plate curves and max ratings questions

[EddieRich]

I am building a computer controlled curve tracer and noticed something while I was looking at some plate curves from various tubes datasheets. All of the curves far exceed the maximum plate voltage and power. So, I have some questions about how to obtain my data.

1. Can I exceed the max plate voltage for a short ( < 1 sec ) amount of time ?
2. Can I exceed the max plate power for a short amount of time ?
3. For pentodes, the data sheet screen voltage is fixed. So is it alright to operate a tube with the screen voltage many times higher than the plate voltage ? For example, 250 volts on the screen and collect data at plate voltages of 30, 40, 50, etc. ?

I've attached a basic test circuit. All the supplies are computer controlled and have computer feedback as to voltage and current.

Any enlightenment would be appreciated.

Ed

[kevinkr]

1 may or may not be an issue depending on the electrode geometry - arcing is the concern here. I would not exceed the rated plate voltage by much more than 10%. (Swag on my part)

2 should not be an issue with reasonably fast acquisitions..

3 may be of concern with some types of pentodes/beam tubes.. I think if it is done very quickly so that average screen dissipation is not exceeded it will probably be OK with most types.

[TheGimp]

One trick you can use is to alternate sweeps of high dissipation with sweeps of low dissipation.

For example you do not sweep the anode voltage with grid voltage steps in the order 0V, -0.5V, -1V, -1.5V, ...-15V.
Instead you step the grid in the order 0V, -15V, -0.5V, -14.5V, -1V, -14V, ...

This alternates low dissipation sweeps with high dissipation sweeps to keep the average lower over the short period of time.

[EddieRich]

Thank you for the replies. I hadn't thought of a staggered step on the grid, I was thinking linear. I think if I plan my sequencing, I can achieve what I want.

I'm just wondering how the original tube manufacturers got their data. But then again, I'm never going to operate the tubes above the max, so I don't really need data at those points.

[MrCurwen]

On most common audio tubes this wouldn't be an issue, but with some tubes the good part of the curves is above max ratings. On some tubes the max ratings are for one reason or another super conservative, and in ICAS use are not necessarily important.

For example the 1626, used in Darling amps and sometimes as a preamp tube, the sound starts to get good way above the max ratings. I ran mine at basicly double the max plate dissipation - 5W is quoted in the specs, I had them making music at 9W. No plate colour or anything strange. Tube life is one or two years (at 12h per day), so in comparison to the price quite reasonable still.

[artosalo]

Are you going to build something like this:

MODULE LAMPEMETRE ANALYSEUR - VACUUM TUBE ANALYZER | eBay

You can exceed the max. ratings by using very small duty cycle during the measurement. Some 1...5 ms active and 100 ms "wait".
Then the average power dissipation stays minimal, even with high peak current and the tube will not suffer.

For example, this Lampemetre operates so that the control grid is allways kept at max. negative voltage, except at those small periods when the measurement is done. The anode voltage is generated and the measurement "syncronized" to half wave rectified mains voltage, which is not filtered. So the wave form is a pure positive half of the AC-cycle.

When the anode voltage begins to rise from zero, the measurement period starts.
The anode current is monitored at different control grid voltages and at different anode voltages during the 90 degrees when the voltage rises from zero to max. This cycle is then repeated few/some times depending on the how many steps/grid voltages are selected. Typically the test cycle takes less than 5 seconds.

Typical result looks like this (6j9p triode connected) :

[Merlinb]

Originally Posted by EddieRich

I am building a computer controlled curve tracer and noticed something while I was looking at some plate curves from various tubes datasheets. All of the curves far exceed the maximum plate voltage and power.

Do you have some examples? Off hand, I can't think of any I have seen where the curves go beyond the max peak voltage. Rarely does the horizontal axis even go beyond 500-600V.

[TheGimp]

Some do and some don't.

6GH8A is spec'd at 350V max anode voltage but the curves to to 450V.

http://www.mif.pg.gda.pl/homepages/f...35/6/6GH8A.pdf

This 6BQ5 shows anode plots out to 600V with a max rating of 550V.

http://www.mif.pg.gda.pl/homepages/f...127/6/6BQ5.pdf

Unless you are flogging the tubes, there is no reason to go beyond the mfg specs.

That said, the 6DX8 is only spec'd at 4W plate dissipation with 250V anode max and I've run them well beyond twice that dissipation and 350V with no problems. I didn't see any sign of red plating till I got to 12W!

So, I'm creating my own data sheets for this tube for which I intend to do some long term testing at stressful levels.

[EddieRich]

@artosalo : I have looked at that, but I is ejicated in the U.S. public school system, so I don't know how to read French. I had considered using a half wave unfiltered anode, but I want to be able to do static tests as well, like maybe biasing a tube and injecting signals so I can view it on a scope, so I want the exact voltages to be under the control of the computer.

In addition, this is a learning process for me. It's not that I want a curve tracer per se, I want to get the experience of working with high voltage electronics. When my learning journey is over, I will have a curve tracer / programmable power supply I can use.

@Merlinb: Examples, well just about every curve I look at. I have a 'lot' of tubes I got from ebay, and one of the first amps I am building will be an iPod dock using a 12AB5. The plate curves go out to 500V, yet the max plate voltage is 330V. The triode preamp will use a 2C22 which has a max plate of 300V and the curves extend past 600V. When I draw the power curve and load line on the plate curves, all the curves extend way beyond the max power. For instance, the 12AB5 0V curve goes out to 450V @ 120mA, that's 54 watts and the max power is 12 watts.

I am reading the 'design center' max values. Is there some other spec I should be looking at instead?

Just a little background on me, I'm 47 years old and have been working with electronics since I was a teenager, both professionally and as a hobby. This is my first attempts at anything over +/- 15V. I am a software engineer, but my specialty is controlling hardware and data acquisition.

[dcgillespie]

One thought to help you approach your question is to look at real world performance. Let's take 6550 tubes with 600 volts on the plate, developing 100 watts into a 5000 ohm load. Under full power output conditions, the tube that is conducting at maximum excitation is passing nearly .43 Amps, while the tube that is cut off at the same point has effectively over 1100 volts applied to its plate. These of course are well above published limits for the tube. The point is, those limits are based on continuous DC values (or average AC values within limits). A curve tracer -- although not an audio signal -- is still a very brief application of excitation, much as an audio signal is at any given point in time. Therefore, it is quite reasonable to exceed the published limits for the brief period required to develop the traces.

Dave