ATX PSU, 13.8v

[Pinhead]

Re: ATX PSU, 13.8v

The input rectifier is a bridge, though it's not heatsinked and I couldn't read the model/brand. I plan on replacing it with one that can be mounted to a heatsink, though I seriously doubt the input rectifier is currently the weak link.

It's difficult to say how warm the components run under load, since removing the cover also stops the airflow over the heatsinks and transformers. I know I haven't "felt" any real heat blowing out of the power supply even during a long TX session. Especially when comparing to the ambient temperature seen inside the typical PC, it runs appreciably cooler.

The PSU will typically see a duty cycle much less than 50% so I doubt "heat soak" will be a problem.

[Pinhead]

Re: ATX PSU, 13.8v

I did a little more work last night and seem to have eliminated the shutdown issues when transmitting on SSB. However, I'm concerned if the PSU's short circuit protection is still active.

Pin 6 is the "12v monitor" on the ATX-2005Z PWM controller. I tied this pin directly to ground potential which made the PSU supremely more stable; when hammering it on SSB the voltage only dropped from 14v to 13.9v and didn't shut down after 20 minutes of transmitting on max power talking to an operator in Australia. The duty cycle was better than 50% and I couldn't feel any heat coming out of the back of the PSU (80mm fan is running directly on 14v).

However, I was able to get the PSU to shut down by transmitting in AM mode at max power, so I assume the PSU still has some type of Over Power Protection enabled (AM uses more average power than SSB).

Here are some pics of the guts of this particular PSU.

[Pinhead]

Re: ATX PSU, 13.8v

I replaced the primary filter caps with 680uF/105°C capacitors, though they're admittedly not a brand that I've ever heard of (K-Eagle).

The rectifier is a PBL45 bridge rated at 4a @ ~70°C, 2a @ 100°C. With a 1.1v rated voltage drop, this bridge will have to dissipate approximately 4.4 watts at full tilt, though I believe the components further down in the chain will prevent this bridge from coming close to its max values.

Here is a datasheet for the D13007K transistors on the primary side. I'm not sure what values to refer to when finding their max current in a SMPS so I don't know whether they need to be replaced/upgraded. I haven't been able to locate any information on the "CET 6588Z CEF062N6A" primary transistor.

Secondary transistors:
SBL2045CT: Power Schottky Rectifier -20 Amp 45 Volt
SB1660CT: Isolation Schottky Barrier Rectifier - 16 Amp 60 Volt
SBL3045PT: Schottky Barrier Rectifier - 30 Amp 45 Volt

The ERL-35-2005 transformer in this supply is much bigger than the CD-EC-35B transformer in a Logisys PSU that I'm disassembling which is "rated" at 30, 38, and 25 amps on the 3.3, 5, and 12v rails, respectively. I seriously doubt that PSU's credentials, however, as they claim it can draw 10a from the mains at 110v with individual diodes as the main rectifier.

I found a THREAD claiming a real "35" transformer can supply 250-350 watts depending on various factors. If this is true, then it would seem that the main limiting factor in this PSU is the secondary Schottkys.

[Pentium4]

Re: ATX PSU, 13.8v

I replaced the primary filter caps with 680uF/105°C capacitors, though they're admittedly not a brand that I've ever heard of (K-Eagle).

Ahhh, K-Eagle Comic sans font and a "u" instead of a "µ" ? Those will work though and I doubt they will fail. Whether they are actually 105C caps is another story.

The rectifier is a PBL45 bridge rated at 4a @ ~70°C, 2a @ 100°C. With a 1.1v rated voltage drop, this bridge will have to dissipate approximately 4.4 watts at full tilt, though I believe the components further down in the chain will prevent this bridge from coming close to its max values.

You're right, I don't think the 4A bridge would be a factor but I always like to use a 6A minimum. (On 120V....)

Here is a datasheet for the D13007K transistors on the primary side. I'm not sure what values to refer to when finding their max current in a SMPS so I don't know whether they need to be replaced/upgraded. I haven't been able to locate any information on the "CET 6588Z CEF062N6A" primary transistor.

