Kenwood KA-3500 distortion left channel only! Guidance needed.

I get distortion at Qe5 on the collector only when turned up past halfway. Base and emitter are -38v. See pics 1 and 2 below.

Also emitter of Qe11 is very high up above 10vac at full tilt and it is supposed to be -1.2 in the manual. See 3rd pic below.
 

Attachments

  • IMG_20170903_093708.jpg
    IMG_20170903_093708.jpg
    46.3 KB · Views: 14
  • IMG_20170903_093720.jpg
    IMG_20170903_093720.jpg
    41.7 KB · Views: 13
  • IMG_20170903_094712.jpg
    IMG_20170903_094712.jpg
    49.3 KB · Views: 13
Qe17 is bad. Do I need to replace qe19 as well didn't notice but legs are dark and corroded on both but 19 passes. What would be a good replacement and if replacing both should they be matched?
 

Attachments

  • IMG_20170903_111203.jpg
    IMG_20170903_111203.jpg
    59.6 KB · Views: 13
Last edited:
32/34mV dc offset is a little high however it's not going to cause you any serious problems, can sort it out/improve it later.

With distortion on Qe17b and Qe19b, need to follow the bold line back in the service manual. Means distortion at Qe5c
Suggest check at Qe5b (measure at Re9), think we've been here before.

Assuming no distortion then Qe5 (2SC1885) or Qe7 (2SA620) suspect faulty.

As part of general house keeping you should check the bias, see service manual, 40mV across the emitter resistors.
 
32/34mV dc offset is a little high however it's not going to cause you any serious problems, can sort it out/improve it later.

With distortion on Qe17b and Qe19b, need to follow the bold line back in the service manual. Means distortion at Qe5c
Suggest check at Qe5b (measure at Re9), think we've been here before.

Assuming no distortion then Qe5 (2SC1885) or Qe7 (2SA620) suspect faulty.

As part of general house keeping you should check the bias, see service manual, 40mV across the emitter resistors.

Yeah I don't know why offset is so high I gain matched both transistors with a gain of 371. I was reading and it looks like matching Vbe would give better results. I ordered 30 of those transistors so I'll try to bring that down later.

I decided to remove one transistor at a time to test and Qe17 was faulty so I'm replacing Qe19 as well with a beefier to-126 package. Echowars seems to have a lot of great reference posts on the Kenwood amps as well as replacement recommendations.

Yeah I saw the 40ma bias in the user manual. So would 40mv get me close enough without removing a leg of the resistor to test current?
 
If Qe17 tested faulty then certainly replace it. However earlier post
indicated distortion of Qe5c so likely source is before Qe17 eg, Qe5 and 7.
Qe17 probably taken out by the faulty Qe.

The SM has a typo where they say to use a current meter. You need a dc mV meter.
Aim is for 40mV across the emitter resistors as per the pix. Certainly don't lift
a resistor leg.
 
If Qe17 tested faulty then certainly replace it. However earlier post
indicated distortion of Qe5c so likely source is before Qe17 eg, Qe5 and 7.
Qe17 probably taken out by the faulty Qe.

The SM has a typo where they say to use a current meter. You need a dc mV meter.
Aim is for 40mV across the emitter resistors as per the pix. Certainly don't lift
a resistor leg.

Yeah I ended up just removing and testing every transistor on the left channel only one that was defective was 17. Some had some gain in the 140s but didn't look to see what the specs we're. Signal was clean as long as volume was below half but voltages where off after q7.
 
Well replaced q17 and q19 adjusted right channel to 40mv bias but no matter where I turn the pot it sits at near 0mv. I do get a clean signal out both channels according to the scope.
 
Recapping?
3Vdc offset on lect channel, all outputs, zeners, electrolyics and bias resistor replaced.
Left channel now working but heavily distorted with 1KHz input. Right channel good.
Problem with power amp, preamp ruled out.
Qe1,2,3,4 replaced with kSA992.
Distortion on positive half cycle of sine wave.
Distortion at Qe17b and Qe19b

Please note voltages shown in service manual are for dc bias condition not AC signal.

Distortion at Qe5c only when turned up past halfway.
DC offset 32-34mV

Qe17 suspect faulty, replaced. Qe19 also replaced.
Bias trimmer has no affect, 0mV at emitter resistors.

Putting this all together, key points are,
- distortion is volume dependent, probably needing a few "volts" for b-e(?) junction
to fail. Static/cold testing probably won't show the fault
- bias trimmer has no affect, 0mV at emitter resistors

Candidates for distortion are Qe5 and Qe7
Qe7, VRe1, THe1 and bad solder joints could cause the bias issue.

