartment building, wired to a huge switch, and then an Ethernet connection ran to every customer. On the plus side, it was super-fast for torrenting. On the down side, during a thunderstorm, many people lost their computers due to those Ethernet cables running around and between buildings without protection. Add old buildings with no grounding and cheap Deer power supplies, and you can start to imagine how many PCs bit the dust.

I'm using FF 48, and the latest versions of Palemoon, MyPal, and Basilisk. After a while the RAM (probably when i have bitchute or screwyoutube window open, but not playing) goes to just over 2GB and the browser wont let it go (even after closing all of the windows) until i restart the browser. And it still does it on a different computer.

Ah yes, that's pretty much standard not just for the latest/last Firefox, but Chrome/Chromium and IE/Edge as well. On systems with less than 2 GB of RAM, you just have to close and restart the browser after a bit of use. Heck, even with 4 GB of RAM, I've ran out after several hour-long sessions on YouTube.

i threw the adapter into the rubbish bin, which was collected about 2hrs after i read your post - yes, i should have at least opened it up to see why.
Should have let it sit for a while longer, until that idea popped into my head.

Ah, sorry I didn't post that earlier then. It's been pretty hot here for the last few weeks, so I've been avoiding the computer during the day or limiting use.

This happens occasionally, and it was the total RAM usage, not just Firefox. I'm now monitoring it more carefully, to see when it happens.

It's Firefox, I'm pretty sure.
I always keep Task Manager and regularly keep an eye on my resources - especially on my slower systems, like my main PC.

[socketa]

Re: Task TK-940TX-DF Blowing fuse

Correction:

I'm using FF 48, and the latest versions of Palemoon, MyPal, and Basilisk. After a while the RAM (probably when i have bitchute or screwyoutube window open, but not playing) goes to just over 2GB and the browser wont let it go

This happens occasionally, and it was the total RAM usage, not just Firefox. I'm now monitoring it more carefully, to see when it happens.

[socketa]

Re: Task TK-940TX-DF Blowing fuse

Looks like we may have got it!
Took out the PBU805 bridge rectifier and replaced it with a RS405L and the bulb didn't come on. It's a downgrade to 600 Peak Reverse Voltage, is that still perfectly acceptable?
Back at the start of this process i did remove the bridge rectifier and it tested good (and still does) according to this video:
https://www.youtube.com/watch?v=K-oNaQvMQco
and i also tested to both AC Legs, which over all equates to 4+4+2=10 tests in total, that all tested good, according to the diagram in the video.
So the rectifier tests good but it is bad! (i double checked that the bulb lit, by tacking the original rectifier onto the reverse side of the board before attaching and soldering the replacement in)
Looks like it shorts out only when voltage is applied - thought that there was something odd going on when there was no contact between the power plug pins, but the light came on.
I've put the MOV's, heatshrinked resistor, small blue cap, and 5VSB transformer back in, and the bulb remains unlit.
Should i keep replacing the parts ( in any specific order?) and doing the bulb test, after each component?
I've also lost that big pink resistor that you can see in post #5 - searched high and low for it - maybe it self-dissipated back into the quantum vacuum. What value would you say that it is?, as i can recall that one of the coloured bands looks a different colour in the photo than 'reality'.

I don't know about dial-up. I get chills when I remember how bad ours was back in late '99 with its dial-up connection re-try noises before finally getting a connection... only to get lost after 5 minutes max

That's either an earth contact or a resistance fault, often accompanied with audible line noise, on the copper phone line, or suspicious.
My dial-up was great at around 3kb/s, and only occasionally was disconnected when there was a problem with the copper lines or the exchange equipment (and i rang to find out), or my computer. Also, with a 56k modem, but only a few times, it'd reach speeds of close to 20kb/s when i accessed a certain FTP site to download files, or doing bit torrent. So i suspected that dial-up could go a lot faster than what they were giving us.
There is others that verify this as well:
https://forums.whirlpool.net.au/archive/408223

Pretty much all versions after Firefox 32 have this "bug" where at startup and also at random intervals, they like to hog CPU resources for 10-30 seconds. Firefox 24 was the last good version that didn't do this and also wasn't too heavy on resources

I'm using FF 48, and the latest versions of Palemoon, MyPal, and Basilisk. After a while the RAM (probably when i have bitchute or screwyoutube window open, but not playing) goes to just over 2GB and the browser wont let it go (even after closing all of the windows) until i restart the browser. And it still does it on a different computer.

