Grid current characteristics of 12AX7

[Alan0354]

Anyone has graph or article on the grid current of 12AX7 with different grid voltages and plate voltages?

I read in chapter 2 of RDH4, it gave all the contribution and graphs of a tube, but I want to know the real number of 12AX7.

[jazbo8]

Here you go, please note this is from a single sample and un-likely to match those tubes in your collection.

[Alan0354]

Here you go, please note this is from a single sample and un-likely to match those tubes in your collection.

Thanks

But I need the more detail info. This is because there are different small sources in uA or even lower range from gas ionization, leakage between plate and cathode, and emission current. That's the numbers I am looking for. obviously those are quite significant as people use it to set up bias for grid leak bias. It must be close to 1uA or so.

[jazbo8]

have not seen one yet, let me know when you find it...

[Alan0354]

have not seen one yet, let me know when you find it...

I know, I have been digging already. This is not important on normal operation. But it is vital for noise calculation if people work on low noise design. This produces shot noise current and create noise voltage with the input resistance. This can turn out to be a major noise contribution in low noise design. that's why people use either BJT or MOSFET depends on the input impedance. People put up with MOSFET and it's high 1/f( flicker noise) because it does not draw gate current. If you have a high input impedance, the MOSFET is a big win.

From what I understand, the grid leakage and ionization current can add up to uA and that is a lot of noise current.

[jazbo8]

I think the data you are looking for may be contained in Merlin's recent paper, but you will have to pay to get the paper or head on over to your local university library to see if they have a copy. But the following quote from Merlin provides an indication of the triode flicker noise, with the caveat that every tube (type) will be different, since there is no known good noise model for the triode at audio-frequencies in the present time:

"1uV is a good ballpark figure to keep in your head, though high-gm types can consistently get below 0.5uV." - Merlinb from diyaudio

[Alan0354]

I think the data you are looking for may be contained in Merlin's recent paper, but you will have to pay to get the paper or head on over to your local university library to see if they have a copy. But the following quote from Merlin provides an indication of the triode flicker noise, with the caveat that every tube (type) will be different, since there is no known good noise model for the triode at audio-frequencies in the present time:

"1uV is a good ballpark figure to keep in your head, though high-gm types can consistently get below 0.5uV." - Merlinb from diyaudio

You are of big help. 1uA is my wild guess from the grid leak article. I did use that for noise calculation in the other thread. I was showing the noise from the grid stop resistor and the grid current is 12uV peak to peak for single triode and 14uV for parallel. That's not trivial at all. But still need to calculate the other contribution to complete the picture.

Also, I can't thank you enough for finding the FP PI from RCA, that confirm my assertion. Before, it was just my guess only as I don't have the long running knowledge like you, JMF and Chuck H etal. I absolutely don't have the historical knowledge of tube amps.

Many thanks.

[jazbo8]

You are quite welcome! Nothing beats going through the numbers as you are fond of doing, but for guitar amps like many before me, I have found that good hi-fi and/or engineering practices do not necessarily translate into good sound/tone, so often "the theory" has to take a backseat to what actually works in the field. Nonetheless, for geeks like us, the number crunching is good to keep our noodle working a bit harder.

[Alan0354]

You are quite welcome! Nothing beats going through the numbers as you are fond of doing, but for guitar amps like many before me, I have found that good hi-fi and/or engineering practices do not necessarily translate into good sound/tone, so often "the theory" has to take a backseat to what actually works in the field. Nonetheless, for geeks like us, the number crunching is good to keep our noodle working a bit harder.

The long post on the noise has nothing to do with the sound quality. Those are try and true to find the right value, there is no theory for that. I don't even touch that part. This is just limited to the noise........hissing and popping sound from a high gain front end. That you need theory to do. After that, then it is all touchy feeley.

It's like you need theory also for ground loop, but that again, just eliminating noise, you still need the whole process of getting the sound. I am still learning the sound part. No, there will be no math for that.

[jazbo8]

I think you will like this paper by van der Ziel, although his thesis on 1/f flicker noise is not widely accepted by his peers (don't ask me why, it's over my head), anyhow, the paper does contain a lot of references to prior research done on active device noise.

[Alan0354]

I think you will like this paper by van der Ziel, although his thesis on 1/f flicker noise is not widely accepted by his peers (don't ask me why, it's over my head), anyhow, the paper does contain a lot of references to prior research done on active device noise.

Ha ha, these material physics is above my head too. It is difficult to understand the origin of the Flicker noise and how to predict it. That's the reason they put a "K" in the equation. I think it is more a measured parameter than calculated parameter. Someone can sit down and measure the flicker with known input resistance.

But, so far, we pretty much using 12AX7 for the better or worst. So you can spend all the time and at the end of the day, you are stuck with the tube.

Thanks for your information. You are pretty good in researching.