• Please review our updated Terms and Rules here

SE30 repair caps

clh333

Veteran Member
Joined
Feb 22, 2015
Messages
1,443
Location
Cleveland, OH, USA
Last week, spurred by interest in the Mac SE, I acquired an SE/30 to add to the collection. The seller represented the machine as "working", but that may have been optimistic: It powered up exactly three times before refusing to boot. Internal inspection revealed the usual problems: dead battery, dirty FD, etc.

However, when it DID boot I noticed two "peculiarities" that after some research I now believe indicate failing mobo capacitors. The signs were: very faint chime on power-up and occasional distortion of the video. The video distortion was not regular, and had to do with horizontal rather than vertical sweep. "Tearing" might describe the phenomenon, and it was most prevalent when first started.

I took out the RAM and ROM and washed the board. There was no sign of leakage on the traces or any indication of lifting, for which I am thankful. So: I can send out the board to someone who will recap it for ~ $80 or I can take on the job myself. I have performed through-hole replacement on other boards but this is SMT, a different (and foreign) kettle of fish. Don't have a hot-air rework station, never made such a repair.

Any advice?

-CH-
 
Don't have a hot-air rework station, never made such a repair.

You don't want to use hot air on any electrolytic capacitor, it'll cause damage to them and potentially make them vent.

I'd recommend getting a very fine pencil soldering iron and using needle pliers or tweezers to rock the capacitor gently off the board one side at a time. I've used this method on an LC III and Performa 600 and it works fairly well in most places. There were a few cases where I had to lift one side up and snip the cap off because the other side was cramped next to a connector. Removing the remnants was easy after that.

You'll want to use at least a 40W iron though because pencil tips don't have much heat capacity and tend to get drained by large ground planes fast.
 
You don't want to use hot air on any electrolytic capacitor, it'll cause damage to them and potentially make them vent.

I'd recommend getting a very fine pencil soldering iron and using needle pliers or tweezers to rock the capacitor gently off the board one side at a time. I've used this method on an LC III and Performa 600 and it works fairly well in most places. There were a few cases where I had to lift one side up and snip the cap off because the other side was cramped next to a connector. Removing the remnants was easy after that.

You'll want to use at least a 40W iron though because pencil tips don't have much heat capacity and tend to get drained by large ground planes fast.

Thanks to all for their suggestions. "Knowing your limitations" is a good idea, and in this instance I'm going to bite the bullet and send the board to someone who recaps for a living. I just don't want to screw it up after all I went through** to acquire it.

However, recognizing that this is not the last time I will face this issue, I have also purchased a rework station and will practice on old junk until I feel I have the capability to perform the work myself.

-CH-


** I was looking on eBay for SE-related stuff and contacted a seller in the Chicago area who was selling a defunct SE. He mentioned that he had a "working" SE/30 and we struck a deal. I drove out to Chicago to pick up the unit and bring it back to Cleveland: 400 miles each way + $50 tolls + $60 gas + $114 hotel (wife was with me) + $80 meals = expensive. The unit lit up in Chicago and twice more since, then the HD failed. Sometimes you're the windshield, sometimes you're the bug.
 
It's really not THAT hard to recap the SE/30, once you remove the old caps, the pads are quite large, my SE/30 was the first SMT work I ever did, re-did mine with radial electrolytics, it's not pretty but it does work. I just twisted the old caps off, fluxed and tinned the pads, then soldered the bent legs of the radial caps down.

My first SMT solder job on the SE/30, done with a $10 iron, note that the burnt crap is just the contaminated flux from the old leaked electrolyte, I took the pic before cleanup:
IMG_0203.jpg
I have got a bit better since that first one a few years ago, quality tools, quality parts. Now I use a nicer temp controlled hakko clone, and SMT tantalum caps:
IMG_0134.jpg
 
re-did mine with radial electrolytics, it's not pretty but it does work. I just twisted the old caps off, fluxed and tinned the pads, then soldered the bent legs of the radial caps down.

Clearly you are ahead of where I am on the curve.

I had read someone's account of taking this approach and considered it. But if anything mysterious were to occur I knew I would always suspect my soldering job. I figured it was better to get it back to spec before tinkering. Might help to have a backup board, too...

I'll let you know how my apprenticeship fares. Thanks for the encouragement!

-CH-
 
I have seen photos of folks who have done radial electrolytics...I worry about it trying to pull the pads off from the weight/tension...but I've also not heard of any problems. My preference are the little MLCC capacitors - they just barely bridge the distance between the pads, leaving plenty of room to solder, unlike the equivalent electrolytics to the originals.

Unsoldering the originals has not worked well for me...pulled a pad on one of them while attempting it. I've done it two different ways, one more conservative than the other. Take a pair of wire cutters and cut the cap apart at the base, then take a pair of pliers and gently remove the remainder of it, and when you're done, you'll have two legs standing up. The other method, which I have heard several others cite good success with, is to grab it with a pair of pliers, push down, and rotate. May have to spin it a few times, but the legs will snap off, leaving just a little sliver of metal attached to the pad. Yes, it sounds like something only a lunatic would do, but it seems to work fine. May not be a good idea if the board has any damage from leaked electrolyte, however!

Once removed, I put on a little flux over the pads, then heat up and remove the remaining parts of the legs. Then come back with a bit of solder braid and wick up the remaining solder from the pads. Follow that up with some high-grade isopropyl alcohol and q-tips and clean up the flux residue. You'll also need to clean up any remaining electrolyte from the old leaking capacitors.

As mentioned, I use the MLCC capacitors. Flow some solder onto one pad. Drop the capacitor on the board, and I use a toothpick to put it in position - one side will stand up a bit on the solder blob you just created. Push down on top of it with the toothpick and re-flow the solder onto the capacitor, at which point it will be held in place. Spin the board around and solder in the other side, then come back to the first side if it doesn't look good enough.

