Memorex MLT3221 - No Power

[Dgtech]

This one is stumping me for the moment.

THE PROBLEM - No power at all. No power LED. No visual or audible changes when plugged in or power button pressed. "CON3" shows no output voltages to the digital board. IC8 is

CHECKED COMPONENTS - Both fuses, all electrolytics (under HS3 - heatsink 3) were checked and confirmed. MOSFET's all checked good.

CHANGED COMPONENTS - IC4 (KA431AZBU Shunt Regulator), IC8 (PWM chip), IC1 (on back)

If you notice Q1 wasnt on the board, it will be shortly. I just removed it for the moment. It is a 2n4403. It checks ok as well.
I have no documentation on this power supply. I find lots of people fix these on ebay. This board is listed on there either as IPOS 150 or 120. This one is the IPOS 150 Rev.3.3. (200-P00-HIVI150-CH).

[retiredcaps]

Originally Posted by Dgtech

THE PROBLEM - No power at all.

Do you see 5V standby voltage on CON3?

[Dgtech]

Nope, no 5 volts there.

[momaka]

Originally Posted by Dgtech

Nope, no 5 volts there.

That's a problem. The pin labeled "5vsb" should always have 5v present on it.

Seeing that there are HEC caps used on this power supply, it's very possible that the PSU start-up capacitor has dried. I see 2 small capacitors under the primary heatsink right behind Q6. One of them (or possibly both) is responsible for starting up the PSU. If they are of a known unreliable brand, change both.

Before you do that, though, check the voltage across the big primary capacitor (EC2?). There should be around 165v. If present, then change the two caps I mentioned above.

[Dgtech]

Here are some more readings:

165 VDC across EC2 (the big electrolytic)

68 VAC at J7 - near upper 2 mosfets - between C72 and C73.

Con3 - on the 12vdc pin I show a 1 Hz pulse that only goes as high as 9.8

VDC ( using fluke, min-max function)

80VDC on gate of Q17 - which I think is a driver mosfet for the lamp circuit.

On 5VSB there is a 1Hz pulse of a high of 1.5v

[PlainBill]

Originally Posted by Dgtech

Here are some more readings:

165 VDC across EC2 (the big electrolytic)

68 VAC at J7 - near upper 2 mosfets - between C72 and C73.

Con3 - on the 12vdc pin I show a 1 Hz pulse that only goes as high as 9.8

VDC ( using fluke, min-max function)

80VDC on gate of Q17 - which I think is a driver mosfet for the lamp circuit.

On 5VSB there is a 1Hz pulse of a high of 1.5v

Those measurements are consistent with either a bad start-up capacitor (those mentioned by retiredcaps) or an excessive load on one of the outputs.

PlainBill

[Dgtech]

I checked and , just for kicks, actually changed the caps under the heatsink on the primary side. No change in result. There is 80VDC on gate of one of the mosfets that drive the lamp circuit. This is making me think that the primary side should be ok. Does this judgement seem sound?

There was a mosfet that I question. This one is a Fairchild -FQPF 27po6. I checked it using two different fluke meters.

Fluke 1 shows
OFF = 336K OHMS ON = .2 OHMS

Fluke 2 shows
OFF = 886K OHMS ON= CANT SEEM TO TURN ON

With the fluke 1, I'd think that the off should have been higher resistance S-G.
With fluke 2, maybe the battery is dying?? I just changed the one in #1 last week.

[retiredcaps]

Originally Posted by Dgtech

Fluke 1 shows
OFF = 336K OHMS ON = .2 OHMS

I realize there are multiple ways of testing, but can you explain how you are measuring the mosfet with the terminology ON?

[Dgtech]

I connect the source to the meter negative and touch the gate with the positive. Then move the positive to the drain. Should be on at this point = very low resistance. I made a typo above on the other post when I said I measured from source to gate, I meant source to drain.

