512k sweep/power supply help

[JDW]

According to page 332 (Appendix C) in Larry Pina's Macintosh Repair and Upgrade Secrets, Q11 is the following type of component:

T-NPN, 12A, 100W, TO-220

Apple OEM Replacements:

2SC2827

2SC3039

2SC2810

3rd Party Replacement:

Philips ECG 379

(12A, 100W, High Voltage, High Speed, Switching NPN-Si transistor)

 

[Patnukem]

thank you I might have to get that book.

 

[tomlee59]

it does not seem like Q11 is working when I test it.

How did you test it?

(And Pina's information about the 2SC2810 seems to be in error. I am looking at the datasheet now, and it says that the 2810 is a 50W device; 7A peak Ic, and 500V BVcbo; it's the same info that I had at the time I put together my own schematics, and summarized there -- patnukem, see the schematics at the end of that pdf; this info is there for every semiconductor I could identify)

 

[Patnukem]

Multimeter with transistor tester. it worked to correctly test the bad one in my lisa. yeah I was confused because I mentioned the Q11: 7A, 4-500V, 50W, 18MHz up top from your pdf as you said. Id rather use one I have here to see if that is whats wrong for sure then have to order something.

 

[JDW]

Tom, I wouldn't be surprised if Pina was in error because I've spotted other errors in his books. But for the most part, his information is quite accurate.

Pina says Q11 is to be rated at 12A. So when I looked up those Apple OEM "replacements" listed in his book, I too thought he would be in error because the data I see shows them to be rated at only 7A. However, I believe Pina is simply saying that those 3 parts are "what he himself has seen in various Apple analog boards through the years," and he therefore deems them "OEM replacements." Pina then offers the Philips ECG379 part as a proper 12A replacement.

So I think going with the Philips part or similar is the most prudent choice.

 

[Patnukem]

http://www.moyerelectronics.com/ seems to agree that ECG379 or NTE379 (they have a sheet for this one) is the replacement. they have a nice cross reference for ecg and nte, I had been looking for one this is very helpful. I may just stick with buying a new analog board since I do not have one that will work here and if I have to order anything I may as well just get the whole analog board. I will keep looking I may have one buried.

 

[Patnukem]

I replaced the two 100uf 200v caps. and reinstalled Q11 then started it up and It sparked just like before but dimmed slowly. I also noted that the brightness knob on the front seemed bad and would flicker brighter and then darker as it faded slowly, I also noted that the sides were waving in and out (like water) I will replace more caps I have here and see if it gets better.

 

[tomlee59]

Tom, I wouldn't be surprised if Pina was in error because I've spotted other errors in his books. But for the most part, his information is quite accurate.
Pina says Q11 is to be rated at 12A. So when I looked up those Apple OEM "replacements" listed in his book, I too thought he would be in error because the data I see shows them to be rated at only 7A. However, I believe Pina is simply saying that those 3 parts are "what he himself has seen in various Apple analog boards through the years," and he therefore deems them "OEM replacements." Pina then offers the Philips ECG379 part as a proper 12A replacement.

So I think going with the Philips part or similar is the most prudent choice.

I heartily agree that Pina is to be trusted in most matters. I suspect that he was subtly (or not so subtly) guiding the user to an upgraded part. In switchers, however, there is a tradeoff among the various parameters, and it's not always better to seek a higher-current part. In switchers, there is considerable dissipation in the transition between states, so switching speed is important. Higher-current parts can be more sluggish, and using them can actually be self-defeating. Given that Q11 seems not to have any unusual failure modes, I'd normally be inclined to leave well enough alone.

That said, it appears, from what you've found from Pina, that Apple used a variety of transistor types for Q11 (I myself have not noticed this, so thanks for letting me know), so evidently any of the ones he cites have been proven to work well.

 

[tomlee59]

Multimeter with transistor tester.

Yes, but how did you use it? Did you remove the transistor first? Or did you leave the transistor in circuit? The only reliable way to test a transistor is out of circuit. If you do it in situ, the test can be confounded by other things tied to the device.

From the intermittent behavior you are reporting, it's highly unlikely to be a transistor problem. Transistors rarely "fade out" repeatedly in this manner; they either work, or die.

 

[Patnukem]

I removed the transistor completely and used it as the instructions say to and it had no response at all not even a blip, I used it to test a good one the day before so I know it works and its only 4 days old (the multimeter). It is odd though my board did not have a connector to the Collector or middle pin is was just in a hole on the board with no solder or leads anywhere near it so I assume it is not meant to be connected. is there a proper schematic for this analog board, that pdf shows Q11 having all 3 pins connected.

 

[tomlee59]

Yes, all three terminals of the transistor are used.

The middle terminal is electrically connected to the tab.

I still have lingering doubts about the test. It is very rare for a transistor to behave in the intermittent way you describe. As I mentioned, either a transistor works, or it is dead. It is extremely unusual for a transistor to go back and forth between these two states. I cannot recall ever having encountered this myself. Indeed, I can't recall even hearing about anyone else encountering such a phenomenon. Depending on the precise method used by the multimeter, it is possible to get a false result. In the case of power devices, it is frequently adequate to do the following:

1) Check for a good diode between base and emitter.

