AC or DC for 300b Heaters

[sgerus]

Hello All,

I'm bulding my first amp and need advice on how to heat the 300B's
(and the input/driver tube as well)

For the 300B, should I:
1. use 5vac with the pot and Resistor/Cap on the fillement
2. use 5vac with the Resistor/Cap on the center tap of the 5vac transformer
3. use a solid-state 5DC -wlth slow startup (and the pot/Res/Cap setup)

same general question for the 6sn7
1. use6.3 vac
2. use 6.3 DC solid-state -wlth slow startup

[leadbelly]

Lots of good info here:

http://members.aol.com/sbench/humbal.html

[kmtang]

I would go for 5Vac for simplicity and good sounding.

The hum pot should be a shielded one. I used one with non metal case which cause a lot of buzz/hum noises. I found that out after replacing it with a shielded one lately.


Johnny

[kevinkr]

I guess the question to ask is how efficient are your speakers, how far down do they go and how much hum can you tolerate on the output of the amplifier as a consequence.

My speaker system efficiency is over 100dBsplw/m and the low frequency cut off is around 25Hz, (reflex loaded JBL D130's) and 2mV of hum on the output of an amplifier is quite audible.

I use constant current dc heating in my 300B SE amplifier and residual broadband noise is less than 400uVrms, and there is no discernable hum or ripple component in the output.

My 45 SE OTOH used AC heating, and had roughly 1.5mVrms broadband noise which was dominated by 60Hz hum, which was audible, but not annoyingly so from a meter away.

Some people have had good success with AC heating 300B's, but the best I have been able to do was about 5mVrms hum on the 8 ohm tap with a 300B, this is not acceptable with my speaker system, but might be with yours.

There is a raging argument about what sounds better, and I can't draw any specific conclusions - they all seem to work fine if well implemented. I have used both ac heating, constant current dc, and constant voltage dc with chokes - all sounded good to me.

[Tweeker]

Should not a constant current heater help extend tube life if nothing else?

[kevinkr]

Hi Tweeker,
Interesting that you mention that, yes in a sense it should greatly reduce the warm-up stress on the filament as the current is limited to the normal operating current. A cold filament depending on the source impedance of the supply might momentarily draw far more current, in 300B I have measured cold inrush currents >5A which was the limit of the supply doing the heating.

Another very important consideration which is problematic with constant current heating is possible cathode stripping and arcing if the plate voltage comes up to operating levels before the 300B filaments are warm.

I use CC heating in my 300B and it takes more than 30 seconds for the filaments to heat, which is also about the time that the power supply pass regulator tubes are warm enough to start passing current (they're 6550's) - it's kind of a guess as to whether the dht filaments are actually hot enough when plate current starts to flow and I am seriously considering a 45 second thermal delay relay on the B+ to the outputs. So far I have not had any reliability issues with the JJ 300B's I use.

Unrelated perhaps: I really like the JJ 300B's sonics, and they are rugged, well matched, and pretty consistent in my experience. They don't look as good as a WE300B, OTOH they cost me a 25% of what the WE cost.

[kevinkr]

One other consideration when using constant current heating with 300B, the filament currents from one brand to another aren't that consistent.

For example to properly heat a JJ 300B requires ~1.5A as evidenced by the voltage across the pins when fully warm, and a slightly annoyed gander at their data sheet confirmed this.

OTOH the WE300B really needs about 1.2A as do some of the Chinese Valve Arts types I have tried.

Boutique brands often require even more current than the JJ so I recommend making the constant current source adjustable - its quite easy and does not require the pot to handle any current. (I will eventually post the circuit to my website.) To set it t properly you must patiently measure the filament voltage as well as the current (measure current too - initially anyway) when the tube is fully warmed up.

edit to fix typo and minor additional clarification added

[sgerus]

Thank you to all that have replied so far.

My speakers are 98dB@1watt/1meter

Here is what I'm thinking:

Start with ac... with the current setting restistor on the center tap of the 5vac transformer.

If that results in too much hum...

Use 100R pot and 2 22R bypass resistors


If that still not good enough... DC

[JimW]

Kevin, does your implementation still use a humpot, perhaps to equalise bias current? I built a pair of 300b amps using dc on filaments. One amp uses a pair of bias resistors and caps (one on each side of the filament, the other uses a single cathode resistor and cap connected via a humpot used to balance the bias. I cannot hear any difference between the two arrangements, nor can I measure any difference in hum (both are less than 1 mv)
Jim

[Eli Duttman]

Segerus,

I agree with your decision to put the bias network on the 5 VAC winding's CT. There's minimal fuss that way and you may just be fine.

Your speakers are efficient enough to work WELL with a 2A3 based amp, which would cost you LESS to build. What made you choose the 300B?

An easy way to lower overall hum is to put DC on the heater of the 'SN7. If the idea appeals to you, I suggest you use either a 12SN7 or a 14N7 for voltage amplification. They are electrically equivalent to the 6SN7 and a 12 VDC heater supply is easier to execute than a 6 VDC supply is.