When these small transformers are over-loaded, the primary typically tends to open. The better-built ones will have a small fuse built into the transformer's primary (typically under a few layers of tape.) Change that and the adapter/transformer will work fine again.

Thanks for that info.
i threw the adapter into the rubbish bin, which was collected about 2hrs after i read your post - yes, i should have at least opened it up to see why.
Should have let it sit for a while longer, until that idea popped into my head.

[momaka]

Re: Task TK-940TX-DF Blowing fuse

...
4) Still glowing (why am i not surprised?) after removing the transformer and the resistor(1Ω)

Just brought some keyboard cleaner today - so that was good timing.
Cleaned it as much as i could, and couldn't spot any blobs or spashes that would have caused a short.
How the heck is the bulb glowing when all that i'm getting between phase and neutral (of the power plug socket), with fuse inserted, is 1MΩ (that "discharge resistor")?

Good question.

I double and triple-checked everything from the last photo you uploaded and there are no possible low-resistance/impedance paths left to ground. That is, again, assuming you have removed the MOVs behind the bridge rectifier.

This only leave the bridge rectifier and/or possibly something wrong with your incandescent bulb setup.

Do this first: take a known good ATX PSU and connect it to your series incandescent bulb. Make sure there is nothing connected to the output of the PSU (i.e. no motherboard or anything.) Now plug it and see what happens. Does the bulb still glow?

If not, then remove the bridge rectifier on the Task PSU and replace with another bridge rectifier, if you have a spare one. Otherwise, you can use 4 regular diodes too in place of the bridge rectifier (but they need to be rated for at least 400V DC reverse blocking voltage, or more.)

If i just connect the bulb in series with a 1MΩ resistor, and nothing else, would it glow like that?

Well, try it and see.

Hint: before trying, think about the resistance of the bulb relative to the 1 MOhm resistor and the total current that will flow through this circuit. That might give you a clue of what results to expect.
Hint 2: the results will be rather uneventful.

So it seems like the way to destroy a power adaptor is to make sure that its output is AC then plug it into a DC device.

Yup.
If it's an AC adapter, it's probably just a simple line-connected transformer with no diodes or capacitors on the output, hence outputting at 50/60 Hz AC (depending on country line frequency.)

Your router probably has a reverse-voltage protection diode inside, which likely shorted the AC adapter on every reverse-polarity cycle it output.

When these small transformers are over-loaded, the primary typically tends to open. The better-built ones will have a small fuse built into the transformer's primary (typically under a few layers of tape.) Change that and the adapter/transformer will work fine again.

So i had to sort out another computer, which i was going to do anyway because either firefox (and it's few forks) or websites these days demand 2 Cores and more than 2GB of RAM (good grief, all that i want to do is read posts and watch internet videos at 240p or 360p) or they will cause the computer to lock up.

Windows XP 32-bit?

2 GB of RAM should be enough (though 1 GB isn't), but pretty much all versions after Firefox 32 have this "bug" where at startup and also at random intervals, they like to hog CPU resources for 10-30 seconds. Firefox 24 was the last good version that didn't do this and also wasn't too heavy on resources. Unfortunately, it's too old and won't render correctly on many websites. You really need version 49 or newer to make sure you don't have problems with websites.

That said, I'm still using a s478 Pentium 4 Prescott HT as the CPU in my main PC, where I do watch YT @ 360p once in a while (using FF 52 ESR Portable.) It's a pain, though (takes a few seconds for YT page to load fully, and then a few more to start the video buffering.) Adblock + custom hosts file helps quite a bit - at least with cutting out ads from loading and hogging more resources. But all in all, I'm getting very near the point of retiring that PC.

So i've got something like 30 462/754 boards that are now practically non-internetable.

The socket 462 boards... yeah, unless they have Athlon XP 2500+ or better with at least 1.5 GB of RAM, they won't be much good on the internet, mostly due to websites being too bloated and not optimized to run on the (few) older instruction sets that these CPUs support (SSE at best on Athlon XP.)

