Antec SL350P PSU

[c_hegge]

Re: Antec SL350P PSU

PS: I had no helmet in the end, but I did wear a pair of goggles. Incidentally, they are laser protection glasses, which I thought would be handy to at least protect me from UV light if sparks should occur...

I never bother wearing protective gear, although when I re-built the Rexpower PL-400, I did use it on the end of a 10m extension cord and stood way back.

[Toasty]

Re: Antec SL350P PSU

LOL!

Happiness!

>>It seems that the light bulb in series was at the root of problems after all.<<

I use a 100w and 200w bulb. If it won't fire with the 100w, I switch to the 200w. It usually does the trick and pulses for a few seconds until the caps charge fully. On the other hand, I've had them not start with the 200w also. Then it's a "fire test" with just a circuit breaker (old TV style) and a jumper across the fuse.

Have you tried running the system with it?

Toast

[amartolos]

Re: Antec SL350P PSU

I wear eye protection since the inrush limiter disintegrated upon power-up. I guess I am too lazy to extend the mains cord farther. Actually, when I'm soldering I wear it also, in case I am too clumsy and liquid solder starts flying around. My safety glasses are reasonable to wear, they are made of transparent polycarbonate.

>>Have you tried running the system with it?<<

I only had time for a short test, before calling it a day. The clean-up and removal of the too long lead ends took quite a while (there were so many). I want to install also the DM311 circuit, and it's such a mess to solder the wires coming from the mains receptacle (too short) to enclose the PSU, that I would do it only once. One of those on-board connectors for mains power would have been handy, but I can't get any.

[amartolos]

Re: Antec SL350P PSU

Finally... the computer runs again, after nearly two months in service. It is now in its prime...95, running at full load since almost an hour. The voltages are 3.30 V, 5.00 V, 12.27 V and 5.12 Vsb, measured with the DMM (different from the m/b reported voltages). Pictures with the PSU with mounted heatsinks are here and here. Sorry again for the bad color balance. The list of cap replacement, for reference, looks like this:

	Type		Value		Size	Pos.	Replacement
___________________________________________________________________
2x	LP NEG		330uF 200V	22x25	1,2	Panasonic TS-HC 105°C 250V 24x25
	TNR		10uF 50V	5x11	5	Nichicon PW
	TNR		22uF 50V	5x11	14	Nichicon PW
	TNR		47uF 25V	5x12	8	Nichicon PW 50V 6.3x11
	NP (85°C)	1uF 50V		4x7	34	Nichicon VP 5x11 (85°C)
2x	SS		1uF 50V		4x7	33, PFC	Nichicon PW 5x11
	SS		0.22uF 50V 	4x7	PFC	Panasonic KG
	TMR		4700uF 10V	10x35	31	Panasonic FM 12.5x35
2x	TNR		3300uF 10V	10x25	36,37	Nichicon PW 12.5x30
2x	TMR,TNR		2200uF 16V	10x30	30,35	Panasonic FK
	TNR		820uF 16V	10x20	25	UCC LXY 10x25
	TNR		100uF 10V	5x11	26	Nichicon PW
	TEAPO A3	3300uF 6.3V	10x25	32	Rubycon MBZ
	CapXon KM	330uF 16V	7x16	42	Panasonic FM 8x11.5
2x	CapXon GL	1000uF 10V	8x20	23,28	Panasonc FM

For the positions see this picture from one of the first posts. The replacements are written with what is different from the originals. The only well-done capacitor was the odd 4700/10 one. Some size problems were posed by the 12.5 diameter caps instead of the 10 mm originals in the secondary filtering. The 4700/10 had anyway a bigger silkscreen, so it fitted quite snugly, but the other ones around it had to be squeezed. It is the reason why I chose Panasonic FK for 2200/16, and it was manageable, without needing to drill additional holes.

I successfully applied the DM311 mod, thanks to the comprehensive instructions from everell in post #5 of this thread (dealing with a similar power supply, but without APFC.) Pictures of the DM311 board are here and here, not the best soldering job but it works. This picture shows how it sits in the PSU. I had to abuse the secondary heatsink, but it's quite stable. My only concern is how it changes the airflow.

