How low is too low (heaters)?

[EFK]

This amp is a 2XEL34 amp and the PT can be set for 110, 120, 200 etc. If I set it for 120 VAC, and measuring 120.5 out of the wall, the power tube plates are @ 435 DC and the heaters are at 5.7 V AC. Isn't this *too* low? I know there's leeway, but that seems really low. If I set the PT for 110 and use it w/ the same 120 out of the wall, the heaters jump up to 6.3 VAC but the el34 plates then jump (after rebiasing) to 481. I actually could live with this, but I know there are times when the wall voltage is going to jump up to 123-125 and at that point I think I'd be pushing into the 500 VDC zone on the plates, and the filter caps are rated at 500 which I think is pushing my luck.

What to do? I suppose I could use a bridge rectifier and big cap to run the filament AC to DC, which usually bumps it up about .5 to .7 volts once it goes through the diodes, which would put it where it ought to be and probably be quieter to boot. I did this once before with a new transformer, but this PT is 40 years old and I don't know if the bridge rectifier and the typical 15000 uF cap stress the PT in any way?

Any thoughts?

[Chuck H]

The filament requirements are rated at +/- 10%. 10% of 6.3 is .63. 6.3-.63 is 5.67 so your right there on the line. And since, as you say, the wall voltage tends to go up rather than down your likely to stay within spec. But...

Maybe I'm blowing smoke up my own a$$ but I've noticed that tubes with the filaments on the low side seem to sound kind of grungy and less dynamic. I can't back it up with any testing though. Just an observation from personal experience.

Why is it low? You say it's an old PT so I'm guessing salvage? Is it possibly meant for a pair of 6L6's which draw less filament current and is low because it's stressed? This would be my concearn. Most PT's filament winding is ample for the task and more often than not you can make a 6L6 to EL34 change without issue. But sometimes the spec is close and the voltage gets dragged down with the change. Also...

IIRC changing to a DC supply will reduce the current capacity of the supply proportionate to the voltage increase through the rectifier. So you could actually be doing more harm than good if the wind is already over capacity.

Have you noticed any funny smell (like a hot transformer) or the PT getting hot? Have you checked the voltage with the tubes pulled?

[EFK]

This is and old Selmer TnB50. I have quite a few of them as for years they were very cheap. The transformers look like new, but they are original to the amp and the PT is the typical Selmer PT they used on the MKIII amps. This one seems a hair weaker than others, but otherwise everything seems all right. It doesn't seem to get hot. Good advice, I think I will yank all the tubes and measure. I'd like to see if something is pulling it way down.

[Steve Conner]

The Selmers were designed for the British market, the 120V tap could be a bit of an afterthought.

435 is a bit low for the plates too.

DC heaters will stress the heater winding more than running the same tube complement directly, because of the high RMS-to-average ratio of the current in the winding.

Edit: Maybe changing from EL34s to 6L6s would help.

[EFK]

Edit: Maybe changing from EL34s to 6L6s would help.

In a British amp? The horror!!!! ( just kidding! )

I have found those voltage selectors to be all over the place in voltage, so I suspect you are right - it's likely an afterthought and not terribly accurate. The schematic design voltage for the plates on a MkII was 460V. They didn't stick those selectors on until the MkIII models, but those usually do run the B+ a little lower despite the diodes. I'll have to think on this a bit.

[R.G.]

5.7 is fine for preamp tubes, maybe even giving a softer distortion and longer life. Power tubes need all the heat they can get to get enough emission. Low plate voltage just lowers output power a bit. Generally not a big deal.

[Amp Kat]

Many design circuits to keep the heaters around where yours are to save their tubes a bit and even use ramping circuits to turn them on slower. Also depends on what the actual wall voltage is putting out and have seen swings of anywhere from 6.3 to 7.5 VAC depending on if the AC outlet is around 122 or 128 VAC and they vary from state to state.

[MWJB]

I'm with Chuck in that low heater voltages makes amps less dynamic & grungy...if your heaters are under 6.2VAC you should do something about it. PT's are usually designed to run the heaters high with no load, so that they still drop to above 6.3VAC with the full tube compliment & current draw.

I would look for a fault if I found a tube amp with just 5.7VAC on the 6.3VAC heaters, your amp will not sound right. I'm not one to quote data sheets normally (hey, it just happens to suit my purposes this time), but they nearly all give operating perameters of "6.3 to 6.9 VAC". Vintage Fenders typically run up to 7VAC and do not display signs of unreliability with regards to heaters.

Also, as Steve says, I would expect a higher B+ on a T'n'B...say 470vdc +/- 10v? You normally need to rebias.

[Chuck H]

You could always use the primary that gives you the heater voltage you need and use a 30V zener string on the HV wind CT to drop it down.

[EFK]

Well since I generally tweak these MKIII amps back closer to a MKII (not many changes, pretty easy save the tube rectifier), I set the PT at the 110 setting, which is putting the B+ off the PT in the 480s. Since I was in there, I replaced the old original BY128 diodes dangling off the PT just for the hell of it (I guess 40 yr. old diodes are fine) with a few higher rated diodes and I'm almost embarrassed to say it but as a quick and dirty way to drop a few volts I stuck a 100 ohm 25W resistor inline between the diodes and the first filter. At idle it drops about 7-8 volts, putting the tube plates where they ought to be, and I guess you could say I'm semi-emulating a GZ34 anyway. I'm sure it drops a little more when going full tilt. So everything seems just peachy now, all is well in heater and voltage land. For the moment. I think the zener would be a more 'elegant' solution but I don't have one.