Kenwood KAF-1030 magic smoke. Worthy lesson by trying to fix?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
So I made a previous thread on a Kenwood KAF-3030R and jinxed myself saying I had a working KAF-1030.

I only payed £25 for it yet was working fine with unusual discolouration at one end of 90% of main board's resistors and tracks, never seen that before. The only sign of previous work was a resolder of the speaker terminals to the board, on both Kenwoods so I'm thinking that was hand done from factory. It had board discolouration around two transistors and I understand discolouration can be normal yet felt they were ready to fail.

I had this running for a week or two with no issues yet yesterday I turned it on, the audio started popping within a minute so I went to switch it off as a sizzling noise and pop from the amp while black smoke came out, that smoke isn't meant to be in there right? Least it got released lol

I opened it up, the discolouration around them two transistors was worse and two fusible resistors burnt out next to them. I've replaced those resistors with the same type and a few 100uF caps after just for the hell of it, removed all the transistors in circuit around them. I was thinking of changing the diodes, main filter caps, 470uF caps too. Until...

Today amongst further inspection I check to board on the secondary transformer board to find a blown ceramic capacitor.

Is the transformers thermal fuse likely to have blown?
All fuses have continuity still.

Is it needed to replace all the transistors around the burnt resistors on the main board?

The only work I did to this is reflowed a few suspect solder joints and used contact cleaner on all usual pots, connections and switches etc.

These Kenwoods aren't anything I like yet I'm learning so much as I enjoy this hobby. I don't mind discarding them but would love to get them going again so any advice would be much appreciated


  • IMG_20191210_122816.jpg
    620.7 KB · Views: 219
  • IMG_20191210_122824.jpg
    822.4 KB · Views: 223
  • IMG_20191217_140213.jpg
    224.1 KB · Views: 216
  • IMG_20191217_140337.jpg
    679.7 KB · Views: 210
Looks like there may be excessive current consumption (even a short) on +/-15 V somewhere, causing the regulator pass transistors to overheat. (Check Q146/148 for shorted junctions. If none found, they're probably still OK.) R279 should normally be dropping about 1.4 V, so <30 mA. Check resistance between both + and -15 V and ground, anything suspicious? It should be around 500 ohms at the very least, probably higher. I would also check ZD101 just in case, it should behave like a diode.

It is unusual for a leaded ceramic cap to fail like that. Normally these are very reliable (their surface-mount brethren have a substantially worse track record). I would check its circuit position (looks like it may even be ahead of the rectifier?) and whether it may have been subjected to excessive voltage. If not, you may be dealing with a spontaneous failure / a rare bad part, rather than anything to do with your other problem.

I think I found it, C660, it's across the secondaries for main power (+/-42.5 V, so ~30-35 VAC x2?). 100 nF Z class dielectric. This part needs to sustain roughly 85 V on a regular basis, so should probably be rated something like 160 V at least. May be rather hard to decipher now that the magic smoke has been let out.
In any case, this is a part not essential to operation itself. It is merely needed for EMI suppression. Yank it out, clean up the mess, look up / order a replacement and move on to the more pressing issues.

The thermal fuse is unlikely to be bothered. You can confirm this by measuring resistance across the power plug's L and N terminals while having the power switch turned on. Should be a few hundred ohms or so, certainly not open.
Thanks again sgrossklass, great detailed advice.

I've search since you posted lol all I found is 103 etc hundreds of them and only a handful of 104's, 4 have 25v printed on them the one other 104 0.1uF is exactly the same as the melted one, just 104 printed underlined.

I took the melted one out and put that in at least for now. Cleaned up the carbon to avoid problems. I will order some 250v's but was thinking is it better to put a blue poly X2 capacitor there instead? Or maybe a square yellow big capacitor like the one pictured as it can fit nicely soldered to the wire ends? Are these more safe?

I'll do the other checks now, thanks again


  • IMG_20191217_202801.jpg
    689.9 KB · Views: 118
  • IMG_20191217_202838.jpg
    421.6 KB · Views: 86
  • IMG_20191217_203132.jpg
    546.1 KB · Views: 91
I'm only getting a resistance of 11.5 ohms from the plugs live and neutral with the power button on. Doesn't look good does it?
Well, at least it isn't open, so the thermal fuse seems to be OK.