Before I tried the first time, someone told me they actually liked SMT replacements more than through-hole...I thought they were crazy, but after doing a few, it's really not that bad. I've been thinking I should try putting this together into a YouTube video... Done succinctly, it might actually be a good use of YouTube... :)

Wesley
 
As mentioned, I use the MLCC capacitors.

The MLCC or multilayer chip capacitor is non-polarized, where the SE/30 electrolytics are polarized. Not a problem?

MLCCs come in three classes depending on what use they are intended for; some with higher capacity, other with higher accuracy. What were the specs you used?

Resin-core solder or something else?

Thanks for the tip,

-CH-
 
I use tantalums now, they are polarized, though the line/mark on them marks "+" and the line/mark on the electrolytic marks "-", so you have to watch that when working with them. They come in all kinds of size packages, so you should be able to find one that fits the pads. Working with smaller parts like this you would need a tweezers or something to hold the part, unlike the radials that are large enough to hold with your hand while you solder.

Practice up, and you can fix the next one you come across ;-)
 
... Working with smaller parts like this you would need a tweezers or something to hold the part, unlike the radials that are large enough to hold with your hand while you solder..
Try using a small lump of Blu Tack to temporarily hold parts in place.
 
Try using a small lump of Blu Tack to temporarily hold parts in place.

Adds Blu Tack to my shopping list......

Actually, I just bought like 30 yards of polyimide Kapton tape, I could probably just tape it in place ;-)
 
I'd recap the analog board and look for AC ripple from the power supply before sending the mobo out.

I obtained a spare SE to practice on. This one was sold as DOA / FPO. But it has a mobo, CRT, drives and a power supply so I figured I would take your advice and test the power supply. I have the supply removed and detached from the computer, and have removed the lid. I can see that I will have to desolder the 120 VAC input and the power switch in order to remove the board. Peering inside I can see a number of electrolytics and one looks (to me) as though it has spewed. Take a look and tell me what you think:

Thanks,

-CH-

IMG_2240.JPG
 
The picture is too small and taken from too far away to tell what the substance is.

None of the caps in that area appear to be bulging (tops are flat) and none are rocked to one side which would indicate the bottom plug has blown out, so I'd think that isn't electrolyte.

Vented electrolyte tends to be a darker brown color and have a smooth look to it, whatever that is looks to have lots of bubbles in it. Some kind of glue perhaps?
 
Took the power supply apart and removed the suspect caps (at the end of the board where the supply wires are located. Measured the capacitance of each removed electrolytic: Some looked suspect; most were below spec but within 10% of rated value. I will replace them anyway just for practice.

But what about the others? There are some tiny caps that are apparently in switching circuits and one huge 640 uF at the input end. Given that nothing has failed, should I replace those as well? Thanks again for your replies.

-CH-

Pictures attached.

SE_PS_1.jpgSE_PS_2.jpgSE_PS_3.jpg
 
Measured the capacitance of each removed electrolytic

Capacitance isn't the only thing you need to check. They could check fine capacitance wise but have high ESR or high ripple. I would replace all capacitors on principle because they are 28 years old.

most were below spec but within 10% of rated value.

Electrolytic capacitors usually have a 20% rating.
 
Capacitance isn't the only thing you need to check. They could check fine capacitance wise but have high ESR or high ripple. ...

I will go ahead and replace all the electrolytics. Is there anything else I should replace on general principles, e.g. age?

My little capacitance checker does not have an ESR measurement capability but I ordered a tester from an eBay reseller. I have a scope and can presumably identify ripple with that. The question is: What values are acceptable or unacceptable? This must be published somewhere...

Thanks again.

-CH
 
I will go ahead and replace all the electrolytics. Is there anything else I should replace on general principles, e.g. age?

Everyone seems to hate old tantalum capacitors, but I don't think there are any on the Mac SE/30. Other usual culprits are components that look like they got really hot (like the PCB under them is dark brown) or burned. The most common things I find other than caps are burned resistors, diodes or transistors (usually as a result of a failed cap.)

My little capacitance checker does not have an ESR measurement capability but I ordered a tester from an eBay reseller. I have a scope and can presumably identify ripple with that.

Cheap Ebay capacitor testers are really only good for weeding out really bad capacitors because they use such a low testing voltage. Bad capacitors can appear fine at low voltages, but start showing how bad they really are when you get close to the working voltage of the capacitor.

True capacitor testers which can test at the working voltage of capacitors are very expensive:
www.ebay.com/itm/272578851054

The question is: What values are acceptable or unacceptable? This must be published somewhere...

Capacitor data sheets usually have the ripple current listed, but that's a hair pulling exercise when you have a mishmash of different brands and series, especially for 28 year old caps. If you need some generic metric to go off of, the ATX power spec is a good start.

Max allowed ripple current in millivolts:
+12V - 120 mV
+5V - 50 mV
+3.3V - 50 mV
-12V - 120 mV
 
Cheap Ebay capacitor testers are really only good for weeding out really bad capacitors because they use such a low testing voltage. Bad capacitors can appear fine at low voltages, but start showing how bad they really are when you get close to the working voltage of the capacitor.

Aha.

True capacitor testers which can test at the working voltage of capacitors are very expensive:

Noted. Maybe for Xmas.

If you need some generic metric to go off of, the ATX power spec is a good start.

The problem with being as uninformed as I am is that you know THAT you don't know but you don't know WHAT you don't know. Thanks for the heads-up.

-CH-
 
Most computer capacitors are low voltage (outside of power supplies), so a 9V battery powered cheap Chinese ESR tester is decent for the job.
 
Back
Top