The socket 754 boards should be a bit better if equipped with Athlon 64 2800+ CPUs or higher. They have SSE2 and clock-for-clock, Athlon 64 should be faster than Athlon XP. Don't expect miracles, of course.

Last time I used my OC'ed s939 Athlon 64 3200+ (@ 2.5 GHz) GPU test system on the web (around a year ago), it still managed to (barely) handle YT @ 720p with the help of an nVidia GeForce 8600 GT (for H.264 acceleration.) It ran OK elsewhere on the web, though. So at a minimum, you don't necessarily need 2 CPUs cores, but a little newer CPU will help. For AMD, I'd say socket 939 @ 2.4 GHz (1 core) or 2 GHz (2 cores) would be a lot better. For Intel, Pentium 4 Prescott HT on socket 775 would be minimum. Even though P4 on s775 is not much better than the 32-bit s478 P4s, the change from AGP to PCI-E GPU could allow you to render some websites faster.

Also, if you need a PC just for surfing more basic (and sane) websites, like Badcaps and Wikipedia, try the last Opera 12 release (version 12.18 on Presto engine, -not- Chromium garbage). I'm still using that on my bench "helper" laptop and a few other retro XP systems. In the case of my bench laptop, it has a Pentium 3 @ 700 MHz with 512 MB of RAM. I use it now only for viewing/downloading PDFs, surfing BCN occasionally, and once in a while looking info up in Google/Wikipedia.

i was happy with dial-up, buffering, and waiting for the video to load.
Take me back 20years and push hold.

LOL
I don't know about dial-up. I get chills when I remember how bad ours was back in late '99 with its dial-up connection re-try noises before finally getting a connection... only to get lost after 5 minutes max.
But I'll be pretty happy if time turned back 10-13 years. Back then, websites still weren't anywhere near as bloated as they are now. And 15 years ago, I think the "bottom-less scrollbar" was still not a thing. I despise websites with bottomless scrolling. We have smartphone users to thank for that "feature".

[socketa]

Re: Task TK-940TX-DF Blowing fuse

Took an extra day to do this because my computer wouldn't talk to the router, so i found a spare one and couldn't be bothered checking the volts and amps, and used the same power adapter and killed the adapter because it's output was AC and i plugged it into a router that needed DC (Some derivitive of Murphys law, since this was the first time that i've ever noticed AC output on a power adapter) - So it seems like the way to destroy a power adaptor is to make sure that its output is AC then plug it into a DC device. So AC and DC are not compatible unless you are Bonn Scott (or Brian before his voice broke down).
After sorting out another router, the same problem occured which pointed to the computer being the source of the problem.
So i had to sort out another computer, which i was going to do anyway because either firefox (and it's few forks) or websites these days demand 2 Cores and more than 2GB of RAM (good grief, all that i want to do is read posts and watch internet videos at 240p or 360p) or they will cause the computer to lock up.
So i've got something like 30 462/754 boards that are now practically non-internetable.
i was happy with dial-up, buffering, and waiting for the video to load.
Take me back 20years and push hold.


1) 210pF
2) Still Glowing
3) Removed but didn't test because the bulb still glows after its removal
4) Still glowing (why am i not surprised?) after removing the transformer and the resistor(1Ω)

Just brought some keyboard cleaner today - so that was good timing.
Cleaned it as much as i could, and couldn't spot any blobs or spashes that would have caused a short.
How the heck is the bulb glowing when all that i'm getting between phase and neutral (of the power plug socket), with fuse inserted, is 1MΩ (that "discharge resistor")?
If i just connect the bulb in series with a 1MΩ resistor, and nothing else, would it glow like that?

[momaka]

Re: Task TK-940TX-DF Blowing fuse

Perhaps there is a "free radical" inside of the circuitry?

Maybe.

I think we are just getting closer, though. Hope you still haven't given up yet.

So the last test with the new primary caps confirmed the old ones are probably fine.

