Maximum cathode current for an EL84?

[mpegjohn]

Hi all,
I am a little confused over the specs for the EL84.

I had a brilliant idea, until I read them. I wanted to make a variable constant current sink 0 - 100mA. However I see that the Ik max = 65mA.
I think I must be looking at this incorrectly, as the graphs show Ia > 150mA.

Can someone explain please why Ik is much lower than the IA values, as I was under the impression the cathode current = Ia + Ig2.

Thanks,
John

[kalee20]

Sometimes valve data does show voltages, currents, power etc beyond the maximum values. It gives an indication what the valve will do, though it is not a good idea to operate it there long-term.

Transistors die fairly quickly if you operate beyond maximum ratings, though even there, it is sometimes possible to run at maximum Vce and maximum Ic for a short time (1msec perhaps) which grossly exceeds maximum dissipation rating, but as the thing has a certain heat capacity, it won't reach destruction temperature if the time is kept short. Valves are even more forgiving.

[ms660]

So far as I can make out the data for the EL84 for the limiting values are for the design centre rating system and not the absolute maximum rating system (see EL84 data in the Philips book in the link below) details of the systems can be found in the Mullard or Philips valve manual, in the one below it's in Section 4:

http://www.nj7p.info/Manuals/PDFs/Tu...art-4-1972.pdf

Lawrence.

[Restoration73]

Many output valves need to be derated when used as dc regulator, and EL86 would be
a better choice; https://frank.pocnet.net/sheets/129/e/EL86.pdf

I would tend to use an EL34 or even a KT66 if you have the space.

[G8HQP Dave]

Some limits are instantaneous limits (go beyond this and it immediately breaks) while others are average limits (go beyond this and it won't last very long). Signal peaks can often go beyond the latter, provided that the quiescent current stays within the limit. Graphs will sometime show points outside limits but this does not mean it is safe to go there. The valve maker can afford to destroy a few samples when getting his data.

For regulator duty you need a spec for maximum continuous current, which may or may not be given for most valves.

[mpegjohn]

Thank you all.
I think then, after looking at the Philips Manual, I will take Restoration73's advice and go for an EL34.

Regards,
John.

[Heatercathodeshort]

A pair of EL84s or a single EL81 might be a much cheaper option. J.

[G8HQP Dave]

PL81 or PL36 would be even cheaper.

[Heatercathodeshort]

Agree but you have to deal with the odd heater voltages. J.

[RF Burn]

Originally Posted by mpegjohn

I see that the Ik max = 65mA.
I think I must be looking at this incorrectly, as the graphs show Ia > 150mA.
John

Hi John,

The Ik max figure is the maximum continuous DC current that the cathode should be allowed to provide for normal operation of the valve, and it is indeed the sum of Ia + Ig2 (+ Ig1 in some circuits).

When the manufacturers prepared the operational graphs for their valves they would show curves that went into areas which exceeded not only the maximum DC cathode current for the device but also the maximum DC anode voltage (65mA and 300V for the EL84 - Mullard Data sheet Feb 1961). However the important point here is the 'DC' notation.
The graphs were obtained from measurements performed on average valves operated under 'pulse' conditions where the average current is much less than the DC current, i.e. the valve is not stressed from the excessive cathode current or anode dissipation. The reason for this is that quite often valves would be used in circuits where the signals are of a pulsed nature, so the designers of such circuits would need to know how far they can 'push' things before the valve reliability is threatened, otherwise they could have been forced to use a more powerful, possibly larger and almost certainly more expensive valve type and then underrun it!.

The upshot of this is - if you are designing a circuit which has fairly constant Ik such as Power supplies, Class A amplifiers, anything 'DC' then abide by the maximum (limiting) values in the data sheet - If you are designing a pulse circuit, or, one which spends more time 'off' than 'on' (SSB Transmitter PA for example) you can use higher values as long as the average current (and anode dissipation) does not exceed the limits.

Adrian

[GMB]

I have a power supply that actually uses EL81 (pair in parallel) to stabilise the output. The xL81 valves have a very high maximum current of 180mA - but they don't have a big maximum disipation to match it. A PL36 gives best of both.

The problem with a variable current sink is the voltage it is expected to take. Unless you are talking about quite low voltage it will be the dissipation that will be the problem.

[mpegjohn]

Thank you.
I have ordered an EL36 off ebay.

I know this is basic stuff but how can I work measure the resistance of the current sink.
I thought using two resistances and then measure the change in current. Then using a Norton model how do I work out the internal resistance?

[rontech]

Many years back I re purposed an ex military pa amp and made a 300v DC regulated power pack using a KT66 as the main tube and a neon / 6J7G regulator as the reference.

The circuit was from an old book ( Bernards Radio Manual No 93 ) It specified a 6L6 power valve but the KT66 seemed to work just fine.

[GrimJosef]

Originally Posted by mpegjohn

... I know this is basic stuff but how can I work measure the resistance of the current sink.
I thought using two resistances and then measure the change in current. Then using a Norton model how do I work out the internal resistance?

The Norton model of your real-world current sink consists of an idealised perfect current sink, which always (because it's perfect) carries current I0, in parallel with a (hopefully large) effective resistance, and it's that effective resistance, let's call it R0, that you want to measure.

You're right, loading the sink with two different resistors, R1 and R2, and measuring the currents through each of them, I1 and I2, will allow you to work out the value of R0. The circuit model and the relevant equations are shown in the picture.

Your current sink circuitry is shown inside the dotted-line box. It has two terminals across which your load resistors are connected. The voltage at the terminals is V1 for load resistor R1 and V2 for load resistor R2. If you know the two resistances and you measure the two voltages then you can calculate everything else.

Kirchoff's Current Law says that the sum of all the currents entering any node has to equal the sum of all the currents leaving that node. The first two equations, (1) and (2), reflect that for node A. The only current entering comes from the load resistor and is I1 or I2. The two currents leaving are I0, through the perfect current sink, and V/R0, through the real-world current sink's resistance.

Ohm's Law for the load resistors gives us equations (3) and (4). The rest is algebra - sorry, I just noticed I forgot to label the 7th equation as (7). The final equation will tell you R0 if you substitute the known values of R1, R2, V1 and V2 into the right hand side.

Cheers,

GJ

[Mike. Watterson]

Shell / Bomb / Fuze ratings of valves can be exciting. Instead of 1W and 8,000 hours, you have maybe 800W and a minute. Probably proximity radar. Russian military valve data sheets.

I'd use big MosFet on a heatsink. Some of those are very cheap and can do 20A and 1000V, but not at the same time, though they are in BIG packages and certainly some can do 200W or more.
Tend to have like 2nF to 20nF gate capacitance depending on current rating because the chip is loads of paralleled FETs.

No awkward heater supplies, though you need op-amps and drivers etc.

[G6Tanuki]

Practically, you need to consider what service your EL84 will be used in.

Simple single-ended Class-A - its cathode needs to be able to deliver twice the emission-current of its static-current, because if it can't you get 'clipping'.

If you're running a push-pull pair in Class-AB2 or well into proper full-on Class-B you may be only running 20 Milliamps of static current per valve, but on peaks more like 150 Milliamps - so your maximum cathode-emission needs to be good enough to handle this.

I built a bunch of guitar-amps and transmitter-modulators using a pair of EL84 and with proper AB2 biasing they could do 40W output for ages.

Though I rather preferred the 5763/QV03-12 because I could run a pair of them in near-class-C [static anode current 20 milliamps, peaking at 300mA at 350V HT] to give 50W and despite the anodes glowing they would survive for 100+ hours.

It was never Hi-Fi, we always had things TURNED UP TO ELEVEN!!!