2) Check for a good diode between base and collector.

3) Verify that there is no continuity between collector and emitter.

You also mention sparking. Is this something you see, or is it something you hear?

And have you looked very carefully at C1? Is there *any* bulging on the top?

Have you carefully inspected the yoke connector? Not only should the connections on the analog board be resoldered, but you should also unmate the two connector halves and inspect all 8 of the pins for signs of arcing, melting, oxidation, etc.

 

[Patnukem]

Yes, all three terminals of the transistor are used.
The middle terminal is electrically connected to the tab.

ah ok I should have figure.

I still have lingering doubts about the test. It is very rare for a transistor to behave in the intermittent way you describe. As I mentioned, either a transistor works, or it is dead. It is extremely unusual for a transistor to go back and forth between these two states. I cannot recall ever having encountered this myself. Indeed, I can't recall even hearing about anyone else encountering such a phenomenon. Depending on the precise method used by the multimeter, it is possible to get a false result. In the case of power devices, it is frequently adequate to do the following:
1) Check for a good diode between base and emitter.

2) Check for a good diode between base and collector.

3) Verify that there is no continuity between collector and emitter.

yeah It is a cheap multimeter has worked every other time I have used it, but who knows. I will recheck it with another meter that has a diode check.

You also mention sparking. Is this something you see, or is it something you hear?

Well its when I turn it on so I can not see it, but the other day I had it opened and discharged the crt and it was just sitting there and it sparked

around C35 C36 and Q11 area that one I saw out of the corner of my eye and I could hear it.

And have you looked very carefully at C1? Is there *any* bulging on the top?

Yeah It looks ok I do not think I have a replacement (or 4 1uf disk caps)

Have you carefully inspected the yoke connector? Not only should the connections on the analog board be resoldered, but you should also unmate the two connector halves and inspect all 8 of the pins for signs of arcing, melting, oxidation, etc.

is this J1 you are talking about, I did test from the solder up to the yolk for continuity and they all were ok with no resistance. (The pins are not real but are nails I used that fit snug into the other side since I do not know where J1 went after it was removed from the board a long time ago) I did not think these had enough juice to ark to other parallel pins (I assume you mean ark to the corresponding connector and not to parallel pins) I could use some electrical tape to insulate them more and see if that helps.

Also should I try to find a real replacement for the C6 I made and posted about?

Quick stupid question I am in need of C6 and its a 22uf 50v and I have 2 10uf and a 2.2uf that are 50 or higher volts will the extra 0.2uf going to be too much extra tolerance for the 512k if I install them in parellel? I can always use 2 1uf ones but I would like to put as few in there as possible and I do not want to order one capacitor online and get charged a bunch for shipping. Radioshack has them but only in 35v (they are so lame) I do not think it will be but I just wanted to check.

 

[Patnukem]

I put some more caps (noticed one was a bit swollen at the top not enough to notice unless I looked real close) I had in it, took out the transistor and tested it again it was good, (even the transistor tester that said it was good worked so it must not have been connecting properly)

re did some of the solder and it turned on and seems to be stable (now I have to fix the (floppy and the brightness dail on the front) and it should be useable.

Thank you everyone that posted here to help me. I feel accomplished this week (fixed a lisa and a 512k mac) I must be stuck in the 80's.

 

[JRL]

Congrats! That must have taken quite a while...

 

[Patnukem]

Thank you, I am still going to get a new analog board as a backup since all these parts were donated from other machines, or made.

 

[JDW]

So that Q11 transistor was good after all?

 

[Patnukem]

YEAH STUPID $3 MULTIMETER, IT WORKS WELL FOR THAT MUCH.

 

[Patnukem]

it does seem to fade slowly still but does stay on. I have a feeling its just the brightness knob as I said above. I have some knobs from the IIc monitor maybe one will work.

 

[JRL]

Here are the rest of the things you need.

400k drive(needs to be relubricated)

Brightness knob

 

[Patnukem]

I actually fixed the drive today. I used silicone lubricant (the store was closed to get the lithium white lube that is suggested) but I cleaned it first (with alc.) then lubed it with the silicone cleaner/lube spray I had. I worked it in until it became mechanically sound. I also replaced the missing felt pad with one from a disk ][ drive and put it back together and read the first disk I put in it. I am going to wait on the brightness knob since I am getting a brand new analog board (along with a mouse and keyboard) and I am going to use this repaired one as a back up. I did adjust the max bright and it seems to be nice and bright (probably a bit too bright) I will adjust it later. Also I seem to be getting weird feedback from the floppy drive. It does make some cool noises though. Thanks for the links I did see the 400k drives but It seems I have fixed mine I do not know yet if the one in my lisa works but we will see when I get some disks made on my 512k. It should work great for making the disks (lisa websites state the best machines besides lisa's to make disks on are the 128k, 512k, plus since they have the 400k drive)