25k Bias pot not enough range??

[RockinRocket]

I seem to be running out of room with my 25k Piher bias trimmers in my Plexi and Wreck builds.

Typical 430v and 400v respectively on the plates. Same power amps with the 47k and a 25k Piher trimmer with the Wiper and other end ties to ground.

I cant get all my tubes to bias colder with a 1ohm resistor at ground pin 1/8. Both pair of Mullards are at 46ma 80%. JJ down to 35ma 60% on the Plexi/430v. I can get a slightly lower reading on the Wreck due to 30 loss volts.


Seems odd to me.. Everyone uses 25k in these builds.
what am I doing wrong?

[dbeasley]

You might need to tweak the 47K resistor to accommodate a different bias range with the pot.

Go higher in value let's you bias colder, go lower in value let's you bias hotter.

Sent from my Moto G (4) using Tapatalk

[R.G.]

The short answer is that there are four resistors, one capacitor, and one AC supply voltage that determine bias, and one (or more) of them has varied in a way that makes your 25k pot less effective.

The bias voltage is being generated by connecting a high-value resistor from one end of the high voltage secondary on the PT. It runs through a diode then to the bias cap to ground. The bias cap is loaded by the resistor divider string that makes up the bias adjuster, plus whatever grid current happens to flow in the output tubes. The diode is set up to make the voltage on the bias cap negative.

The high value resistor from the B+ AC source determines how much charge can get into the bias filter cap each half cycle. The three resistors after the bias filter cap determine how much of this charge is lost in the 16mS between AC cycle rechargings of the bias filter cap, which means that the voltage on the bias filter cap depends on the values of the three "load" resistors as well.

The bias filter cap could be a culprit, but only if it's too small. As you increase the value of the bias filter cap, beyond a certain capacitance, its value stops changing the raw bias voltage much.

The bias voltage that gets out to the tubes is the raw bias voltage on the bias filter cap as diminished by the three-resistor divider of the nominal 15K, the 47K, and the variable resistance of the 25K pot. The pot can be between 0 and 25K, so the resistor divider is between
(47K+25K)/(15K +47K+25K) = 0.83 of the raw bias voltage and 47K/(15K+47K) = 0.76 of the raw bias voltage. The added loading of the 25K pot being set to zero ohms also makes the bias voltage a bit smaller in magnitude, so this end of things is an even colder bias than would otherwise happen.

With what affects what in mind, we can start looking at what might be getting you.
If the total value across the 25K pot is too small, it can't get big enough to let the divider on the raw bias voltage be big enough (in the more-negative is bigger sense) to bias your tubes cold enough.

If the "47K" resistor is actually too small, again the loading of the 47K plus maximum value of the 25K pot can't get big enough to make the bias to the tubes big enough.

If the 15K resistor is too big, the bias voltage can't get big enough, again because of its effect on the voltage divider.

If the "330K" resistor is actually too big, your overall raw bias voltage is too small (not negative enough) and again you can't bias the tubes cold enough.

Finally, and most unlikely, if the bias filter cap is way, way too small, again the overall bias voltage won't be big enough (negative enough) to turn the tubes off enough.

How do you tell? If I were sitting at the bench, I would measure the raw bias voltage. If that's not at least -55V to -60V, the 25K pot can't bias the output cold enough even if the bias pot itself is OK. If the raw bias voltage isn't negative enough, decrease the value of the "330K" resistor from the high voltage secondary a bit. You might parallel it with a 3.3M, 2.2M, or even a 1M resistor to let a bit more current through. This will shift the whole bias voltage to be more negative.

If the raw bias voltage is big enough, I would then measure the actual resistances of the "15K", "47K" and "25K" pot. If the 15K is actually too high or the other two are too small, fix that. If all these check out, I'd worry about the raw bias cap.

As a final note, I disagree with single bias adjustments per output stage. I know it's not "original", but I personally would change the 47K/25K pot setup to two 100Ks and two 50Ks, one per output tube. But that's just me.

[pompeiisneaks]

R.G. Stop doing all this amazing stuff, you're going to teach me a thing or two!

Thanks again for an outstandingly well written explanation. I think we may need a new sticky just called R.G. said ;D

~Phil

[R.G.]

That might be a mistake. I also have a large assortment of limericks memorized.

[pompeiisneaks]

There once was a man from nantucket... wait back to the thread.



~Phil

[R.G.]

Said the plumber, still plumbing, it's me.

[RockinRocket]

Yes RG you should write a book

I just dont get how every one uses these typiacl 25k and 47k reistor and are actually able to have a useable range??

Should I just use a 50k bias pot? I like to stay original to the clones I build but hate to have to use a 56k or 68k Iskra in the Bias.

I feel like I did some thing wrong.. trimer is just about all the way clockwise for a 35MA ish bias..

[M Fowler]

I see said the blind man!

[dbeasley]

I love RG's exposition.

Sent from my Moto G (4) using Tapatalk

[R.G.]

Yes RG you should write a book

I have - just not that one.
I did a book on how to lay out PCBs for making your own effects, and - well, actually, it's a whole string of books or booklets, depending on how you look at it. I did "The Vox Owner's Safety Net", which is a technical stroll through the Thomas organ Vox amps and how to fix them, then did a family of booklets, one per model, on the details about that specific model, which are enough alike to be confusing, but not identical.

