Colossal,
Thanks for taking the time to answer my questions. First, the equation for scaling the the screen curves is in deed the same as RDH mentioned earlier. No surprise there, as the procedure is well documented in the past.
And your effort to curve fit the tetrode and pentode curves is admirable, it is a probelm that many have tried to tackle - it is not that easy! Since each tube type has a different characterstic, some form of curve fitting would still be needed at the end. OTOH, it has been suggested that we do not want to have perfectly fitted curves because of the wide manufacturing tolerances of tubes, so +-20% is usually more than sufficient.
Most of the tetrode/pentode models (for SPICE simulations) are based on the improved Koren triode model, which is itself based on the Langmuir-Childs equation. I won't bore you with the details, but if you google "pentode spice model", you will have a wealth of material to check out.
How do the "big boys" do it? (Tetrode load lines)
Moderators: pompeiisneaks, Colossal
Re: Pentode Models
I completely agree with this. Hahah, I guess my quest for the model that predicts tube behaviour is probably folly given tube variance.jazbo8 wrote:OTOH, it has been suggested that we do not want to have perfectly fitted curves because of the wide manufacturing tolerances of tubes, so +-20% is usually more than sufficient.
Yeah man, I've checked out as much of this exact kind of stuff as I can find. I coopted some equations specifically from a tube model software package user manual (I'll have to dig around for that PDF). I think these were the same guys though that referenced some very obscure variables that said "refer to the data sheet" with no further explanation as to what purpose they served. I think they were weighting terms but found that really you could only vary things so much and the pentode shape never really conformed all that well (as you can see from my load line). Probably plenty good enough, although I really wanted to get this right as I want to tackle plotting the composite curves next and don't really want to build on a weak foundation.jazbo8 wrote:Most of the tetrode/pentode models (for SPICE simulations) are based on the improved Koren triode model, which is itself based on the Langmuir-Childs equation. I won't bore you with the details, but if you google "pentode spice model", you will have a wealth of material to check out.
Next-Tube Programs
For those that have Matlab, Eugene Karpov at Next-Tube http://www.next-tube.com/programs.php has kindly provided several Matlab programs for analysing and designing single-ended, and push-pull stages. If you don't have Matlab, Stephie Bench has several excel worksheets for plotting loadlines for SE and PP output stages. Google is your friend... 
Jaz
Jaz
-
martin manning
- Posts: 14308
- Joined: Sun Jul 06, 2008 12:43 am
- Location: 39°06' N 84°30' W
Re: How do the "big boys" do it? (Tetrode load lines)
In my opinion, plotting composite curves is an academic exercise, unless you are interested in the finer details of harmonic distortion.
My strategy is to digitize only the Vg1=0 curve, and then scale it to the desired Vg2 to get the new knee location. Also of interest is the idle grid voltage that will be required to set a given idle plate dissipation at the new Vg2 and Va-k. To get this, the Ia and Va-k from the load line can be scaled back to the reference Vg2 so that the reference Vg1 can be read from the un-scaled data. Then, that reference Vg1 value can be scaled to correspond with the desired Vg2.
After some study, I tossed out what I got from Kuehnel’s power amp book and went back to the basics as described in RDH4. As proof that this technique is easily close enough for rock and roll, here are two examples using plots of Ia @ Vg1 = 0 for a wide range of Vg2 from GE data sheets. In each case I digitized the Vg1 = 0 curve at the highest Vg2 and scaled it to match each lower Vg2 curve on the data sheet. As you can see, the accuracy is really quite good.
My strategy is to digitize only the Vg1=0 curve, and then scale it to the desired Vg2 to get the new knee location. Also of interest is the idle grid voltage that will be required to set a given idle plate dissipation at the new Vg2 and Va-k. To get this, the Ia and Va-k from the load line can be scaled back to the reference Vg2 so that the reference Vg1 can be read from the un-scaled data. Then, that reference Vg1 value can be scaled to correspond with the desired Vg2.
After some study, I tossed out what I got from Kuehnel’s power amp book and went back to the basics as described in RDH4. As proof that this technique is easily close enough for rock and roll, here are two examples using plots of Ia @ Vg1 = 0 for a wide range of Vg2 from GE data sheets. In each case I digitized the Vg1 = 0 curve at the highest Vg2 and scaled it to match each lower Vg2 curve on the data sheet. As you can see, the accuracy is really quite good.
You do not have the required permissions to view the files attached to this post.
Re: How do the "big boys" do it? (Tetrode load lines)
Very nice as usual Martin.
-
chief mushroom cloud
- Posts: 429
- Joined: Tue Jan 18, 2005 5:42 pm
- Location: Peenemunde CA
Re: How do the "big boys" do it? (Tetrode load lines)
"close enough for rock and roll"........martin manning wrote:....After some study, I tossed out what I got from Kuehnel’s power amp book and went back to the basics as described in RDH4. As proof that this technique is easily close enough for rock and roll, here are two examples using plots of Ia @ Vg1 = 0 for a wide range of Vg2 from GE data sheets. In each case I digitized the Vg1 = 0 curve at the highest Vg2 and scaled it to match each lower Vg2 curve on the data sheet. As you can see, the accuracy is really quite good.
I surprises me how people just LOVE to complicate things....after all....they are, indeed, just rock n roll guitar amps....
Don't overthink it. Just drink it.
Re: How do the "big boys" do it? (Tetrode load lines)
Many here also work on hifi audiophool amps, where we need to get into more of the nitty gritty 