[Dgtech]

OK- I found my error with testing this mosfet. EARLY MORNING BLUNDER. This one is a P - channel vs an N -channel. REVERSE THE POLARITY OF THE METER LEADS. LOL.

I need sleep, or alcohol, which ever comes first!

This mosfet tests ok now. I still have no output on this power supply.

[momaka]

Check diode D6? and also small SMT diode D2 on the backside of the board.
If those are good, check resistance between 5vsb and ground. If it's not shorted, change the output capacitors on 5vsb - that would be EC15 and the two capacitors to the right of the inductor L6 in the second picture.

[Dgtech]

D2 looks fine - .5 Vdc on diode check however, does .158 sound right for D6(both diodes)? I'm used to diodes checking .5 Vdc or .7 but not .158. Do these fail like this? or do they usually burn open? Also, the one above it (on the same heatsink) measures .252 as well.

Here are the numbers on these:
D6 = FYPF2045DN
Above component = MBRF10100

Both datasheets are attatched. I cant seem to find the information from these to tell if both are good or not.

[Dgtech]

.

[momaka]

Originally Posted by Dgtech

however, does .158 sound right for D6(both diodes)?

I think there was a misunderstanding. The diode D6 I was talking about is located on the primary side, close to the transformer and Q6. Perhaps I misread the numbers on the board.

Originally Posted by Dgtech

Do these fail like this? or do they usually burn open?

Diodes *usually* become shorted when they fail.

However, the reading you got for those schottky diodes attached to the heatsinks (0.158 for both diodes) sound normal since schottky diodes have lower voltage drop (and hence better efficiency, which is why they are preferred for the outputs in PSUs).

Originally Posted by Dgtech

Both datasheets are attatched. I cant seem to find the information from these to tell if both are good or not.

For FYPF2045DN, if you look at Figure 1. Typical Forward Voltage Characteristics (on page 2) you will see a graph of Forward Curent vs. Forward Voltage Drop. As the Forward Current increases, the Forward Voltage Drop becomes larger. Since your multimeter is not putting a lot of current through the diode, the Forward Voltage Drop you get displayed on your multimeter's screen is fairly low. Hence the low readings.

[Dgtech]

It was my mistake, you were right about the part number. I checked out those components and found that they were good. I changed the cap just to be sure - same result.

I did find a shorted diode at D10. This diode is marked as HER304. I have circled it on the attatched photo. It is attatched to a heatsink and is standing upright. As also seen in the 2nd photo. This is the classic case of "I thought I checked that before". Taking notes would have made all the difference.

Now I have to find a solution for a match if I cant find the exact component.

[momaka]

Originally Posted by Dgtech

I did find a shorted diode at D10.

Is this out of circuit or in-circuit measurement? If in-circuit, measure it again out of circuit just to make sure.

Originally Posted by Dgtech

Now I have to find a solution for a match if I cant find the exact component.

Looking at the HER304 datasheet, just about any fast-recovery diode capable of 3A @ 50C should work.
Basically just match the Forward Current (Amp) rating and the Maximum Reverse Recovery Time rating. HER304 has 50 ns (nanoseconds) recovery time. As for the voltage rating - I doubt this PSU has any voltage on the secondary side above 30v, so even a 100v diode would likely work fine.

*EDIT*
Looking at second photo above, the solder joint between the heatsink and the diode seems cracked. Perhaps this is what caused the diode to short?

[Dgtech]

I apreaciate the help momaka. Yes, the readings were out of circuit. After checking Digikey and Fairchild, I finally found it at Allied. I will post the result once I get the part. Thanks again.

[Dgtech]

Everything works well after replacing that diode. All powered up and scanning for digital channels now. Thanks again momaka. I apreciated your help and guidance.

[momaka]

Hey don't thank me. After all, you found the shorted diode and did all of the work . Good job!

By the way, what diode did you use as a replacement? (you know, just in case anyone needs this as a reference.)

[Dgtech]

The replacement part was an NTE equivilent. NTE576