Besides, this is all really just applying Ohm's Law in perverse ways.

I just dont get how every one uses these typiacl 25k and 47k reistor and are actually able to have a useable range??

Fact is, it works mostly. Something you learn about making user-adjustable controls is that you have to make the range of the control not be a disaster if it fails. Most users will be careful with a control, but some of them use vice grips or hammers to tweak a setting, or step on the knob. And if you're not careful, you can make a control that causes smoke, flames, or other unwanted excitement if the part fails or if the adjustment range is too wide.

Those are possible issues with bias pots. If you just use the pot wiper to run to the grids, and the wiper goes open, then the grids have no bias voltage and you get overheating, possible smoke, blowing fuses, other nasty stuff. So the Marshall circuit properly rigs the pot so that if the wiper goes open, the bias just reverts to the coldest possible bias. You really need that.

Marshall also did a reasonable job of limiting the control range, so that a non-discriminating adjuster didn't have enough pot range to make things really burn up. The pot range of full on to full off only lets you set a bias voltage of 0.76 to 0.83 times whatever the raw bias voltage is. That's reasonable. I think that in this case, something else was making the raw bias voltage too small (that is, not negative enough) so that even 0.83 times whatever was there was not enough to get a really cold bias. Marshall could probably be forgiven for not designing enough range into their amps for amp builders fifty years after the amps were manufactured.

But a little more range would be OK. This thing works by adjustably loading down the current through the 15K resistor. You can diddle the range by making the "15K" bigger or smaller. That changes both the end points of the adjustment and the range of the adjuster. A little work setting up the Ohm's Law stuff in a spreadsheet will let you tinker it in with stock resistor values without even solving simultaneous equations.

Should I just use a 50k bias pot? I like to stay original to the clones I build but hate to have to use a 56k or 68k Iskra in the Bias.

I'm a bad guy to ask that. My personal opinions lean toward making it work no matter how "original" it is. I don't hold a pursuit of originality against someone who knows what limits they're putting on themselves and does it willingly, but for me, correct operation trumps "original".

What to do with this circuit is pretty open. If you change the "15K" resistor to higher values, the bias to the grids is made lower, and the adjustment range is wider. If you make the "15K" smaller, the bias to the grids is both more negative/colder and has a smaller range. If you make the "330K" larger, the raw bias voltage is smaller, moving the bias voltage to the grids and the range to less negative/hotter ranges. Smaller "330K" values do the opposite, just moving it up a bit. Changing the total of the 47K +25K against the 15K also affects the range. Trading off some resistance in the "47K" against the "25K" pot gives you more range. If you made the "47K" be 27K and the pot be 50K, you get a wider range mostly on the hotter bias side. If you make the 47K be 68K and the pot be 10K, you get a smaller range in a much colder are of the possible bias voltages.

So you can go any way you want - but you have to guess which way you want to go.

I feel like I did some thing wrong.. trimer is just about all the way clockwise for a 35MA ish bias..

I don't think you did anything wrong. I think the tolerances on the parts you're using just added up in the wrong way.

My first advice would be to parallel a resistor with that 330K and see if that cools off the bias. It should. I pick that resistor because you are worried about too-hot bias, so an expedient fix would be to move everything to a cooler (more negative) bias range. It's easy to parallel a resistor across one that's already there. The suggestion of a 3.3M as a starter was a valid one. That lets 10% more current through, and ought to make the raw bias voltage be more negative by a few percent. As a bonus, the added resistor only needs to be rated at 1/10 the power of the 330K it's paralleling, as it carries only 10% of the current. You could even think of this as "330K" resistor with a -10% tolerance. When you've done that and it works or not, you can decide what to do for the actual production unit you ship. You could hunt around for an extra-low value of a "330K" nominal resistor, or decide what to do about messing with the other resistors. But if you can fix it with one resistor, all the better.

And in the end, you can cheat. Who says you can't paint different color bands on whatever resistor you leave in there?

[Reeltarded]

I agree with RG in perpetuity because he's a freaking genius and he's always right and so am I for agreeing.

[roberto]

Go higher in value let's you bias colder, go lower in value let's you bias hotter.

Just one point on this: with negative numbers, going lower means increase its module, so the bias current will be lower, and bias colder. EG -45 is lower than -40. So the results are the opposite of the quoted statement: Lower voltage on g1 means colder bias and vice versa.

[pompeiisneaks]

I also agree with R.G. on the whole 'making it work' vs 'being more original' I've built quite a few clones now, but almost always I deviate a tiny bit or a lot and still get great sounding amps out of them (to my ears anyway). I think there's plenty of room in the world for both methodologies. Those that love the painstaking process of exacting replicas, and those that just like to build one that does what they want and gets their own personal touch .

I'd say adding a jumper from the wiper to the other end of the pot helps ensure that if the wiper does lose contact you have maximum resistance and thus guarantee a bit colder bias if it does lose continuity, Otherwise, I personally had to adjust my bias resistor outside of the pot because the dumble expected value didn't work. If I recall it was like 24k, and all I ahd was 33k or 21k, I think I ended up with 21k and it was perfect for the tubes I had.

~Phil

[RockinRocket]

Thanks for the insights fellas.
I guess from now on I will just use 50k bias trimmers so I can use a wide variety of El34s in my builds.