Keep in mind that primary resistance is being transformed over to secondary side (in addition to secondary side winding resistance). So that's about 1-2 ohms per secondary, plus x. When your load is 2x 4-8 ohms, you don't actually want that much more if you can help it.

That said, it's a rather low value for what I'd guess can't be much more than a 120 VA transformer. Rod Elliott lists values in this range as typical for a 160 VA toroidal, and toroidals have lower than average primary resistances to begin with.

Still, if we assume that for some reason the primary were partially shorted (which in turn should not be happening before the thermal fuse blows), then:
1. The xfmr would get very warm indeed, with corresponding excessive power draw
2. secondaries would be substantially higher than spec
3. the first ones to get upset would be the big filter caps, which rarely have more than a 25% voltage headroom to offer
4. eventually either the thermal fuse or the mains fuse would blow
None of that is happening, is it?

If you have a reliable power meter (effective power / power factor display, not plagued by bad capacitors), you could check mains power consumption in idle. Such a little amp should be around 25 watts or so.

Now, as far as your presumed short on + or -15 V goes, the amp has a bunch of opamps to offer, and an electronic input selector (KIC9164AN, presumably a clone of the TC9164AN). An electronic input selector without too much in the way of input protection, to be precise - a mere 150 ohms in series after 150 pF in parallel. Pioneer had 470 ohms at least. I wouldn't rule out the possibility of the input selector getting zapped and suffering latchup. Usually that doesn't short out the rails though... but neither do most dead opamps.
Sorry for the late reply, I had a Kenwood KA-3020 SE arrive which I only paid £9 for on eBay for S or R as it powered on but no audio, only hum on both channels, and candle wax spillage inside. The picture didn't seem like the wax had burnt, shorting components, so presumed it only acted as an insulator. The preamp board was cracked from what looked like the input switch had a bad whack. Turned out good experience, after painstakingly repair about 40 tracks on the board I got audio in one channel through all inputs, a common issue with this model was bad solder joints on the headphones jack. Restored both channels fixing that and cleaned all the wax out and sounds great, really surprised. I can see why it had top reviews at the time. Anyway that's finished and I didn't want to get mixed up with too many units on the go yet wanted to get a list of parts sorted for one order.

Well impressed with your knowledge and great detailed reply. You make so much sense. I understand the microprocessor IC for the input switching can be suspect and the opamps. I'll keep this in mind. The microprocessor is common to fail on these, from what I've researched.

I misplaced the resistor I usually use to discharge big caps and used pliers on this kaf-1030, I don't like doing that. One had lots of voltage and the other had nothing. Does this mean anything?

I only have a extension lead with a power meter on I use for my PA gear. No numeric meter, just bars. Ironically I had to replace a cap in that to get it working the other week lol.

I have some new Nichicon 4700uF 63v caps doing nothing which will give some extra voltage headroom, so I can use them. I've yet to test the transistors after it, is it likely to replace them all and the bridge rectifier regardless? I won't mind trying to replace most suspect components and opamps (likely I have used replacements of those) before the microprocessor to see if its failed. I feel I can get a clean + and - supply.

So much respect for the help.
I misplaced the resistor I usually use to discharge big caps and used pliers on this kaf-1030, I don't like doing that. One had lots of voltage and the other had nothing. Does this mean anything?
Possibly. Usually the power amps should be depleting most of the charge in the big caps after power-off. If that is not happening, you might have a bad trace or solder joint or a corroded jumper - check for continuity between the big caps and power amp (e.g. power transistor) supplies.
I did reflow suspect solder joints when it was working, as you'll know easy to miss one or one not visibly in perfect.

A corroded jumper not passing current is very likely. Corroded jumpers all over it with one end burnt marked the board around the hole, which could lead to broken tracks I guess. Resistors too. I'm not sure if it's had an ground surge previously or something.

I'll check for that continuity soon and take some pictures.
I have continuity from the main filter caps to the output transistors. I guess that rules out any broken tracks. I understand one or two of the overheated transistors joints are broken, may of been a cause of the problem if they were weak previously.

I'll put the new 4700uF 63v caps in, a new ceramic cap near the fuses, new 470uF 50v caps (C211&12) to 63v types, new diodes D101-104, 5 new transistors and possibly a new bridge rectifier although it does test well.


  • IMG_20191219_215806.jpg
    627.4 KB · Views: 52
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.