8A is pretty low for BJT's, especially with the low efficiency you will have. I would get some 13009's in there, they're usually rated at 12A which is a significant improvement.

Secondary transistors:
SBL2045CT: Power Schottky Rectifier -20 Amp 45 Volt
SB1660CT: Isolation Schottky Barrier Rectifier - 16 Amp 60 Volt
SBL3045PT: Schottky Barrier Rectifier - 30 Amp 45 Volt

Which ones are you using on which rails?

The ERL-35-2005 transformer in this supply is much bigger than the CD-EC-35B transformer in a Logisys PSU that I'm disassembling which is "rated" at 30, 38, and 25 amps on the 3.3, 5, and 12v rails, respectively. I seriously doubt that PSU's credentials, however, as they claim it can draw 10a from the mains at 110v with individual diodes as the main rectifier.

It's most likely good for at least 350W, but can we see a picture of it? Super Flower in one of their half bridge designs got over 500W out of a 35 transformer. It's pretty well designed though for half bridge.

Some of those YC caps may have high ESR and you might want to consider replacing them.

Sorry if I missed it above but what fan are you using? Usually the old Deer PSU's have really low quality fans with dry bearings.

[Pinhead]

Re: ATX PSU, 13.8v

Ahhh, K-Eagle Comic sans font and a "u" instead of a "µ" ? Those will work though and I doubt they will fail. Whether they are actually 105C caps is another story.

Yep, them's the ones. I just hope they're an upgrade from the 330uF 85°C Koshin caps that were there.

You're right, I don't think the 4A bridge would be a factor but I always like to use a 6A minimum. (On 120V....)

I may eventually replace it just for peace of mind, though.

8A is pretty low for BJT's, especially with the low efficiency you will have. I would get some 13009's in there, they're usually rated at 12A which is a significant improvement.

I'll have to look around and see if I can find a replacement.

Which ones are you using on which rails?

The secondary transistors are still in the factory configuration so I expect the SBL3045PT is 5v, the SBL2045CT is 3.3v, and the SB1660CT is 12v (this corresponds to the label's ratings). I haven't done any tracing/measurements yet, however, to verify which transistor is controlling which rail. I assume I could easily put the SBL3045PT on the 12v rail, correct?

It's most likely good for at least 350W, but can we see a picture of it? Super Flower in one of their half bridge designs got over 500W out of a 35 transformer. It's pretty well designed though for half bridge.

I'll get a pic. It seems pretty beefy.

Some of those YC caps may have high ESR and you might want to consider replacing them.

I've only recently discovered ESR unfortunately and am not caught up on the nomenclature. What does YC refer to? And what brand capacitors would you recommend?

Sorry if I missed it above but what fan are you using? Usually the old Deer PSU's have really low quality fans with dry bearings.

It's still got the noisy Deer fan. I've got a pile of old 80mm fans laying around so when this one starts growling I'll just swap it.

[Pentium4]

Re: ATX PSU, 13.8v

The secondary transistors are still in the factory configuration so I expect the SBL3045PT is 5v, the SBL2045CT is 3.3v, and the SB1660CT is 12v (this corresponds to the label's ratings). I haven't done any tracing/measurements yet, however, to verify which transistor is controlling which rail. I assume I could easily put the SBL3045PT on the 12v rail, correct?

I would put the SBL3045PT on the 5V, the SBL2045CT on the 12V, and the SB1660CT on the 3.3V and see how it does. Are there by chance any slots to add more rectifiers in parallel?

I've only recently discovered ESR unfortunately and am not caught up on the nomenclature. What does YC refer to? And what brand capacitors would you recommend?

Sorry, YC is the brand of most of those caps, they have a saturn logo on them. ESR is the resistance through the capacitor. In PSU's you want low ESR, but not ultra low ESR. Over time from being used, heat, or just sitting for long periods of time the ESR will slowly start to rise, especially with lower quality capacitors in an old design like that. My favorite for PSU's is nichicon PW, Chemi-con KY, Panasonic FC.

It's still got the noisy Deer fan. I've got a pile of old 80mm fans laying around so when this one starts growling I'll just swap it.

Is it one that just says "Sleeve Bearing" on it? Might want to add a drop of oil under the sticker sleeve bearing fans running at 100% speed wear out pretty fast especially if the bearing is dry.

