Octal Preamp Tubes and Filament Noise, or... Help! - Fixed!

[Alexo]

Hello all,

Well, I've been boasting about my 100 watt B15 for a little bit now, and it seems I'll have to eat my words because there's a pretty annoying buzz (i.e. upper harmonics of 60 hz, across most of the audio spectrum) that I can't seem to get out of my first preamp tube - a 6SL7. This problem has occurred in other builds of mine that used 6SL7's in the preamp, and if it wasn't for all the quiet amps out there using 6SL7's, I might be forced to conclude that this was inherent in the old pre-humbucking-heater tubes of the 40's and 50's.

The amp uses a solid state rectifier - UF5408's, the psu is very well filtered with 1uf Solens across the big caps at the first stage in the PSU, which is at 400uf. There are six, count' em six, zener diodes between the first and second nodes on the B+ chain, dropping the screens down to 300 volts (plates are at 570 volts) and .22uf orange drops across the screen filtering stages which are at 50uf - two of them forming a pi filter before hitting the screens. Following that, in parallel, are the preamp filters and a 220K/68K voltage divider forming a + voltage reference for the heaters at 70 volts, with a 50 uf filter cap there.

I've isolated the buzz/hash to the heater element on V1 by temporarily running its heaters off a 6 volt battery - noise gone! So here's what I have tried:

Elevated the heaters at various voltages between 10 and 70 volts.

Put smaller caps across the cap that's connected to the HCT voltage reference.

Tried .0001 through .1 caps across the zeners to act as snubbers.

Poked endlessly at lead dress - which should be sound.

Reversed the phase of the heater leads to V1.

Run a non-polarized cap across the heater leads - 120uf did quiet it down a bit, but this is just a band aid fix.

The thing is, the noise isn't really at 60hz, it doesn't hum, it just buzzes. Could this noise be created by the rectifier? Dirty power supply? I did use a filtered IEC socket, and they are UF diodes, so I thought I would be safe here, and I tried snubbers across the zeners, which, as I said, did nothing. The PT is very robust (400ma) so I don't think it's being strained either.

I'm beginning to think my only option is to build a DC heater supply for V1, but the PT I'm using has only one 6V3 tap at 6A, (no 5V tap) I'm using 5A, so I couldn't do the bridge rectifier I'd need to do... I could do a bridge just for the heaters for V1, but then I don't think I'd be able to elevate my other heaters, nor could I use the center tap. Urgh.

Any Ideas??

Thanks,

[Structo]

Sounds like you did most of the common fixes for it.
Kind of weird it would be only on one tube in the chain.

I'm sure you have tried tightly twisting the heater wires and different locations such as overhead and down flat against the floor near the back?

Curious, when you reversed the phase on V1, did the hum get worse or better?

[Alexo]

I'm using zip cord for the heaters, which keeps them as close as they would be if they were very tightly twisted, might sound suspect, but it's worked great for me in higher gain amps than this one. I didn't try running them in different places because I'd have to redo quite a bit of layout, but when I had V1 running off a battery, the pair of heater wires that were previously connected to V1 were just floating in the air, and I poked them around with a chopstick and couldn't get anything to hum or buzz by pushing them near other components, I also pushed them around a lot when they were connected and noticed no change.

I think the root of the issue is that there's this hash running along the heater filaments, it's not just 60 hz, which doesn't seem to be an issue in itself, but a ton of stuff above 60hz. Where does it come from? I don't know, maybe I should shield my heaters, but what a pain that would be..

Thanks for the suggestions though, it might be worth it to try running the heaters against the chassis, instead of in the air, for V1.

And I think the noise is along the whole heater line, it's just noticeable in V1 because that's the most sensitive point in the circuit. I also tried large cathode bypass caps on V1, as RCA recommends, but that only increased the gain and made the noise louder.

Reversing the phase had no audible effect, although it's helped in some other builds.

[Jana]

Here is a way to do ac and dc heaters and elevate them, all from one 6.3 volt winding.

[Structo]

I am not an amp guru but from my understanding, when using ac heaters, the tight twisting is done to eliminate the hum as it causes phase cancellation.

Just running the zip cord seems to me that there would be inductive transfer from wire to wire. Not sure why it would only effect V1.

