Power Mac G4 MDD "Mirror-Door" won't power up

[Toasty]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

You should check the IC that drives Q1 and Q2 since when MOSFETs shorted out they will take out the IC with it, it will damage the new MOSFETs if the Gate drive outputs of the IC stuck on.

A good precaution, but after several dozen repairs, I've never found this to be bad. ML4824IP1 - It's on the daughterboard next to the main heatsink.

https://www.fairchildsemi.com/produc.../ML4824-1.html

On the opposite side of that board is the quad comparator KA339.

https://www.fairchildsemi.com/produc...ors/KA339.html

T

[Toasty]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

Yes, have all the parts in stock.

See this post: https://www.badcaps.net/forum/showpo...8&postcount=49

I'm in NJ.

T

[Toasty]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

Wow, this one's back from the dead.. lol

I probably have all the parts. I'll take a look and post back.

T

This is the Samsung unit, correct?

[lti]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

It looks like the filter cap fails open on these. From page 2 in this thread:

IF the mains cap is a CapXon HP, it -MUST BE REPLACED-. There is a marked history of failures in this series of capacitor (even from a Ukraine electronics site via kc8adu) in which the cap either goes open or shorts then opens. It invariably causes the main switchers Q1 & Q2 to fail (STW20NM50FD) and the primary fuse (8A/250v, Slo-Blo, 5x20) to open (usually violently) as the blackened glass indicates. I recommend a replacement with a ceramic bodied fuse here (Schurter 0001.2513). Another failure can be the destruction of the inrush limiter TH2 (SCK-048).

Check everything else budm and Agent24 mentioned as well.

[haroldblum]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

Thanks for the info I will replace Q1 and Q2 also

By the way would you happen to know where I can find these parts?
ntc thermistor sck-048 4ohms 8amps
I will try looking up the transistors after I figure out how to get the nut and bolt off of them to de-solder
Also the fuse size was 240v 8amps can I use 120v 8amps

Can I use a 5ohm 8amp for the thermistor?

[budm]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

You should check the IC that drives Q1 and Q2 since when MOSFETs shorted out they will take out the IC with it, it will damage the new MOSFETs if the Gate drive outputs of the IC stuck on.

[Agent24]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

Check the bridge rectifier and filter capacitors for shorts too, use a series lightbulb when powering up for testing after you change parts

[haroldblum]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

TH2 is a NTC type thermistor. Its job is to limit the inrush current of the PSU. It should be replaced with a new one after you replace switching transistors Q1 and Q2.

Since the PSU blows fuses, then the switching transistors are almost certainly shorted. The thermistor burning is just a side effect from having a short up ahead, which puts very high current thru it. It goes in series with the circuit, so it cannot be the cause of blowing fuses.

Thanks for the info I will replace Q1 and Q2 also

By the way would you happen to know where I can find these parts?
ntc thermistor sck-048 4ohms 8amps
I will try looking up the transistors after I figure out how to get the nut and bolt off of them to de-solder
Also the fuse size was 240v 8amps can I use 120v 8amps

[haroldblum]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

I fund the lettering on the part
SCK
048

SPECIFICATION

Max. Approx.
Zero Power Steady Resistance Thermal Thermal Operating Part No Resistance State at Max. Dissipation Time
Temperature
at 25ºC Current Current Constant Constant (ºC) (Ω) at 25ºC at 25ºC (mW/ºC) (Sec.)
(A) (mΩ)


SCK-1R38 1.3 8 64 21 59 -40~+200
SCK-1R58 1.5 8 62 21 66 -40~+200
SCK-2R58 2.5 8 75 20 63 -40~+200
SCK-037 3 7 91 21 73 -40~+200
SCK-046 4 6 117 20 62 -40~+200
SCK-048 4 8 87 27 76 -40~+200
SCK-056 5 6 121 20 66 -40~+200
SCK-057 5 7 98 20 65 -40~+200
SCK-065 6 5 159 20 74 -40~+200
SCK-075 7 5 161 20 79 -40~+200
SCK-078 7 8 108 28 57 -40~+200
SCK-086 8 6 130 20 68 -40~+200
SCK-105 10 5 178 20 79 -40~+200
SCK-125 12 5 185 19 59 -40~+200
SCK-152 15 2 704 21 77 -40~+200
SCK-154 15 4 261 19 79 -40~+200
SCK-164 16 4 261 19 79 -40~+200
SCK-204 20 4 283 18 90 -40~+200
SCK-224 22 4 308 20 80 -40~+200
SCK-253 25 3 425 21 76 -40~+200
SCK-403 40 3 511 22 83 -40~+200
SCK-404 40 4 360 22 86 -40~+200
SCK-473 47 3 501 20 77 -40~+200
SCK-802X 80 2.5 693 19 71 -40~+200
SCK-1202 120 2 1010 19 65 -40~+200