Suggest,
With amp distorting, measure dc voltage at Qe17b, repeat for Qe19b
expect +/-1.2Vdc as per SM.
Measure Vdc at Re17/Re19 junction, expect 26Vdc as per sm
Measure voltage drop across Re25

I would be inclined to replace both Qe5 (2SC1885?) and Qe7(2SA620)
 
I know for sure qe17b and qe19b are not 1.2/-1.2. also the emitter of qe1 is not .65 like mentioned in the se I've replaced all caps with nichicon. Replaced power resistors and a few of the bigger diodes.

Qe5 and 7 tested good when I pulled them
 
Please note the bold line from Re3, Ce1, Qe1b, Qe1c, Qe5b etc...
this is the audio path. Previously you've noted distortion at Qe5c.
3(?) possibilities
- bs in bs out, distortion received at input (Qe5b)
- Qe5 faulty and "introduces" the distortion
- Qe7 faulty and "introduces" the distortion

Other Qe's 17,, 21,, already replaced.

Regarding unable to bias. VRe1 is used to adjust Qe7 Vb-e, controlling
how hard it works. However if there is no (or very low) current flowing
through Re27 and VRe1 then the change in Qe7 Vb-e will be very small and
have no impact on bias, or if Qe7 is faulty then it may not respond correctly
to the change in Vb-e. Also note the current through VRe1 comes from Qe26, Re17
Re19, Re27,,, Failure/breakdown in any of these components is possible.

Certainly the voltages mentioned in the sm are for guidance. Qe1,3 emitters may be
0.55-0.75V, I'd be concerned if they are outside this range. Voltages at Qe7 emitter
and collector (+/-1.2V) may be different however they should be the "same", ie equal
driving of pos and neg half cycles.
 
Just replaced every transistor on the left channel plus q26 and q27 and base of q19 and q17 are still not +-1.2v in fact they are in the 20v range
 
Last edited:
Hav
What happens to the voltage if you start switching things on and off (input selector, loudness, speaker selector)? Or have you already done that? And what have you not replaced in the L channel? All replacement transistors same pinouts?
Have not tried that yet but I cannot adjust bias at all and base of q19 and q17 are up near 20v at full volume. All electrolytic caps and all transistors. And power resistors.
 
Weird thing is I get a clean output on the speaker terminals at 18v. Only switches that effect base of q19 and q17 is the tape rec switches.
 
High voltage at Qe17b indicates low current flow through Re19 (26V one side)
Probably cause by low current through Re27, VRe1, THe1, Re25, Re23. This is the bias circuit.
With amp powered off/unplugged suggest measure resistance of Re27, VRe1, Re25 and Re23, in circuit
should be ok, however Re25 and Re23 may continually climb corresponding to charging Ce17.

One of these Re's has failed open or drifted very high, VRe1 is probable candidate.
 
I'm about to systematically pull inspect and replace every component in the left channel. I've even inspected diode 13 14 7 and 9 all check out with my tester. Even the transistors I pulled and replaced were all good except q17 19 21 and 23. First I just replaced 21 and 23 and that got rid of the major dc offset but still had distortion at the out put over half volume so I replaced q1 and q3 still same thing. Removed q17 and 19 one of them was bad so replaced. Now output looks clean according to scope but no bias adjustment and base of q17 and 19 is in the 20 volt range both positive. Down around 5v at idle. Also replaced emitter resistors and all electrolytic caps along with the zeners on the power board. Even checked the two 330 ohm resistors. I tested each transistor to determine orientation before installing so as long as there are no mistakes on the PCB as far as labeling that should be fine. Only thing I can think of is a resistor or ceramic cap somewhere.
 
High voltage at Qe17b indicates low current flow through Re19 (26V one side)
Probably cause by low current through Re27, VRe1, THe1, Re25, Re23. This is the bias circuit.
With amp powered off/unplugged suggest measure resistance of Re27, VRe1, Re25 and Re23, in circuit
should be ok, however Re25 and Re23 may continually climb corresponding to charging Ce17.

One of these Re's has failed open or drifted very high, VRe1 is probable candidate.
Perfect I'll start here and report back
 
Just measured re27,25,23 only one out of spec was 25 at 1.5k vs 1.8k
Ve1 was 40ohm to 1k ohm appears to be fine
BTW the1 is 11k when I disconnected the other day
 
Last edited:
Last measurements look reasonable, maybe THe1 is a little high but need to check datasheets, would no explain symptoms.

With parts replaced etc... I thinking that this might be a protection issue, it does match the symptoms of distorttion when volume
increases past xx.

Qe9,11,13,14 form an over current protection circuit driven by Re45,47. These are the "bias" emitter resistors. Could you measure the
voltage drop across each resistor/one probe on each side. Anything above 500mV would be a concern. Maybe then increase the volume
to distortion and redo measurement.

When you try to adjust the bias, is the volume at minimum? The protection circuit should be off.
 
Back
Top Bottom