6.3 Volt Heater Confusion

[codamedia]

I am aware that to reduce hum, it is recommended to add 2 x 100ohm resistors (1 on each lead) of the 6.3 line to ground, usually done at the lamp.

When I tried to do this, it immediately blew the 1 amp fuse. I removed the resistors, and installed a new "1 amp" fuse and made sure everything was working properly again – which it was.

When I checked the heater lines more closely I saw something different on this amp. At the end of the chain, just before the last tube (the tremolo tube), one of the lines connected directly to ground. From ground it went to its filament connection on the tube, then to the centre pin of the tube. There were no resistors or caps in this chain. The other 6.3 line went straight to its tube connection.

I found this odd, as all my other tube amps simply connect the 6.3 lines to the tube filaments. There is never a connection to ground unless done through the 100ohm resistors.

1: Has anyone ever seen this, and can they explain it to me?
2: Would this connection to ground be the reason the fuse whenI tried to install the 100ohm resistors?
3: Is it safe to remove that entire ground connection, and simply connect the second 6.3 line directly to the tube as it normally would be.
4: If 3 is safe to do – can should I then be able to install the 100ohm resistors?
5: Should I just leave well enough alone – LOL!

Thanks,
Sean

[jonnyeye]

The heaters (which are powered by AC) develop a small region of charge which can be coupled to the cathode causing hum, and this effect is reduced by referencing them to some DC voltage instead of leaving them floating. In theory the best hum reduction short of using DC to power the filaments is to reference to some positive voltage, by taking a voltage divider off the B+ to give a stable 30-50V (someone, maybe RCA or GE did a bunch of research in to this). If you're going to do this, you should put a biggish cap (100uf or so, rated for at least the reference voltage) from the reference point to ground, to decouple the filament AC from the power supply. Like this: (resize the voltage divider to provide 30-50V at the reference point, and tie it to either side of the filament winding.)

More commonly the reference is provided by grounding the center tap of the filament winding; if there is no such center tap, one can be created artificially using two 100 ohm resistors. A method found in cheaper amps (for example, my own Beltone AP-14) with no center tap on the filament winding was to simply ground one side of the filaments. Both methods accomplish approximately the same thing, but the center tap version is claimed to have less hum as the filament voltage is centered around zero volts. (I've never tried both versions in one amp, so I can't comment.)

(Really cheap amps ground one end of the filament winding at the transformer and run just one wire for the filaments, grounding the other side at the sockets! You end up with big AC currents running through the chassis, which is also used for signal grounds! Nasty.)

If you aren't having hum problems, I'd just leave it alone for now. If you are, can you remove the ground connection? Yes - if you want to reference the filaments to any other voltage, you must! You are shorting out something if you don't, and that's what'll cause your fuse to blow! Be sure to leave the center of the socket connected to ground (especially if it's being used as a tie point), then add your 100 ohm resistors. If you're still having hum problems, then you can try the reference above - you will have to lift the center of the 100 ohm resistors from ground, and you can use that as the tie point to the reference voltage. Still, if this was made as a cheap amp, there may be other things you can do to reduce/eliminate hum - my Beltone is quiet, and all it took was a new set of tubes! Read up on hum sources at R.G. Keen's excellent site (there's a ton of other great amp information if you poke around!) - start with the star grounding page at http://www.geofex.com/Article_Folders/stargnd/stargnd.htm

Hope some of this helps - sometimes I just like typing it seems...

[codamedia]

Thanks you for that very detailed comment. Your description of your AP-14 seems to be exactly what my AP-22 has. One side of the filament is tied to ground. This confused me as I have not seen that on my other amps (BF Fender, Hiwatt, and MusicMan).

This amp is relatively quiet, and really sounds great. I am in the middle of doing an extensive rebuild on it. It's got new tubes, new PT (the old one had a great smell when it got hot, but it got way to hot for me to trust it even if the readings were good - LOL), new filter caps, coupling caps and a few resistors.

The heater lines are a mess on this amp though (I think someone did a horrible job of putting in new tube sockets sometime in the past) and I want to clean it up. That is likely going to be my last stage of the rebuild.

Your post has been a great help - thanks!

[cartoonweirdo]

A great trick for new heater leads is to take long (6-10ft) pieces of wire (yellow and green), clamp the ends in a power drill or screw gun, loosely hold the other side and gently pull the trigger. Winds them quite tight and seems to give the overall assembly enough rigidity to turn corners more sharply. When you put them in, run them well away from everything you can, usualy along the back of the chasis.

Carl

[jonnyeye]

I just came across this document today which explains why creating a virtual centre tap for the filament reduces hum.

http://greygum.net/files/radiotronics/hum.pdf (warning - math heavy, although the main points can be grasped through the pictures and the accompanying text)

Reading closely nets the following points:

- Hum is more likely caused by capacitive coupling of the heater and the grid (than the cathode), since the cathode is grounded through a small resistance;

- Hum can be lowered by decreasing Rg, and by reducing stray capacitance between the grid and heater connections;

- One side of the heater winding will reduce hum more than the other, but not necessarily by much;

- The article finds hum to be about 2mV for grounding one side of the heater for a "typical" tube; 2mV is only 34dB down from a "typical" 100mV guitar signal. That said, I don't know for sure if 1pF is a good approximation of heater-to-grid capacitance, as no datasheet I have looked at (with the intent of finding this information) lists it. A typical input (grid-to-cathode) capacitance for small tubes is between 1-2pF, so expect the heater-to-grid capacitance to be somewhat less than that.

- Hum can be eliminated better by forming a virtual centre tap for the heaters, as this causes phase cancellation of the AC hum at the grid (see figure 4 in the article). The article explains that since the two capacitances are almost equal, an equal divider will work fine in most cases.

And much more! Now go squash some hum (at least from one source...)!

[cartoonweirdo]

if unequal grid to heater capacitances are worrying you, why don't you ground the wiper of a two turn 250 ohm pot, put the heater lines on the two outside terminals and adjust it till you're satisfied...

If you want to go to just a hair more trouble get a pre packaged full wave bridge rectifier (the little square ones), some caps, and fire the whole thing on DC.

[dhuebert]

A trick I learned from Gar Gillies is to form a virtual center tap with 0.1 uF caps. Tie a cap from each side of the filament wires to chassis ground as close to the power supply ground as you can. Seems to work and 0.1 uF ceramics are cheap and plentiful. And don't let the audio phools convince you anything more expensive is needed here.

Don

[nigelwright7557]

In my designs I usually go for DC heater supply and that gets rid of hum.

To add to previous replies, some dual valves have a noisy half and a quieter half. I put the quieter half on the input side to reduce noise.