DIMENSION




Unit. mm

Disc Ø D max. L. min. d. nor. P. nor. T. max.
15 16.5 29 1.0±0.02 8±1.2 6
V-I CHARACTERISTIC CURVE


POWER DERATING CURVE





R-T CHARACTERISTIC CURVE

[Th3_uN1Qu3]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

TH2 is a NTC type thermistor. Its job is to limit the inrush current of the PSU. It should be replaced with a new one after you replace switching transistors Q1 and Q2.

Since the PSU blows fuses, then the switching transistors are almost certainly shorted. The thermistor burning is just a side effect from having a short up ahead, which puts very high current thru it. It goes in series with the circuit, so it cannot be the cause of blowing fuses.

[haroldblum]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

[QUOTE=haroldblum;515716]

[haroldblum]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

I fund the lettering on the part
SCK
048

Max. Approx.
Zero Power Steady Resistance Thermal Thermal Operating
Part No Resistance State at Max. Dissipation Time Temperature
at 25ºC Current Current Constant Constant (ºC)
(Ω) at 25ºC at 25ºC (mW/ºC) (Sec.)
(A) (mΩ)
SCK-1R38 1.3 8 64 21 59 -40~+200
SCK-1R58 1.5 8 62 21 66 -40~+200
SCK-2R58 2.5 8 75 20 63 -40~+200
SCK-037 3 7 91 21 73 -40~+200
SCK-046 4 6 117 20 62 -40~+200
SCK-048 4 8 87 27 76 -40~+200-------------------------------------------------
SCK-056 5 6 121 20 66 -40~+200
SCK-057 5 7 98 20 65 -40~+200
SCK-065 6 5 159 20 74 -40~+200
SCK-075 7 5 161 20 79 -40~+200
SCK-078 7 8 108 28 57 -40~+200
SCK-086 8 6 130 20 68 -40~+200
SCK-105 10 5 178 20 79 -40~+200
SCK-125 12 5 185 19 59 -40~+200
SCK-152 15 2 704 21 77 -40~+200
SCK-154 15 4 261 19 79 -40~+200
SCK-164 16 4 261 19 79 -40~+200
SCK-204 20 4 283 18 90 -40~+200
SCK-224 22 4 308 20 80 -40~+200
SCK-253 25 3 425 21 76 -40~+200
SCK-403 40 3 511 22 83 -40~+200
SCK-404 40 4 360 22 86 -40~+200
SCK-473 47 3 501 20 77 -40~+200
SCK-802X 80 2.5 693 19 71 -40~+200
SCK-1202 120 2 1010 19 65 -40~+200

[haroldblum]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

The TH2 on my power supply is blowing fuses and shooting sparks. It is green in color.. I don't know if that is because it is burnt and originally was blue. Does any one know what size it is and which way the lettering goes.

I fund the lettering on the part
SCK
048

[haroldblum]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

The TH2 on my power supply is blowing fuses and shooting sparks. It is green in color.. I don't know if that is because it is burnt and originally was blue. Does any one know what size it is and which way the lettering goes.

[jasonsjwou]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

I fixed my G4 power supply by following these posts.

I replaced

the main cap (CapXon 390uf 420v) with Nichicon 390uf 450V
Q1 and Q2 (had hell of time re-inserting the thin leads of thermister on the heatsink)

Powers up, runs fine.

Thank you for all those who walked this path in the past.

[Ed in SoDak]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

Hi to the group,
Revisiting the thread after a long hiatus. The MDD that I fixed due to help from this thread is still going strong some years later in constant use.