Now for some comments: after recapping, the fuse was blown, and the MOSFET of the APFC failed by shorting the gate to drain, while drain-source remained open (not even the internal parallel diode could be detected with the DMM). This short caused probably also the explosion of the current inrush limiter at the second attempt, after replacing the fuse. Looking now at the schematic of the APFC section, and knowing that both attempts were with the PSU in standby, not shorting the green wire to ground, the possible causes for the FET failure could have been either a fault in the soldering of the caps responsible for the standby circuit (not unlikely, as I was impatient then), or just failure due to aging. As seen in the posts that goodpsusearch linked to earlier in this thread, this seems to be a weak point for Antec supplies with APFC. My PSU has quite a long history: I acquired it 8 years ago (I was so proud of it...), and it has seen everyday use in the evenings and longer in the week-ends, with some occasional gaming. During this time it survived the demise of a K7S5A board (guess why... bad caps), followed later by the death of the remaining CPU, on the new board (the current one, a K7VT4A Pro). I can only hope that the new MOSFET, with a lower RD-S(on) resistance, will fare better. Which makes me wonder: should I worry about the other silicons and their age?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


I would have some specific questions:

1. A 8A fuse is required by this supply. Do I really need such a high value? I would say that I could go with only 4A, the 8A being probably required if I want to plug it at 110V, which I will never do. The last tests were done with a 2A fuse and this didn't raise any problem, neither for the initial loading of the mains caps, nor for running with a 44W load. The inrush limiter consists now of two NTC's in parallel rated at 12 Ω, 5 A each (this is what I found, compared to the original at 5 Ω, 8 A) and they seem to handle well the initial surge.

2. There is a power diode on the APFC circuit, D103 in the schematics, on the same heatsink as the switching FET. It sits oposite to the boost coil, and originally had a plastic sheet glued to it, probably to prevent arcing from/to the coil. I forgot to glue it back, so now I wonder, is it really needed? If yes, can I use a spot of crazy glue / super glue to fix it, or it is too corrosive for the diode casing, in which case what kind of glue to use?

3. Here maybe goodpsusearch, if he still reads this thread, can answer: you mentioned a fan mod to increase the fan speed, by placing a resistor in parallel to the RTH2 thermistor, as in this post. In that case you also had a thermistor at RTH3, visible in pictures next to the secondary heatsink. I don't have a RTH3, but I do have a RTH2 directly beneath the secondary heatsink, visible in the picture where the heatsink is removed, close to one of the Schottky rectifiers. Is this the one that has to be hacked? Of course, there is also the option to directly connect the fans to 12 V, but I would still have some temperature control.

There, I hope I am not being too picky.

Thanks!

[goodpsusearch]

Re: Antec SL350P PSU

2. There is a power diode on the APFC circuit, D103 in the schematics, on the same heatsink as the switching FET. It sits oposite to the boost coil, and originally had a plastic sheet glued to it, probably to prevent arcing from/to the coil. I forgot to glue it back, so now I wonder, is it really needed? If yes, can I use a spot of crazy glue / super glue to fix it, or it is too corrosive for the diode casing, in which case what kind of glue to use?


Thanks!

Better safe than sorry. These smart power psus were not designed to have APFC! Antec was forced to add the add on APFC board because of EU limitations. That APFC board is notorious to fail. There are incidents that the coil went so hot, it burnt itself. I have seen photos of it being completely black... So I would not risk it.

3. Here maybe goodpsusearch, if he still reads this thread, can answer: you mentioned a fan mod to increase the fan speed, by placing a resistor in parallel to the RTH2 thermistor, as in this post. In that case you also had a thermistor at RTH3, visible in pictures next to the secondary heatsink. I don't have a RTH3, but I do have a RTH2 directly beneath the secondary heatsink, visible in the picture where the heatsink is removed, close to one of the Schottky rectifiers. Is this the one that has to be hacked? Of course, there is also the option to directly connect the fans to 12 V, but I would still have some temperature control.

Possibly...You have to experiment with it. I realized that the other thermistor in my antec (RTH3) is for overtemp protection, because adding resistors in parallel did nothing to fan speed and suddenly I went too low on the value of the resistor, then the psu simply denied to start!

But, you don't want to connect the fans directly to 12V. The noise is terrible and the psu soon will get dusty...

I believe that the smart power psus really need to be fan-modded, because they get too hot during operation and that can cause problems like this:
https://www.badcaps.net/forum/attach...3&d=1257994643

https://www.badcaps.net/forum/showpo...80&postcount=3

[kaboom]

Re: Antec SL350P PSU

Better safe than sorry. These smart power psus were not designed to have APFC! Antec was forced to add the add on APFC board because of EU limitations.

