Question about 12AX7 parallel - bias

[bluesbear]

I read over and over about many of you straping across a 12AX7... running the tube in parallel. I'd like to try that but I've never understood what affect this has on the bias resistor. It seems as if you'd need to halve the resistor to keep the bias level. For example, if you're running 1/2 a 12AX7 with the standard 1.5k cathode resistor, wouldn't adding the extra half tube mean the resistor should be changed to 820 ohm... or should each side have it's own 1.5k to keep the bias correct? Am I looking at this correctly?
Help a poor confused person, please!
Thanks,
Dave

[HotBluePlates]

Yes, you are right.

Most folks don't bother with that. If the paralleled stage is the input stage, the incoming signal is small enough that the change of bias is unlikely to materially affect stage gain, distortion or headroom.

As always, experiment to see what works best for you.

[jim]

Checkout the gain stage for the Carolina Blues Special.   Its parallel  input stage has much more gain than any Fender gain stage and is very full sounding and pick sensitive.  Hot Plate is right--just about any bias arrangement will work--the tube is one huge maxed out triode next to the small guitar signal.

[tubeswell]

Yes ohm's law states that Voltage = Current x Resistance.  So if you double the current (as in two parallel stages), then you need to halve the resistance in order to keep the voltage the same (if that is your goal).

Similarly, if you parallel two stages without halving the load, then there will be a bigger voltage drop across the plate resistor, and you may want to increase the bias resistance in order to keep the bias voltage the same (if you want 'centre biasing').  That is why some fender pre-amp stages with a 220k plate load have a 2k7 bas resistor.

However, as the others have said, you can bias the stages hotter or colder for different effects. Hotter biasing will tend to result in clipping from grid-current limiting. Colder biasing will tend to result in clipping from plate cutoff.

[HotBluePlates]

However, as the others have said, you can bias the stages hotter or colder for different effects. Hotter biasing will tend to result in clipping from grid-current limiting. Colder biasing will tend to result in clipping from plate cutoff.

What I meant was, while the above is true, in an input stage, the guitar signal is well under 0.5v peak-to-peak. Unless you get to the very extremes of bias, this input signal level will not cause any signifigant distortion in the input stage.

Distortion of the second stage, given paralleled input and proper biasing for maximum output swing (along with minimal loss between 1st and second stages) is another matter entirely.

[bluesbear]

I guess a little more clarity is in order. This is an AB763 preamp with a modified Hot Switch after the tone recovery stage. It's the Hot Switch I would like to try it on... just to see how I like it. I think I'll have to jumper in different resistors to see. I was just hoping for a starting place.
Thanks,
Dave

[jojokeo]

Simpy start off with half the values for what you have that's using a single triode. If it's a 100k/1.5k then go with 47k or 56k load resistor & 820r

[Fresh_Start]

Simpy start off with half the values for what you have that's using a single triode. If it's a 100k/1.5k then go with 47k or 56k load resistor & 820r

I also was thinking that you "should" halve the plate resistor as well as the cathode resistor to keep everything more or less equal.

Please keep in mind that parallel triodes do not have to share a common cathode resistor & bypass cap.  You could have the second "boost" triode in the circuit all the time but with a big, un-bypassed cathode resistor.  Then your switch could reduce the effective size of the cathode resistor and add a relatively small bypass cap (2.2uf or less) if you want to brighten up the boost.

Parallel triodes do need to share a common plate resistor.  Unless you choose to have a completely separate triode with its own post-plate coupling cap.  That would let you simply ground the grid of the "boost" triode when you want it off, leaving you with a completely stock circuit until you engage the boost.

Hope that helps,

Chip

[rafe]

I have been tempted to do this with a 65 supro thunderbolt...I hate to see 1/2 a tube wasted ....but what can I expect in return? Is it going to be a noticeable improvement or is it negligible? Is it not worth messing with on a vintage amp?

[tubeswell]

I have been tempted to do this with a 65 supro thunderbolt...I hate to see 1/2 a tube wasted ....but what can I expect in return? Is it going to be a noticeable improvement or is it negligible? Is it not worth messing with on a vintage amp?

I have done a switchable single/parallel v1 stages on a couple of amps now. see "6H3" schematic. There is a noticeable increase in gain from the paralleled triode mode.

[HotBluePlates]

I wouldn't halve the plate resistor, or the increased current due to the second triode will result in no increase of output signal.

[Fresh_Start]

I wouldn't halve the plate resistor, or the increased current due to the second triode will result in no increase of output signal.

OK - I'll buy that.  Never seen a circuit with 1/2 value plate resistor value anyway. 

What does the parallel triode produce that you couldn't get by doubling the value of the plate resistor on a single triode?  IOW 100K plate resistor shared by two triodes versus a 200K plate resistor on just one of them?

Always trying to learn.

Respectfully,

Chip

[38Super]

For 12ax7 triode gain stage, typical Fender values are 1500 Ohm cathode R and 100K Ohm plate resistor.  Let's say it's 1mA and plate supply is 250V.  So, Vk=1.5V and Vp=150V.

If you parallel two triode stages, halve Rk, but don't halve Rp, you'll have 2mA total bias current (assuming similar transconductance between V1a and V1b) flowing through a 750 Ohm Rk.  So, Vk=1.5V; however, now Vp is (250-2mA*100K)=50V.  So you have dramatically shifted the quiescent plate voltage and with this asymmetric bias point, you'll have less headroom in this gain stage and a different 'feel' in the stage because of reduction in Vp.

Maybe you like this. 

Paralleling triode stages effectively doubles the transconductance of the stage - you get more stage voltage gain, but there are many factors that affect gain.  Consider two separate gain stages, biased the same with same load resistance...  You capacitor couple them into a summing network and end up with two times the signal, right?

rob

[jojokeo]

You get a slightly higher gain paralleling triodes and better signal : noise ratio. Halve the shared plate and cathode resistors for a starting point. This is what was asked in the intial thread..."where to start". The alternative ideas have been shared - awesome.

Overanalyzing everything isn't necessary. Install the values in the amp, turn it on, and play. Very simple and not wasting time contemplating everything. Over analysis leads to paralysis. Get off the computer, heat up the iron, install, and play. Repeat if necessary. Get back and report on results, findings, and likings. LOL

[HotBluePlates]

Get off the computer, heat up the iron, install, and play. Repeat if necessary. Get back and report on results, findings, and likings. LOL

  Now THAT'S good advice!

[sluckey]

Amen. You don't always need a 5 day discussion to try a 15 minute trick.