Marshall Major Schematic Accuracy - Phase Inverter Question

[Krinkle]

Based on the attached Marshall Major schematic and the curves of the 12AX7, how can the grid of V2b be 10V less than than the cathode? Doesn't that mean that the tube is biased at -10V and is way past cutoff? The curves only go as high as -5V.

I've heard that you can't get an accurate measurement with a voltmeter at the grid of a cathodyne PI. It's somewhere on Rob Robinette's site, I think, but I can't find it right now. Can anybody provide any clarity on this?

[martin manning]

I've heard that you can't get an accurate measurement with a voltmeter at the grid of a cathodyne PI.

True, even for high input impedance voltmeters like DMM's, and that applies to measuring the grid voltage on LTP inverters too.

[sluckey]

Something I like to do with these high impedance bootstrapped circuits such a s LTP-PI, cathodyne inverters, some cathode followers, etc... I measure the voltage at the bottom end of the grid resistor rather than directly on the grid. Assuming there is no grid leak current happening, the voltage at the bottom of that resistor will be the same as the voltage on the grid. You get a much more accurate grid voltage reading like this.

[nworbetan]

how can the grid of V2b be 10V less than than the cathode?

This one's simple enough. Imagine V2 sitting there idling and nothing interesting's happening. A small but measurable and steady amount of current is flowing through the tube, cathode plate et al, and also through the 2,700 ohm resistor and both of the 100,000 ohm resistors. The chart says a ten volt difference between the top of the 2.7k and the bottom, and that sounds fine to me. Ten volts divided by 2,700 is .003 amps, or 3 milliamps. That seems reasonable. Is it though? I guess I can't say for sure.

[sluckey]

That ain't the reason. The reason is because the meter is loading the high impedance grid circuit and causing a significantly lower voltage reading than the actual voltage that's really on the grid. Even my high end Fluke 87V dmm with a 10MΩ input resistance gives a useless reading.

[nworbetan]

Here are the notes I have within arm's reach from an Ampeg V-4B and you can see that I don't even bother making useless readings on the grids. I guess I'm also another person that doesn't quite understand exactly why that happens. Can you give us a more detailed explanation?

IMG_20200324_090951_3.jpg

[sluckey]

Read the second part of reply #2...

https://el34world.com/Forum/index.php?t ... 5#msg84005

Then do as I suggested earlier... Measure the voltage at the bottom of the grid resistor and know that the voltage at the grid is the same even though your meter can't accurately measure it.

[nworbetan]

Your point about measuring the voltage at the bottom of the grid resistor being more accurate than directly at the grid is important and I'm learning from that. However, when you say that our meters aren't accurately reading the voltage at the grids is a little bit wrong too. Our meters (and not just the fancy ones) truly can accurately read and display a grid voltage. It's just not the same as the quiescent grid voltage that happens without the meters connected.

No, in fact, the 100% expected and 100% normal behavior is for the grid voltage, when you connect a dmm to it, to drop down below where the cathode voltage was. It's typically not useful on the grids of tube stages with extremely low cathode voltages. I'm not going to insult you and call this a troubleshooting 101 lesson, but if you're being honest you may have just learned a thing or two too.

[sluckey]

However, when you say that our meters aren't accurately reading the voltage at the grids is a little bit wrong too.

I never said your meter is not accurate. I said your meter cannot accurately measure that voltage. There's a difference. You're taking that statement out of context and making a generalization about it. My comment was directed at the LTP PI, AC coupled cathodyne, AC coupled cathode follower, etc. All of these circuits are ***BOOTSTRAP" biased to the cathode and your 87V or my 87V cannot accurately measure the voltage on the grid in that circuit. A Simpson 260 is even worse. Even your scope will not measure it correctly.

I challenge you to find any current production meter that costs under a thousand dollars that can accurately measure the grid voltage of a LTP PI.

[Krinkle]

Thanks for the replies guys, confirms what I thought.

In the meantime I did some poking around the net and it looks like the schematic itself should be correct. This is a "corrected" schematic as there are mistakes in a lot of the other schematics floating around the net, and I was suspicious of the correction. I found that the Princeton and 5E3 deluxe have similar PI's but the output to the lower side of the load is taken from the cathode in those circuits. I was wondering if that was a mistake, as well. Then I read somewhere that moving the output to the bottom of the cathode resistor was a tweak that came from the Radiotron Handbook, V4 to improve balancing. So I'll just ignore the fakeout grid measurement and move on. I'm getting about 2.3mA across the 2k7 cathode resistor so maybe that seems to make sense regards to the comment on idling and tube self bias? I see the note about measuring voltage and turning off the amp and measuring resistance, I'll give that a shot and cofirm the real number.

The reason I'm asking questions is I'm getting a splatty overtone at the end of single G and A notes played on the low E. I am looking at the Paul C mod and a few other ideas, from Lyle Caldwell and Rob Robinette on other sites. LC seem to have had a similar noise as me on a Princeton build and lists a bunch of things he did to get rid of it.

https://www.mylespaul.com/threads/5e3-c ... ts.159635/