I just picked up today a better base MDD system also with a bad PS. As it happens, it's the same MDD I referenced earlier in the thread that had its PS replaced with one bought off ebay, so it was on it's second supply when it failed again.

As a quick check of the system, I plan to swap in the repaired supply from my working MDD. At that point, if it works, I'll have two toasted PSs to experiment on (hopefully not a third), and a far better system to play with in the meantime.

Earlier in the thread I posted about my diagnosis of this MDD's original PS and it follows the typical failure pattern. I didn't bother to repair it then, as my first attempt has worked so well. So I guess I have a backup plan in place if I terminally smoke something.

I can't add much to the most recent replies tech-wise, but do have some advice on basic desoldering techniques and the typical tools we laypersons can easily acquire or already own.

Desoldering braid works best if "fresh,", that is, bright-colored and un-oxidized. If your braid is dull-looking, try buffing the outer parts so they are shiny. It doesn't take much oxidation on copper to really mess with a good and fast-acting wicking job. Tinning the very end of the braid will help it take up solder faster than bare braid. If your pre-tinning of the braid moves up too far, just snip off the fully-soldered part, leaving just a fraction of an inch pre-tinned.

Same goes for your iron or other reworking tools. If the tip of your iron is wasted away and pitted, it won't present a large heating surface to the work. File it down to a nice flat wedge-shaped point and tin it with a bit of solder before going to work. The melted solder and flux on the tip helps to increase the heated mass and transfers the heat to the work more quickly.

Be sure to clean and correct any connections of the tip to the iron itself. The same condition of oxidation, as seen when using old desoldering braid, can happen in the tip's connection to the iron. Just loosening and retightening the tip's fastener can really help heat transfer.

As an iron heats and is flexed under use, the tip may loosen a bit, so I give the tip a little gentle tightening with a pair of needlenose or whatever. You can get a feel for what's loose and flaky, or what's actually twisting the tip's connection instead of getting a snug fit. Heat transfer is everything, even moreso than which specific brand of iron or tip you choose. It's not "more heat" that does it, but "effective" application of the heat.

Soldering isn't like a slow-cooker on your stove. Leaving the iron on a joint while waiting for things to happen and pressing harder on the joint will cause more problems than you're trying to fix. A slow-heating iron can do more damage by all the prodding and pressing and waiting and such you do while trying to force it to do the job. That just heats up the parts you want to insulate from heat and can start to lift circuit traces, cook out the soldering flux and generally waste your time.

It's best to go in with a fresh, clean, hot iron, with tinned tip and get the job done quick. Watch that solder flow liquid, then get off it when it makes a good bond. Move nothing, lest the joint crystallize in a "cold-solder" joint. Those appear pasty-looking and gray instead of shiny and smooth. Desoldering is similar. The sooner you can flow the entire joint and suck it off quickly, the better. There's countless documented cases of failures due to to factory-soldered bad connections. Sometimes the "fix" is nothing more than "reflowing" the factory connections. Not so here with the Mirror Door Mac and its known capcaitor issues, but certainly applies elsewhere. It happens far more frequently than one might think that an expensive flat-panel TV or toehr gear fials for nothing more than a sloppy solder joint. So learn to recognize how they look or at least be aware. High-frequencies used in a circuit, like a Gigahertz Mac, have been known to actually crystallize a solder joint, kinda like how an opera singr can shatter a champagne glass. We don't see that here as a big issue, but it was a major fault in some earlier Macs and plenty of TVs and computer monitors with a tube-based display. Next up on the failure roster is bad caps! So these same techniques can be applied elsewhere and that's a lot of the fun in trouble-shooting. Reapplication of what you've learned elsewhere and it grows from there. For some of us, the "fixing" part is more fun than the using!

Stagger the work. If there's 30 joints to desolder, don't just work on up the line unless t moving quickly. It mught be better to do a few at one end and move to the other end for a spell, in order to avoid concentrating al your work in one small area with chance to coll down in-between. If you work a bit in one area, then move over a bit or waiting a bit before removing more solder, you can avoid overheating some particular part.

There's plenty of junk electronics to practice on to be found curbside or in your own garage. Work on some junk before the all-important Mac Resurrection Project. Time spent learning the proper way will smooth the real goal when you feel you're capable. You're not on the clock, so take your time and make it fun, instead of a high-pressure bound-to-fail March into the Unknown.