[Pinhead]

Re: ATX PSU, 13.8v

I would put the SBL3045PT on the 5V, the SBL2045CT on the 12V, and the SB1660CT on the 3.3V and see how it does. Are there by chance any slots to add more rectifiers in parallel?

Not doubting your wisdom, just trying to learn -- why not use the 30a Schottky on the 12v rail? I'm not using 3.3 or 5 volts at all (and may eventually eliminate both). Is there something in the specs that I've overlooked that would preclude the SBL3045PT from being used at 14v?

I don't believe there are any extra holes on the board for more rectifiers but I'll double-check tonight.

[jordan1611]

Re: ATX PSU, 13.8v

Does anyone know of a substitute A1972 transistor
thanks

[Th3_uN1Qu3]

Re: ATX PSU, 13.8v

Not doubting your wisdom, just trying to learn -- why not use the 30a Schottky on the 12v rail? I'm not using 3.3 or 5 volts at all (and may eventually eliminate both). Is there something in the specs that I've overlooked that would preclude the SBL3045PT from being used at 14v?

It's a good idea to remove the 3.3v and 5v rails if you're not using them - however, make triple sure you removed all monitoring first, otherwise your power supply will not start anymore. You may need to make some resistor voltage dividers that make 5v and 3.3v to keep that 2005 circuit happy - what i do is i remove that altogether and add in a daughterboard with a TL494 and LM339 which gives me maximum flexibility. I can get whatever voltages i want (by rewinding the transformer as well) and cook my own protections.

Now, for that diode, the reverse voltage rating is a bit close for comfort.

The way a SMPS works is voltage is pulsed through the transformer, then smoothed out by the inductor and output capacitors. Obviously, for regulation to work, the input voltage must be higher than the output voltage. The pulsed waveform on the secondary of the 12v winding is actually significantly higher than 12v, if you have an oscilloscope you could check. If you remove the feedback altogether and leave the PSU running with no load (note: i don't recommend actually doing this), you're going to get around 23-27v on the 12v output, and the voltage on the transformer will be still higher than that, because some of that pulse is dropped across the output inductor. And then there will be ringing on top of the pulse because there doesn't exist a perfect transformer or perfect semiconductors, which will make for even higher peak voltage as seen by the diode.

In all power supplies i've encountered so far, the 12v rectifier had a minimum of 60v reverse rating.

[Pinhead]

Re: ATX PSU, 13.8v

The Logisys PSU that I've begun to disassemble seems to have at least used some relatively robust Schottky diodes -- FCH30A06 to be exact.

--- FCH30A06 ---
Repetitive Peak Reverse Voltage - 60v
RMS Forward Current - 33.3a
Peak Forward Voltage - 0.69v
Dual Diodes - Cathode Common

From what I can see, this look to be a suitable replacement. What say you?

The primary switchers in the Logisys PSU are SBW13009-S's which can handle 12 amps so I'll be swapping them, as well.

The next question is, what do you suppose I need to do to eliminate the "squeal" that I've been getting when talking on AM?

RF has to be getting back into the PSU which is causing an inductor coil to vibrate. Am I correct to assume the 12v PI Filter coil will likely be the culprit? I have a few spares that are much taller and use larger wire. Should I upgrade these filters?

Or should I add some small bypass capacitors to the 12v output? Maybe I need to wrap the output wires around a toroid to stop RF from traveling down the wires? Does anybody have any suggestions?

[Pinhead]

Re: ATX PSU, 13.8v

Here's a requested pic of the primary transformer and a pic of the secondary PI filters. Should I replace the 12v PI filter with a larger coil?

[Th3_uN1Qu3]

Re: ATX PSU, 13.8v

The Logisys PSU that I've begun to disassemble seems to have at least used some relatively robust Schottky diodes -- FCH30A06 to be exact.

--- FCH30A06 ---
Repetitive Peak Reverse Voltage - 60v
RMS Forward Current - 33.3a
Peak Forward Voltage - 0.69v
Dual Diodes - Cathode Common

From what I can see, this look to be a suitable replacement. What say you?