I analyzed the new image you posted (of the solder side) and determined there is only one last possible path: shorted snubber ceramic cap on the 5VSB primary side. Actually, there are two snubbers on this 5VSB, it appears. One is the standard series connection of a diode with a cap and resistor in parallel. The other is just a small (blue) ceramic cap connected between Source and Drain of the (removed) 5VSB MOSFET. It's possible that this cap is shorted, so remove it and test it. Let me know what results you get. Also try the incandescent bulb test again after you remove this component, regardless if it tests OK or not. If the bulb still glows, remove the 5VSB transformer and try the bulb test again. Still glowing or not? If still glowing, "free radicals" and small gremlins inside the circuit are confirmed. I doubt it, though. If you had to remove the 5VSB transformer for the bulb to stop glowing, test resistance on the primary side of the 5VSB traffo. Using your last image above of the solder side as a reference, and starting from the top two pins on the 5VSB transformer, these two should show low resistance to each other. Likewise, the bottom two pins should also show low resistance to each other. However, the top two and bottom two pins should NOT show any resistance to each other (or near infinite, if you're touching the metal parts on the pins or multimeter probes with your fingers, which can happen to show a high resistance instead of open circuit ).

So just to recap the steps:

1) Remove blue snubber cap connected between Source and Drain pins on 5VSB MOSFET and test it out of circuit. Post results for this step.
2) Try incandescent bulb test. If it is NOT glowing, STOP. If it is still glowing continue to step 3.
3) Remove 5VSB transformer. Check resistance between pins and post results here.
4) Perform bulb test again. If glowing, STOP and clean all PCB tracks thoroughly with IPA and make sure there are no stray solder blobs shorting things. In addition to that, remove heatshirnk-covered resistor next to 5VSB MOSFET and do the incandescent bulb test again (though I doubt this will change anything.)
.
.
.
5) You've been defeated by the circuit gremlins
.
.
Or at least for now. Let's see what results you get, as that may give us a different clue.

[socketa]

Re: Task TK-940TX-DF Blowing fuse

Try measuring the resistance of both caps with a multimeter instead of ESR meter.

After discharging them and putting the multimeter across them, they both quickly increase in resistance and go up to (according to my meter) just over 40MΩ, then back down slowly to 35MΩ and 29MΩ respectively.

If neither shows shorted, check the resistance across each cap spot on the PSU PCB. See if anything there measures low resistance (please post what resistance you get.)

There's no low resistance (there's actually ∞ resistance). I also made sure that test path wasn't blocked by any flux residue, by checking continuity to another point on the same trace.

If resistance across either spot is several KOhms or more, try putting different caps in the PSU

Got some 680uf 200V caps from another PSU and soldered them in.
But the light bulb still fecken (long live "Father Ted") glows.
Perhaps there is a "free radical" inside of the circuitry?

[momaka]

Re: Task TK-940TX-DF Blowing fuse

Went through all of your 'flow chart' four steps, and the bulb only stopped glowing (thank goodness) when i removed the primary caps.
So it appears that this is the "unlikely" cause of the problem?
Both of the capacitors test about 740uf and have and ESR of 0.05

Try measuring the resistance of both caps with a multimeter instead of ESR meter.

If neither shows shorted, check the resistance across each cap spot on the PSU PCB. See if anything there measures low resistance (please post what resistance you get.)

If resistance across either spot is several KOhms or more, try putting different caps in the PSU (and mind the correct polarity, because otherwise you may get some magic smoke even with the series light bulb ). The caps don't need to have the same capacitance, just the same or higher voltage rating. Anything in the range of 220 to 1000 uF will do fine, though the caps do need to be a matching pair (as in, both the same capacitance and voltage.) After this, plug in the PSU and see if the bulb glows again.

If it does, there is still something shorting across one or both caps and the problem is something on the PSU. Thus, the original caps may be fine, if that is the case. As requested in my last post, again, it would be great if you can provide another PCB shot of the solder side, preferably more focused and looking at a straight angle at the PCB (so I can see if we missed a possible path somewhere that can be causing a short-circuit.)

Thanks for the safety info - when i was at school, years ago, we were charging up capacitors, and they guy next to me discharged one on his face or mouth (i cant remember which) and he said that he saw a flash of light.

Ya, big HV caps can store quite a bit of energy. If charged and shorted directly with a metal object, they will definitely produce a large spark and loud pop.(anything over 30V DC and 1000 uF starts to get loud very quickly.) The problem with that is not so much that the spark can hurt you, but rather the loud noise catching you off guard - with the danger being pulling your hand/head/body away and hitting/scraping/cutting yourself in the process. You can actually see that on Electroboom's intro video. Of course, that is scripted, as are all of his other videos, but it's actually a pretty realistic outcome.

