Are Positive Grid Voltage Traces Necessary?

[crj11]

I am in the final stages of designing a programmable tube tracer. To reduce cost and complexity, I'm thinking about leaving out the ability to generate traces with positive grid voltages.

Are the plate curves at positive grid voltages generally useful, or can they be safely ignored?

[HollowState]

Originally Posted by crj11

Are the plate curves at positive grid voltages generally useful, or can they be safely ignored?

If one wishes to use a zero or near zero bias triode with low plate voltage, then positive grid voltage is needed to get good current draw. Typical tubes would be transmitting types like 811, 805, 838 and 203A to name a few as well as other high voltage types. Since I'm old school, doing that is using the wrong tube. (or the tube wrongly ) But a lot of builders do it anyway because they have the tubes and insist on using them. I think the Japanese started this trend in the late seventies, early eighties in that MJ Stereo Technique magazine. (mu-sen-to-zee-ken) Now it's up to you if you want to cater to that use.

[SY]

I think that it's useful for output tubes, not just zero-bias triodes, but also tubes which are happy being driven into AB2 (e.g., 2A3, 6L6/7027/5881).

[TheGimp]

6P1P-EV data sheets show positive grid drive. I've been wondering if this implies the tube is designed for it?

6AQ5 shows positive grid voltage traces for grid voltage of +5, +10, +15 V. That would seem to me to imply that the tube IS designed for class B operation. and if the 6P1P is designed as a Russian replacement for the 6AQ5 then it too should be capable of class B operation.

So I think in there is good argument that positive grid drive testing might be useful even though the use in class B for audio is limited (but does exist, as I've seen it discussed).

[crj11]

It seems that positive grid voltages could be a useful feature.

I've been looking at less complicated ways of generating the positive and negative voltages. I have a 100V programable floating supply, so I can ground either rail, the question is the cheaptest and most effective way of doing that.

I originally was going to use a MOSFET bridge to switch the output and ground connections, but that seemed to complex, especially since you would need to use isolated gate drivers to turn on the MOSFETS. More recently I have been looking at just using a couple of SPDT dip relays.

The tracer uses a 12-bit ADC to measure the grid voltage. Because of the dual polarities, and the 0-5V input range of the ADC, my though was to use a TLE2426 "rail-splitter" chip to get a low impence 2.5V source and then use a 40:1 resistor divider to generate 5V to 0V as the voltage goes from 100V to -100V.

The other complication which I need to think about is protecting the supply from flash over from the plate in a gassy tube. The plate supply is fused, so it's easy to protect a negative grid. Just put a diode that conducts a positive grid voltage to ground. It is less clear to me how to protect the grid supply when it is positive. I guess I could always just use higher voltage components in the grid circuit. That means more expensive and larger 500V components instead of 100V components. I guess I still have some cost/benefit tradeoffs to consider.

[Yvesm]

Even lo level tubes, specially frame grid, may exhibit non negligible grid current even at negative grid voltage.
If the grid supply is unable to provide it, your curves will be somewhat compressed near 0 Vg like here:

[crj11]

Yves,

I agree that keeping the voltage stable near zero volts will be challenging. I think that it will probably be harder at slightly positive voltages than slightly negative ones.

Chris

[astouffer]

Have you thought of using a TL783 regulator for the grid bias supplies?

Linear Regulators - Standard Voltage Regulators - TL783 - TI.com

[Ty_Bower]

Originally Posted by crj11

I am in the final stages of designing a programmable tube tracer. I'm thinking about leaving out the ability to generate traces with positive grid voltages.

Is your tube tracer for fun, or for profit? If you're designing a product you hope to sell to the public, will you limit your target audience if you eliminate this capability? If you are building it strictly for your own use, ask yourself if you ever design circuits that operate in that region.

[crj11]

Originally Posted by astouffer

Have you thought of using a TL783 regulator for the grid bias supplies?

Thanks for the link. I wasn't familiar with that chip. It definitely has a large input range. I'm actually working on a custom low noise switching supply to keep things as small and efficient as possible.