Starved-tube weirdness

[transient]

I've been experimenting with a single stage, cathode-biased circuit, using a 12ax7 at 12V.

I know that for high-voltage cathode-biased stages, you make the cathode positive in respect to the grid, and set the grid to ground potential. This makes the grid negative in respect to the cathode, which is required to get the tube working.

But by measuring the voltages on my circuit, i've found that the cathode is at 0V in respect to ground, while the grid is actually negative (around a half volts). The starved-tube article at Paia talks about this effect, so i guess it's not that strange.

Now about the weird stuff:
The circuit normally works and sounds fine. But when i add a cap to the input (for low freq filtering), it starts to sound seriously misbiased. And the cap value doesn't matter, it sounds misbiased as long as there's a cap there.

Now here's my (probably wrong) theory:

When there's no cap, the negative DC on the grid can travel all the way to the guitar pickup, making it negative also. And since both are at the same potential, it works as if both were at 0V, thus i get a biased sound.

But when the cap is there, it prevents the pickup from becoming negative. Thus it's similar to having positive DC on the guitar signal while the tube grid was set at 0V, it messes with the tube biasing.

What do you guys think?

...
emre

[A.S.P.]

10M grid>ground

[transient]

Quick reply, thanks But i do already have a 10M resistor at the grid.

I noticed that the bigger the resistor, the more negative the grid becomes.

.
e

[GFR]

Perhaps it's grid blocking - when the grid is driven positive by the input signal, the cap gets charged and then at the next cycle the grid will be more negative and cut off.

You can try a smaller grid resistor so that the current through the resistor will help discharging the cap at the negative semicycle; you can try a diode from ground to the grid so that the cap will be discharged by the diode at the negative semicycle;

[transient]

Thanks GFR

I'll try them and report back.

...
emre

[transient]

I tried adding a diode, it works Thanks again GFR.

Would the diode type make any difference, i'm using a 1N914?

.
e

[luix]

Hi transient,

I had the same experience vith my starve tube distortion, the grid is negative in respect to ground!!!!

I noticed also that the grid resistance to ground affect the anode bias, for example if I use Rg=1M the anode can't polarize at 1/2Vcc (6V), to polarize at 6V the grid must be around 180k (100k 120k 150k).

Moreover I found that a 12AU7 sound better than a 12AX7 for low voltage B+, more distortion so I think more gain  maybe for the low internal resistance.... I will try to use a ECC86 that is a low voltage tube.....

[transient]

Hi Luix,

I'm guessing that you're using the JFET biasing approach, using a 100K trimpot. But as far as i know, you don't need to have half the voltage at the plate. So you can just place a 100k resistor there, not worrying about having 1/2V at the plate. A higher value (like 250K) will give more gain.

And if you're using this tube stage as the *first* thing that's plugged to your guitar, i don't think you should go lower than 1M for the grid resistor. It will load down the guitar signal, resulting in "tone-sucking". But if there's a buffer before the tube stage, then you can go lower.

Take a look at this page for information about tube biasing. It talks about high-voltage operation though, where you pick a value for the plate resistor and calculate the required cathode resistor value according to that. Starved-operation might require a different approach, but we at least can use it as a starting point.

...
emre

[luix]

Yes I know infact this is my circuit:

[transient]

Wow, you've got a pretty advanced design there

I see that you're not using cathode-biasing, that explains why you needed 6V at the plates.

So, how does it sound?

.
e

[luix]

Here are the sample: click here

3CB mean the "valvenstein" (it's name) with the marshall style tone controll, 1CB mean the version with big muff style tone controll...

[wampcat1]

Here are the sample: click here

3CB mean the "valvenstein" (it's name) with the marshall style tone controll, 1CB mean the version with big muff style tone controll...

Here is the correct url:
http://www.diyitalia.com/index/elettronica/effetti/CAMPIONI%20EFFETTI.htm


I like this clip: http://www.diyitalia.com/index/elettronica/effetti/campioni/Valvenstein%20I%20clip%201.mp3
A bit more highs would sound good, IMO, but it could just be recording setup and/or amp used. Clipping sounds really good, there is a bit of an 'odd' attack on some of the other clips it sounds like. I didn't care for the big muff tone control clips.
My opinion, anyway!

[luix]

What do you mean for

there is a bit of an 'odd' attack on some of the other clips it sounds like

sorry for my bad english...

[wampcat1]

What do you mean for

there is a bit of an 'odd' attack on some of the other clips it sounds like

sorry for my bad english...

Not a problem - remember, this is just my opinion! 

The 'attack' is when you first hit a note on your strings, it sounds a bit oversaturated. Just my opinion!

Take care,
Brian

[luix]

remember, this is just my opinion

don't worry, I can't understand very well tecnical english so...

Is a characteristic of this circuit, I don't know why, is a surprise for me and for my beta-tester... I'm not very good also with guitar...

The instrumentation is a small bass amp of a brand called Rogers, the pedal was in the return so I use only the final of this amp (TDA2030 with two power transistor like the data sheet), the microphone was a sennheiser for voice connected to the sound-card of my pc, a sound blaster PCI 128, the guitar was a stagg with original (and awful) pickup...

[transient]

Yes, starved tubes sound a bit weird when heavily overdriven. Maybe that's why most tube pedals don't let the tube distort and just use it for coloration.

Thanks for the clips, BTW

.
e

[luix]

Maybe that's why most tube pedals don't let the tube distort and just use it for coloration

Infact my idea was to use the OPAMP to make a clean booster that saturate the 12AX7, other pedal use the OPAMP to distort the sound and the tube to add some tubish sound...

[GFR]

I tried adding a diode, it works Thanks again GFR.

Would the diode type make any difference, i'm using a 1N914?

.
e

Try a schottky diode if you can.