Left channel hum.

[raymondp1138]

Hi,


Recently my BeePre experienced its first really hot days of summer and didn't like it. Left channel produced a 120Hz hum of top of the music. Did a bit of research on this forum and found that sagging mains voltage can cause the issue. So I measured my mains, and the voltage was 115VAC. When the mains went back to a healthy 120VAC, the BeePre was happy again.


The theory is that sagging mains causes a dropout voltage condition on the LM1085 regulator, thereby preventing it from properly filtering out AC to the 300B filament. I measured ~12.5V on the Vin of the regulator. Normally, I would get 13.4V. Vout was ~9.8Vdc.


A solution was to use a step up transformer to boost the mains voltage. Another was to use a separate filament transformer having greater than 6.3V. But I'd like to also explore other options. I know the dc filament kit for the Kaiju uses a regulator with a lower dropout voltage than the LM1085. Is there a compatible substitute for the LM1085 with a lower dropout voltage? Would increasing the 33uF bypass capacitor hanging off of the regulator Vin help?


I'm also wondering if it is at all possible for the Darlington bipolar transistors (I think that's what they are) to sustain oscillation under these conditions.


Thank you!

 

[Deke609]

I'm very interested in this as well.

 

[Paul Birkeland]

The theory is that sagging mains causes a dropout voltage condition on the LM1085 regulator, thereby preventing it from properly filtering out AC to the 300B filament. I measured ~12.5V on the Vin of the regulator. Normally, I would get 13.4V. Vout was ~9.8Vdc.

Yes, low line voltage is a problem.

A solution was to use a step up transformer to boost the mains voltage.

If your line voltage moves around a lot, a regenerator is a better solution.

Another was to use a separate filament transformer having greater than 6.3V.

When your line voltage goes back up, the unregulated supply voltage will also increase, and this will increase thermal dissipation in the regulator.

But I'd like to also explore other options. I know the dc filament kit for the Kaiju uses a regulator with a lower dropout voltage than the LM1085. Is there a compatible substitute for the LM1085 with a lower dropout voltage? Would increasing the 33uF bypass capacitor hanging off of the regulator Vin help?

We tried a few lower dropout regulators when developing the BeePre, and we soon discovered that lower dropout seemed to come with increased noise for most regulators.  There are two 10,000uF caps in series at Vin already, so taking the 33uF cap up to even 330uF is unlikely to make much of a difference.

I'm also wondering if it is at all possible for the Darlington bipolar transistors (I think that's what they are) to sustain oscillation under these conditions.

Are you talking about the NPN and PNP transistors in the noise filter by the 1085 regulator?

 

[raymondp1138]

Are you talking about the NPN and PNP transistors in the noise filter by the 1085 regulator?

If I'm seeing it correctly, it looks like there's a feedback path from POS OUT back to the input of the Darlington circuit through an RC network. Also, conceptually, this Darlington circuit is trying to assert control of POS OUT, while at the same time the LM1085 is also trying to assert control of POS OUT. I wonder if all of that together under certain conditions would cause oscillation.


I'm not too worried about this hum problem as my line voltage is usually stable at 120V. I suspect this problem would only surface on those occasional very hot days when the electrical grid is strained.


Thanks fo your replies, PB.

 

[Paul Birkeland]

If you started removing capacitors from the board, I'm quite certain you could get it to oscillate.

 

[raymondp1138]

Haha. That I would not do.

 

[aroide]

I had a similar issue.  Family room AV setup and listening room power are all on the same breaker with a long wire run from my panel.  Result is voltage drops to 109-110 at times. 

My painfully expensive solution was to add a PS Audio regenerator to the listening room.  With stable 120V (and harmonic distortion reduced from about 4% to .3), system is stable and sounds even better than before.  Just to try and help minimize the voltage drop, I put conductive silver paste in all the house wiring wire nut connections for this circuit... no change.

It probably would have been cheaper to have an electrician add a new line to the listening room, but I didn't want to risk the results not improving things much.

 

[Paul Birkeland]

A mild step-up transformer or autoformer would've solved the problem for significantly less money as well. 

 

[jjvornov]

Paul

Any recommendations for an autoformer? I'm having this problem this summer as well. Load on the circuit from the window AC or elsewhere in the house will cause hum for a while. There are lots of choices out there and I don't think I have a problem except the BeePre and it's intermittent.

James

 

[Paul Birkeland]

The easy way is to get a 120V:6.3V CT transformer with a 2A secondary, then wire it as a boost autoformer.  You can have +0V, +3V, and +6V with that transformer, just grab a 4P3T rotary switch rated to switch 120V AC.

I just built such a box for Alonzo a few weeks ago.

 

[Deke609]

An even easier and potentially cheaper fix is to use a variac. I've been using one since last fall and it works great. It also helps with keeping the filament voltage on my EML tubes in other amps within Jac's narrow specified range.

But PB's post reminds me that I had intended to build a larger 10A-capable version of PB/Alonzo's box using a huge autoformer and a couple of buck/boost transformers.  Thanks for the reminder!

cheers, Derek