Bingo.

What's more, it was treated as an afterthought. Let's all have faith in Antec to spec something right to their OEM. Ha- dream on!

[c_hegge]

Re: Antec SL350P PSU

I always add a 2.4 or 2.5K resistor in parallel with the termistor for Antec StupidPowers (although I've only done it to SPxxx models. I'm assuming that it would be similar for an SLxxx). I'd say RTH2 is the one to add the resistor to. Your unit may not have OTP if there's only 1 thermistor.

[everell]

Re: Antec SL350P PSU

That DM311 mod looks great! I especially like that classy heat sink. Glad it worked for you and is stable. No more worries about the 5vsb going overvoltage.

[amartolos]

Re: Antec SL350P PSU

Bingo.

What's more, it was treated as an afterthought. Let's all have faith in Antec to spec something right to their OEM. Ha- dream on!

Ok, you all convinced me. I took out the diode (STTH8R06D) to glue the foil on it, but in the process I managed to rip one leg off it. The only suitable replacement I found at the local shops was a rectifier containing two diodes with common cathode, but in the same package (TO-220), so I managed to install it by adjusting the positions of its terminals to the holes in the PCB, and I thus have one spare diode

Concerning the fan mod, a 2.4k resistor was too much, causing the fans to spin really fast and making a jet-engine-like noise. RTH2 is indeed the thermistor controlling the speed of the fans, and (in circuit) it measures about 4k, dropping to about 3.6k when I use a blow-drier on it. So I took some wires outside of the case and connected a 100k potentiometer to them, adjusting it for a higher speed of the fans, but still in the ear-pleasant range. I did not measure, but the current setting should be something like ~30k in parallel with RTH2, resulting in a fan speed of ~1650 RPM (measured by the mainboard) right after startup (initially it was ~1500 RPM), with the difference increasing as the temperature gets higher.

I plan to play further with the fan settings, and also try to replace them with some ball-bearing (both are sleeve bearing), and maybe put a 120 mm fan at the intake (this involves some sheet-metal work )

[Th3_uN1Qu3]

Re: Antec SL350P PSU

Around where that thermistor is connected there will be the transistor controlling the fan. From its collector to ground there's a zener diode which establishes the minimum speed of the fan. If you remove that zener, the fan will start at "off" and spin very slowly when the PSU is cold. It will of course still reach maximum speed when things get crispy.

[Toasty]

Re: Antec SL350P PSU

The 450 Antec's have the thermistor mounted to the heatsink with a clamp and screw. It's positioned just behind the large output filter coil between D11 & D12. You can easily extend the leads with some wire and place it there.

Remove the damn plastic shielding too. It's a factory quickie fix so the coils don't strike the heatsink in transit.
It blocks airflow across the heatsink very nicely.

Secure the coils to the heatsink and each other with "just enough" silicone caulk. Don't overdo it.

[lti]

Re: Antec SL350P PSU

I don't see how that plastic shield will block airflow. It looks like it is in parallel with the airflow path.

[amartolos]

Re: Antec SL350P PSU

I also had the feeling that the plastic sheet stays in the way. If I think of it, it covers the fins of the heatsink, rendering them ineffective and creating in between small (thermally insulating) air pockets.

[Toasty]

Re: Antec SL350P PSU

Correct. The airflow is downward from the top fan, and then out thru the side fan. The fins are completely covered by the plastic for better than 1/2 if the heatsink.

Now, if the fins were on the other side of the heatsink, the plastic would be fine.

I put a 1/4" (6-8mm) spacer in between the fins and the coils then caulked and let the caulk cure. Removed the spacers and there's a nice air gap.

Toast

[goodpsusearch]

Re: Antec SL350P PSU

After reading again this thread that user Amartolos opened...

and this: https://www.badcaps.net/forum/showth...ighlight=K2996
I think K2996 mosfet going short is a common failure of those Antecs.

Searching online brought some similar cases:
http://translate.google.com/translat...-surselor.html
http://translate.google.com/translat...e8%26t%3D46900

But I still can't decide if it's better to remove completely the add on APFC board or just replace the shorted K2996 mosfet.

I need your input here..