That's why I suggested discharging the caps though the plug on a soldering iron or something similar that can take the high DC voltage but not draw much current, and hence discharge the caps very "gracefully".

[socketa]

Re: Task TK-940TX-DF Blowing fuse

Finally got around to doing this (Johnny-ten-jobs at the moment! - trying to get as much stuff done as possible, just in case TPTB have more tricks up their sleeves):
Went through all of your 'flow chart' four steps, and the bulb only stopped glowing (thank goodness) when i removed the primary caps.
So it appears that this is the "unlikely" cause of the problem?
Both of the capacitors test about 740uf and have and ESR of 0.05
With the multi-meter, +ve lead to +ve terminal, -ve lead to -ve terminal, it rises to 37Megohms and then slowly starts to fall, and for the other cap it rises to 40Megohms and then slowly starts to fall.
So primary capacitors can test OK and still arc internally?

Thanks for the safety info - when i was at school, years ago, we were charging up capacitors, and they guy next to me discharged one on his face or mouth (i cant remember which) and he said that he saw a flash of light.

[momaka]

Re: Task TK-940TX-DF Blowing fuse

Attached a photo with the components removed, just to confirm that i removed the right ones.

Yup, you got them.

Additionally, there is 1 Meg ohm resistance measured across phase and neutral - is that OK?

Yes, that is normal.
Usually, there is a high-value resistor between Live and Neutral to discharge the X2 caps. 1 MOhm is a very typical value.

And would that cause the light bulb to glow brightly?

Well, there aren't that many paths left, but possibly there is still something shorting on the DC/rectified side of the primary side.

Discharge the primary caps (putting your soldering iron's wall plug across each cap's leads for a few seconds will do the trick), then measure resistance between (+) and (-) leads on the bridge rectifier. Keep the multimeter connected for a few seconds, as the capacity of the input caps may show a low resistance at first. With this test done, if the resistance raises and is not low (over 200 Ohms), remove the two orange MOVs. I know you mentioned you measured them and they aren't shorted, but perhaps they could be conducting at live voltage. Now try plugging in the power supply and see if the bulb still glows. If yes, there are only two more possible paths from faulty components that can cause this.

The first one is the blue ceramic cap between the 5VSB transformer and 5VSB MOSFET (i.e. next to that big ceramic resistor.) Remove that and try the bulb test again.

Bulb still glowing? If yes, the last possible path is... and this is very unlikely... that one or both of the primary caps are bad and arcing internally when charged. Remove both caps and try the bulb test again. If the light bulb still glows, take a good picture of the bottom side and post it here. I know there is one on page 1, and I've been using it extensively to trace the circuit on this PSU. But just to make sure something isn't shorting there or something hasn't changed in an unexpected way since that last image.

So again, just a run down of the procedure:
1) discharge primary caps and test resistance between + and - of bridge rectifier. Post value here.
2) remove MOVs and perform incandescent bulb test again. If NOT glowing, STOP. The MOVs are likely bad. Otherwise continue below.
3) Remove blue ceramic cap next to white ceramic resistor near 5VSB MOSFET area. Perform incandescent bulb test after this. If NOT glowing, STOP - ceramic cap is likely bad and failing when charged. Otherwise, continue below.
4) Remove both primary caps and perform incandescent bulb test. Post here if bulb still glows or not.

Also, between each step above, it probably IS good idea to try and discharge the primary caps as detailed earlier with a soldering iron plug or something similar. That way, if you get to one of the test steps above where the bulb is no longer glowing because the short on the primary is gone, then the PSU will likely end up with its primary caps charged at high voltage (about 170V DC each) - at least for a few moments/minutes until the discharge/balance resistors across the caps discharge them. Thus, for extra safety, I like to make sure those primary caps are discharged.

[socketa]

Re: Task TK-940TX-DF Blowing fuse

No power selector on this PSU - so that rules that out.
Attached a photo with the components removed, just to confirm that i removed the right ones.
Additionally, there is 1 Meg ohm resistance measured across phase and neutral - is that OK? And would that cause the light bulb to glow brightly?