Load Line, Q-Point, and Frequency Response in Excel

[martin manning]

Here's something I've been working on for a while that others may find useful. This gathers up knowledge I've gained from a variety of sources including this forum and, notably, help and encouragement from the Valve Wizard. Constructive feedback of any kind is welcomed, as I'm sure it is not perfect yet.

Cheers,

MPM

Note: The file includes macros, which must be enabled to open it. Security must be set to medium or lower.

Edit: Attached hereto is the latest version.

[Structo]

Wow!

Looks like a lot of thought went into that.

But, I have no idea what all those numbers mean.....

[Bob-I]

That's amazing. I'm gonna spend some time with this to try to understand all these number.

[Phil_S]

This is way cool. I'm an accountant and I use Excel a lot, just not in this way. Very nice work. Thank you.

[M Fowler]

Lot of data there and my Sylvania Tecnical Manuel has a lot of this information as well but not this organized.

Mark

[martin manning]

Thanks, guys. Feedback and questions will inspire improvements!

Tom, the beauty of the numbers is that you can see where your operating points are, how they move, and what will happen to the frequency response as you change the values of the R's and C's.

Phil, yea, I don't think you or your clients would want any imaginary content in your spreadsheets ;^)

MPM

[David Root]

Martin, that is a real piece of work!! I've been playing with it a bit lately and in many ways it is complementary to TubeCad in the grounded cathode mode, which many of us use a lot. I especially like the freq. response plot, and the clear distinction you make between AC & DC loadlines.

I also like being able to add other tubes, eg 5751 and so on.

Thank you very much for the many hours you must have spent creating it. I see +50 downloads of it so it is getting around.

One comment: your database is all from original fifties datasheets, nothing wrong with that, they should be the benchmark, but people should be aware that modern tubes generally fall short of these specs, particularly in the gain and DC current areas.

[martin manning]

David,

You are most welcome, and thanks for the applause. I don't have Tube Cad, but please let me know if it seems to disagree anywhere.

I used available data sheets that were of good quality, since it is tedious to digitize them. The Sylvania 12AX7 sheets are a work of art. If you or anyone else work up any more, perhaps they could be shared and added.

MPM

[martin manning]

I made a little update on this. A forum member suggested that it might be nice to have some extra slots and buttons for additional tube types available, so this version has that feature. No bugs reported so far in the original version.

Cheers,

MPM

Edit: attachment removed. See latest version posted at begining of thread.

[wscrane]

Neat sheet.
It would be cool if you could get a plot of THD as a function of signal level on your loadline. It seems like a perfect job for a computer and the last time I checked tubcad didn't have it, although that was a couple years ago.

[wscrane]

Thanks martin (for doing and sending the THD calc privately sub rosa.)

I only realized after looking it over that what I asked is a real perplexing conundrum. I think the method you used involves the largest signal that doesn't clip. But since that signal is different for different Q points you can only compare loadlines with the exact same bias point. A hotter bias will always give lower distortion since it uses a smaller signal.

A possible solution might be to pick some kind of standard size signal and permit both cutoff and clipping as you vary parameters.

A good reference on this stuff is Eastman Fundamentals of Vacuum Tubes. He has a really simple way of getting the Fourier amplitudes in appendix B.

I 'm trying his approach right now. But I'm using a 12AX7 spice model to get Ip since I have no clue how Eastman got such a smooth output waveform off his loadline. (Another difficult aspect of this that didn't occur to me.) Its fig 9.46 in the third edition.

Sorry to muck up the thread with all this distortion-talk. The spreadseet is really great.

[martin manning]

No problem; this is all good stuff.

Yes, this is the so called "five equal voltage ordinates" method (RDH4 p. 564), and it estimates distortion at clipping (signal peak at Vg-k=0). Similar methods using more points are shown to get % distortion at higher orders. There is no reason any of these couldn't be modified to use an arbitrary peak-to-peak signal voltage, and in fact it seems it would be quite easy to do. Probably more points would be better if clipping or cut-off is allowed. I'm interested to see what you come up with... I think I mentioned my fear of a Fourier analysis picking up on artifacts resulting from the coarseness of the typical plate curves ;^)

I think what most folks here are doing is varying Ra and Rk proportionally, which keeps the bias point more or less at the same Vg-k. So, the classical approach is fairly representative of that case. Version 7.00 is posted here so others can see it, it has only small cosmetic differences from the beta version I sent off-line.

MPM

Edit: v 7.01 has additional comments and schematic on usage notes tab

Edit: attachment removed. See latest version posted at begining of thread.

[rsi]

This looks cool. I am getting a "can't run macro" type error when I click on one of the triode type buttons (12AT7, 12AX7, etc.).

Any idea why that is?

[martin manning]

Are you running in Excel or some other program (Open Office, e.g.)?

If in Excel, did you check the security level? Tools -> Macro -> Security..., then set to Medium or lower.

The latest Mac Office does not support VB macros, so if that is the problem you are out of luck using the buttons. You can always copy and paste the plate curves from the device library over the ones on the input data worksheet, though. Use Edit -> paste special -> values to preserve the formatting

MPM

[Ripthorn]

Sorry to bump an old topic, but I just downloaded this and am interested in putting in some data for some submini's in the custom slots. Do you have to manually put in all the data for the load lines, etc. on the Input Data tab? Thanks.