I did my MDD repair and all my other R&R adventures with that lowly Radio Shack desoldering bulb. Mine is the type that's permanently married to an iron. I first make sure the screw to the pencil iron's shaft is tight to the desoldering dealy. I find you need to tend to it like an old car-battery terminal to make sure it's getting maximum heat transfer. How soon after plugging it in before it melts some solder put on the tip is a good indicator of heat transfer.

It will heat a good-sized joint without help, but always try to get the tip positioned so it doesn't just suck air instead of solder. It helps to prime the tip or even add some fresh solder to the joint you're trying to clear. It seems counter-productive, but the added heat mass and conjoined flow actually hasten melting of the joint. The extra solder can help with making sure the hole in the tip isn't sucking air, as it will be surrounded with solder and thus make a good vacuum. A little additional assist from a prod or probe while the side of the tip's hole works the lead will help free the component's lead from the hole. You don't need to do much more than make sure the lead is free and clear such that yout can wiggle it a bit.

If it resists these attempts at freeing it, at least you get an idea which lead(s) might be sticking and will take additional work to remove the complicated heatsink assembly. After knowing that, I can take further steps, such as trying again, resoldering the hole only to reapply the desoldering tool to try for a better "suck," or I may use cautious prying, etc. while lifting the loose end of the assembly along with a well-positioned iron tip to the offending junctions. Practice makes...

It's kinda crude, but I often use the side of the hole of the tip itself as a sort of hook or pry to help get under bent-over leads and help get them straight enough to remove the part cleanly. Get the joint melted and try to snag the bent-over lead using the solder sucker's tip.

Get the holes cleared so you can wiggle them to prove they're freed up or at least reveal the ones that need more attention, don't waste time, allow time for a sip of beer or a smoke while an area recovers from your invasion and you'll make headway without unduly stressing the parts or circuit traces.

I browse the forums while the cooling occurs and sometimes I may only get a part or two swapped out i an evening. I'm easily distracted, LOL! But it wasn't a high-stress task to take it slow and easy that way, either, and I can take stock of things better by taking my time.

The only thing I try to do fast is to shorten the effective time the iron is on the board.

-Ed
Solder quick, then inspect. Move on.

[Agent24]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

All lines must be loaded. Loading only one will cause an imbalance and shut it down. Have you tried it plugged into machine?

For my AcBel I did a first test after repair and powered it up according to this: http://haertle.ch/tinker/ps/

Short PS_ON to GND (as usual) but you also need a 10-30 Ohm resistor (you will need more than 1/4watt) between GND and 3.3v Sense line. I used a 15 ohm 5-watt.

The PSU powered on without any other load and all voltages were correct and stable.

I do not know if the same can be done with the Samsung version, however.

[Toasty]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

>>After replace of R122-2/increased resistivity from 10 ohm to 3.5Mohm<<

What does that mean? You changed a resistor? Why?

All lines must be loaded. Loading only one will cause an imbalance and shut it down. Have you tried it plugged into machine?

[gogo75]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

Update-main cap 375V SB and ON.PFC part of ML4824 work fine.After replace of R122-2/increased resistivity from 10 ohm to 3.5Mohm/ PFC "blink" halogen bulb of 150W.
Still no change in secondary side-voltages drops....

[gogo75]

Re: Power Mac G4 MDD "Mirror-Door" won't power up

Yesterday I can't work on it,but today I replaced all small caps with Rubycon and the results:
1. no load 300V main cap
2. load 3.3V/1 Ohm drop to2.95V;5V/4 Ohm drop to 3.6V;295V main cap
3. load 12V bigger than 0.3A-PSU shuts down

Pins 4 and 19 of SUB1 are some feedback from opto's,
8 is PWM to 11N50/switchers of T1/,
9 is feedback PWM,
10 feedback from PFC,
5,11,15,16 not connected
12 +15V from 7815
6,7,13,14 GND
17,18 temp. sensor
1,2 line monitor
3 PFC switchers 20NM50
I want to know voltages on pins 4&19 /sb,min load/ because both changed from "SB" to "ON".Wich one is "PG"/"on" ,"start" command/ and wich one is feedback?