5f8-A build bias question.

[MikeW86]

Got a bias probe and checked the plate current. I am getting around 425.5V on the plates, and plate current on individual 6L6s was around 11.62mA. Does the current split between pairs of tubes in a quad ended amp, or is the bias really just that cold? For 60% dissipation i should be around 42mA.

I don't have a bias pot, so I swapped the 56K bias resistor for a 47K, and got around 15mA.

[Latole]

Bias probe should read right bias ma for each tubes = 60 % power dissipation.

425 volts for 6L6, your are ok with 42 ma on each tube.
Read minus voltage at pin 5 on each tube = not more than -35V IMO
 
From 56 k to 47 k and tubes stay very cold with 15 ma !!! Look something wrong. Or the 15k is short ? Or your tubes are very weak ?

You may have too much voltage from PT for bias circuit.  Is it the original PT ?
1- Check AC voltage at PT bias circuit
2- check DC voltage at diode secondary

[pdf64]

Bias probe should read right bias ma for each tubes = 60 % power dissipation
425 volts for 6L6 your are ok with 42 ma

Just to note that idle anode dissipation just needs to be somewhere in a fairly broad ‘Goldilocks zone’. The lower the HT voltage and the beefier the valve type, the wider that zone can be.
Pretty much any of the valve types we use will be well out of crossover distortion by 30mA.
 

From 56 k to 47 k and tubes stay very cold !!! Look something wrong.

Dunno, provided that the trimmer is changing the bias output, then perhaps the bias range resistor value just needs tweaking.

You may not have enough voltage from PT for bias circuit.

No, I think you may be in a muddle there; as bias voltage increases, anode current reduces.
If the bias supply voltage was insufficient, anode current would be too high.

[Latole]

 

You may not have enough voltage from PT for bias circuit.

You are late, I edit ; You may have too much voltage from PT for bias circuit.

[pdf64]

It’s a lot easier to tweak a voltage down that’s too high, than than it is to tweak up when it’s insufficient (voltage doubler to the rescue  ).

[tubeswell]

Does the current split between pairs of tubes in a quad ended amp, ...?

It's not necessarily an even split - whilst the current in each same-side pair is in parallel (and thus sums within the respective 1/2 of the OT primary), the current is each tube is dependent on that individual tube's bias point. (you can get unmatched tubes)

While some degree of mismatch isn't worth worrying about, it is better not to have too much mismatch. If the bias point is too different, it can cause the hotter tube in each same-side pair to suck the lion's share of current under heavy overdrive conditions. (bias-balancing circuitry is an option)

[MikeW86]

The tubes are pretty evenly matched. So I just add the mA of the two tubes that are paralleled off each side of the OT primary for the total for that side?

[sluckey]

Got a bias probe and checked the plate current.

Which bias probe? Need more info. Link, or pics, or describe.

[MikeW86]

This one.

[sluckey]

OK, that probe is only capable of measuring current for a single tube. The meter reading is for that one and only tube that's plugged into the socket. No sharing, no complicated math. The other tubes have no bearing on what your meter sees.

Yes, your tubes are biased really cold, about 5 watts idle.

[MikeW86]

Gotcha. I'll try a 33K and see what happens. Is there an easy formula for calculating bias resistor value or is it just trial and error?

[sluckey]

happy bias = 50K pot plus 22K resistor.   

[pdf64]

Given the valve type and the approx idle HT voltage, its triode anode charts can be used to derive the control grid voltage (ie bias) for the desired idle anode current.
As that will be for a bogey valve, the range of bias adjustment should be about +/-25% of that.

[tubeswell]

Gotcha. I'll try a 33K and see what happens. Is there an easy formula for calculating bias resistor value or is it just trial and error?

^what Steve said^


(A useful rule of thumb for a starting point is bias voltage output that covers a band around a midpoint that is about 1/10th of the B+ voltage. E.g. If B+ = 425, then aim for -42VDC bias voltage when the bias pot is at mid-rotation with +/- 10V (or =/- 15V). So design the bias voltage divider output accordingly)

[PRR]

...midpoint that is about 1/10th of the B+ voltage. ...

Yes, for most power tubes.

The underlying reason is the Mu of the tube and the G2 voltage. Since most audio power tubes have Mu near 10 and have G2 voltage nearly full B+, B+/10 is right. However EL84 has Mu nearer 18, so you see lesser biases on EL84 rigs. If you go trolling TV or RF types you may find lower or higher Mu or lower Vg2.

An "ideal" tube would cut-off with Vg1 at Vg2/Mu. If it did, that would be too cold. But cut-off is not a sudden thing and Vg2/Mu is often about right.

An old RCA hack says that 0.6*Vg2/Mu is a good hot centered bias. You'd never want to be that hot. In many over-volted stage amps, you can't run that hot or the Tolex would peel. So you want to design the bias network to never give less than this.

[MikeW86]

Well, I'm up to about 14W dissipation with 33mA and a 27K resistor, plate volts down to 396V. I'm going to swing by the electronics supply store on tuesday and get a 50K trimpot and put an end to this madness.