The primary switchers in the Logisys PSU are SBW13009-S's which can handle 12 amps so I'll be swapping them, as well.

Yes, those are perfectly good replacements. Swap ahead!

The next question is, what do you suppose I need to do to eliminate the "squeal" that I've been getting when talking on AM?

It can be a number of things. I don't think it has to do with RF energy entering the PSU tho - squealing means the feedback loop is unstable. Before you go crazy on any RF shielding, try a resistor load of the same power and see if it still squeals.

[Pinhead]

Re: ATX PSU, 13.8v

[QUOTE=Th3_uN1Qu3;390979]Yes, those are perfectly good replacements. Swap ahead!

Good deal.

It can be a number of things. I don't think it has to do with RF energy entering the PSU tho - squealing means the feedback loop is unstable. Before you go crazy on any RF shielding, try a resistor load of the same power and see if it still squeals.

Hmm it may be difficult to find a 200 to 250 watt resistive load. I'll see what I can find.

[Pinhead]

Re: ATX PSU, 13.8v

I got all of the components swapped, plugged the PSU in and turned it on and all was well. I then put the board back in to the case and got it all assembled, fired it up to begin testing in earnest, and as soon as I flipped the power switch I heard a loud SNAP and saw a quick flash of light. I still had my finger on the switch so I immediately turned it back off.

When I took the board back out of the enclosure to take a closer look, I noticed a burn on the plastic film directly below one of the SBW13009's that I had just soldered in.

It seems that, while the solder joints weren't shorted, they must have been close enough to allow arcing. I'm not sure why it worked outside of the case and shorted once I put it all together... Perhaps the plastic film has a slightly lower dielectric strength than air and gave it just enough path to arc.

It seems that some troubleshooting is now in order.

[budm]

Re: ATX PSU, 13.8v

"Hmm it may be difficult to find a 200 to 250 watt resistive load" You can use 4 car hi-beam head lights for load just turn the lamp on one at a time since the cold resistance will be very low (about .2 Ohm) or use 10 Ohms 10W in series with each lamp then bypass the resistor once the lamp starts glowing.

[Behemot]

Re: ATX PSU, 13.8v

Don't forget on thermistors to reduce the startup current.

As for plastic package, it is usually tested for 2-3 kV strength.

[Th3_uN1Qu3]

Re: ATX PSU, 13.8v

Use a "dim bulb tester" next time. Likely saved me hundreds of blown parts by now.

[Pinhead]

Re: ATX PSU, 13.8v

Use a "dim bulb tester" next time. Likely saved me hundreds of blown parts by now.

I assume you mean I should place a light bulb (or some other type of resistance) in series with the power supply to act as a current limiter? Not a bad idea, as at "idle" a working power supply should draw very little current and therefore drop very little voltage across the bulb, but in a "short circuit" situation such as I saw, there is less of a chance of popping, as most of the voltage (and power) will be across the bulb. Good idea.

[budm]

Re: ATX PSU, 13.8v

You can make one of these boxes:
http://s807.photobucket.com/user/bud...tml?sort=3&o=0

http://s807.photobucket.com/user/bud...tml?sort=3&o=2

[momaka]

Re: ATX PSU, 13.8v

Here's a requested pic of the primary transformer and a pic of the secondary PI filters. Should I replace the 12v PI filter with a larger coil?

I notice in the second picture, you ask what is the second toroid for?
- This is for filtering the 3.3V rail only. The 3.3V rail on these Deer power supplies uses a mag-amp circuit to regulate itself. You may notice that between the two heat sinks of the power supply, there is another even smaller toroid. One side is connected to a 5V tap on the transformer and the other is connected to the 3.3V schottky rectifier. This toroid is used for the regulation of the 3.3V rail.

And yes, light bulb trick is a very good idea, indeed. What I've made is a box with 2 plugs and 1 light switch. The plugs are wired in series, so that on one of them I plug in the current-limiting device (usually a desk lamp with a 40 to 60W incandescent light bulb), and the other plug is where I connect the device I want to test. The switch connects/disconnects the Live from the wall (so it's like a master ON/OFF switch) - not necessary, but I like to have it just in case.