[Andy Le Blanc]

It'll be 60Hz if its the heaters, 120hz if its thru the B+.
things to try; as mention'd, twist the wires.
Install a hum-dinger, humpot, 100 to 500 ohm pot. between the legs of the
6.3 v supply, with the wiper to ground or 20 to 60ish V+.
Another thing to try is to observe the polarities of the heaters even with a.c., and wire accordingly.
Another, try if the 6.3 winding has a center tap is to place a large film cap
to ground from the center tap.

Another issue that might be leading to the hum is if the bias resistor is bypassed or not.
You might have to do it even if it is not in the schematic, or the design.
you can get away with some supriseingly small bypass values on the order
of .005 to .02 without radically changing the gain structure of the stage.

[Alexo]

Thanks so much for the suggestions, guys. At this point I have literally tried everything that's been brought up here, and in the archives, and the winner is....

Jana's method for DC preamp heaters!!

Hoorah!!

That tube is actually running much closer to 6.3V than any of the others; there's still a little but of noise from some cc resistors and whatnot, but that grating zzzzxzxxzxxzxzx is vanquished. I used UF4007's, I figure their 1A ratings are probably safe in a bridge that's only dishing out 300ma.

Whew, I think it's time for the celebratory Jameson on the rocks.

Cheers!

[Jana]

cool beans!

[drhulsey]

... the tight twisting is done to eliminate the hum as it causes phase cancellation...

Structo is right. It's not the proximity of the wire, but the noise cancellation you get with the twist.

[davent]

Thanks so much for the suggestions, guys. At this point I have literally tried everything that's been brought up here, and in the archives, and the winner is....

Jana's method for DC preamp heaters!!

Hoorah!!

That tube is actually running much closer to 6.3V than any of the others; there's still a little but of noise from some cc resistors and whatnot, but that grating zzzzxzxxzxxzxzx is vanquished. I used UF4007's, I figure their 1A ratings are probably safe in a bridge that's only dishing out 300ma.

Whew, I think it's time for the celebratory Jameson on the rocks.

Cheers!

My stereo power amp runs 6sl7's for the PI's and the amp is dead quiet. Put those same tubes in my little tube tester guitar amp you get a quiet hum/buzz that isn't there when running 6eu7's or the 12a_7 family. The stereo amp has dc heaters and the guitar amp has ac heaters. May see if i can squeeze in a dc supply (as shown) for the octal socket in the guitar amp and quiet it down completely.

dave

[austinb]

Maybe the tube has a heater to cathode short.

[Structo]

I'll bet those tubes are not made like the more modern 12ax7 tubes that have the heaters cancel out the noise. Seems I read that they are counter wound elements or something that cancels out the hum.

[jjman]

I had similar problems with that tube type in a friend's amp. The solution was to change from Sovtek to NOS.

[Alexo]

When I was debugging this, I thought: Hm, the original Ampegs had two channels, I wonder if they wired the heaters out of phase so you could dial out the hum by setting both channels to the same volume...

Today, my friend who's borrowing his father-in-law's old B15 comes to me and tells me that the amp hums, but he can dial it out by turning up the volume on the channel he's not using.

...coincidence could just be bad filter caps.

Twisting and zip cord, just to be clear, I'm using "lamp wire," two 16 ga. leads fused alongside each other by the insulation. I have to think that this does cancel out emf and hum, as the magnetic fields generated by the heater current are as close together as they can possibly be, but I wonder if twisting them makes each lead act sort of like an RFI shield to the other, I believe the higher order noise I'm hearing, harmonics of 60hz, is probably more on the RFI scale than EMF. But I've consistently had this issue with 6SL7 amps regardless of what method I used on the heaters.

In any case, I greatly appreciate all the thoughtful responses and helpful suggestions.

[billyz]

I had to go thru quite few 6sl7 to find some really quiet ones for my B15n. But I finally did and it stayed quiet for years. I think they are very picky about that tube though. Ampeg must have recognized this because of the hum balance control and tying the filament to the bias circuit as well. I think you need to go thru a bunch of 6sl7's , fortunately I had a stash of NOS. I think I did try some Sovteks and they were noisey.