KRK Rokit 5 G2 repair - 100Hz sawtooth, warm-up issues - where to look next?

[DavidA]

Hi,

After powering off my KRK Rokit5 G2s for a few weeks, I had an issue with one where it would emit a loud "blaart" at power on (from cold), and then a steady audible 100 Hz tone (and harmonics) and a lot of scratchy, random noise every few seconds (in bursts). Note that I'm on mains power at 50 Hz, not 60 Hz.

I have an electronics background but it's more on the digital side, so although I'm comfortable with a 'scope and a soldering iron, I don't really have the hands-on experience with debugging analog amplifiers.

I've watched Dave's videos, and some others, on this topic. I've scraped/chipped out as much of the black gunk as I could, and replaced the four main capacitors on the Amp board with the same kind but next voltage up (35V -> 50V). Unfortunately, this did not solve any of these issues (and I measured the capacitance of the replaced capactitors with a DMM and they all came out roughly as expected). So it has shiny new caps but still has the same issues.

Now I'm trying to track down the source of this 100 Hz buzz, but all I have to go on is the RPG 6 service guide, which I found here. As far as I can tell, it's similar enough to the 5, but I'm cautious. I've checked the 15V supply to the Amp board and it looks reasonably clean (+/-20mV ripple AFAICT)

I've noticed something interesting - the scratchy, random bursts of white-ish noise (like a dirty pot, but they occur without me touching anything) go away after the unit has been powered on for about 10 minutes. Some kind of warm-up thing? Also, the "blaart" when I turn it on doesn't occur when it's in this warm state. If I let it "cool down" for at least 5 minutes, both those symptoms reappear.

But the 100 Hz buzz is very consistent, does not seem to change over time, and is not affected by the input signal or the Volume knob. The Hi and Lo signals from the PreAmp board look OK - they are "correct" but have a lot of very high frequency noise on them (well above hearing range), and perhaps that's normal anyway. At the sub speaker ports, with the speaker connected, the 100 Hz sawtooth is about 100 mVpp.

Actually, as I've been writing this, I've just noticed that the hum is gradually getting quieter, or maybe I'm getting used to it... but no, my 'scope shows that it's dropped to around 40 mVpp... I think it's getting quieter the longer I leave the unit powered, but it's not going away completely.

I'd really appreciate any suggestions or advice on how to proceed with fixing this speaker.

[feedback.loop]

Check all capacitors. Why just four?

[feedback.loop]

[DavidA]

Check all capacitors. Why just four?

Fair point - in almost every video I've watched, it was the one (or more) of the four main power-supply caps that was at fault. There's a handful of smaller electrolytic caps so I'll go through and replace those too. Do you think any of the non-polarized caps might be responsible?

Thanks for the video, I haven't seen that one before. it looks very useful.

[Audiorepair]

There's a couple of 1000uF 35v caps around the centre of the PCB that supply the tweeter power amp IC and possibly other stuff.

Scope the tweeter IC power supply pins and you might see a whole load of ripple due to the non functioning of these caps.

You won't see this ripple on the 15v rails.



In fact don't scope the power amp pins, there is too much risk of shorting pins and blowing stuff up.
The power supply pins are either side of the output pin.


 A clue here is that this ripple only comes through the tweeters, so is not 100Hz, it is in fact harmonics of a 100Hz ripple confined within the range of the tweeter output.  So quite nasty and buzzy.
Just follow the PCB traces back to the caps, and scope the caps, or just replace them.

[DavidA]

A clue here is that this ripple only comes through the tweeters, so is not 100Hz, it is in fact harmonics of a 100Hz ripple confined within the range of the tweeter output.  So quite nasty and buzzy.
Just follow the PCB traces back to the caps, and scope the caps, or just replace them.

Sorry, I perhaps wasn't clear - the 100 Hz + harmonics is mostly on the sub (LF) channel, not the tweeter (HF) channel. I've replaced all four large caps so far, am looking to compare testpoint-by-testpoint with my other (working) speaker.

On the tweeter (HF) there's just a momentary click at power on - no bad tones or scratchy noises or anything like that. It's all on the sub (LF).

Probing around a bit, comparing with my working speaker, I can definitely see the 100 Hz sawtooth AC on top of the +15V and -15V into the TDA2052 LF amplifier. It's about +/- 100mV on both rails. On the working speaker, the output (pin 1) seems clean, whereas on the faulty speaker, the output has the sawtooth.

Unfortunately, the Rokit 6 schematic that I have differs significantly from the Rokit 5, because the former uses a TDA7296 with more pins than the TDA2052. I've been trying to use the 6's HF circuit as a guide to the 5's LF circuit, but it gets really confusing as the component numbering doesn't match in some cases, and half the component IDs are still obscured by the black residue (or scratched off when I tried to remove the black goop).

From a design point of view, what's most likely to be the fault here? This sawtooth is clearly getting into the audio signal from the main rails, it's pretty much identical in shape. Obviously not a shorted cap to a voltage rail, as I'd see the TDA2052 output sitting well away from 0V. And if a cap has gone open-circuit, then why would that introduce noise from the power rails? What part of the circuit could fail causing this AC noise on the power rails to get into the audio signal?

I *could* replace every cap on the board but that's going to be a major operation.

I really wish I could find the schematic for the Rokit 5...

[DavidA]

Scope the tweeter IC power supply pins and you might see a whole load of ripple due to the non functioning of these caps.

It's actually the LF circuit that is bad. So applying this suggestion to the LF IC, I see +24V and -24V going into pins 2 and 4 on the TDA2052. Both of these have +/- 200mV 100Hz sawtooth on top of the DC voltage. The "good" speaker shows the same.

On the bad speaker, the signal into pin 7 looks clean but the output on pin 1 has the 100 Hz noise present, around +/- 100mV. On the good speaker, the output on pin 1 is a nice clean amplification of the input signal.

You won't see this ripple on the 15v rails.

Ok, I do on the +/- 24V into the LF power amp though, on both the "good" and "bad" speaker.

In fact don't scope the power amp pins, there is too much risk of shorting pins and blowing stuff up.
The power supply pins are either side of the output pin.

I was careful and scoped them anyway - didn't short anything

But the output pin isn't in-between the power supply pins, it's on the end - pin 1...

[Audiorepair]

You haven't yet considered ground reference.

If the 15v rails don't have the same ground reference that the power amp power supply to the power amp chips have, then you may well get unexplained hum.

That black goo might have compromised a ground somewhere, or is leaking AC into a DC rail somehow.


One other less likely scenario is that the mute circuit power supply has ripple, so that the mute Fets might be AM modulating the signal at 100 Hz.

[Audiorepair]

Ground reference also applies to your scope probe.

What you may see as ripple might actually be a ground reference issue instead.