June 2001

New - 20 June 2001: additional information on valve codes - see below.

My little 3W valve amplifier circuit - as featured in the March issue of Practical Wireless - resulted in a couple of interesting letters. One writer expressed concern at the amount of negative feedback I'd suggested. Too much negative feedback can cause instability, even oscillation. If you do experience any instability, try increasing the value of the feedback resistor R12 to 20k or more.
Unfortunately, brief bursts of instability - which are often only present when the amplifier is being driven - can be difficult to hear. It's best to view the output of the amplifier on an oscilloscope if you have access to one.

The valve pictured on the right is a Mullard PM2DX 2-volt battery valve. I unfortunately knocked it over and may have damaged the filament. The filament draws 190mA at its rated 2V, which is strange because the data I have for the PM2DX says it should only draw 100mA. However, there's one book I have which gives the filament current as 250mA, but that only applied to the very first PM2DXs that were made. I'd like to know whether I have one of these early versions, or whether I damaged the filament when I knocked the valve over.

12V h.t. Valves

One topic in this month's column concerned those valves which were introduced in the late 1950s for use in car radio sets. They worked from a h.t. supply of just 12V. [More info] Several people have told me that the 12V h.t. ECH83 was rumoured to be just a specially selected ECH81. I found that the inter-electrode capacitances - useful in comparing the physical dimensions of electrode structures - of both the ECH81 and the ECH83 are the same but for one tiny discrepancy. Even more interesting, the 12V h.t. EBF83, and its mains equivalent, the EBF89, have identical inter-electrode capacitances. I finally bought a set of 12V valves and after checking the manufacturers codes on the valves I came to the conclusion that the ECH83 and ECH81 are indeed the same. Likewise, the EBF83 and EBF89 appear to be the same.

While most American 12V h.t. valves (tubes) were specifically designed to run from very low h.t. supplies, many European 12V h.t. valves seem to have been derived from normal mains types. Indeed, at least two are mains types - as described above - albeit they are specially selected on test. One type I've found that seems to have a mains antecedent is the ECC86, a v.h.f. double triode. I have a couple of ECC86s - both made in Philips's Heerlen factory in Holland - and they both look identical to an ECC189 I have which was made in the same factory. Although the ECC189 has a remote cutoff characteristic and the ECC86 has a sharp cutoff characteristic, that only affects the winding pitch of the grid wire.

From what I've discovered, there's clearly good reason to suppose that other mains-type valves could be useable at very low anode voltages. The valve types that might offer the best chance of success are those with a high mutual conductance and a low impedance. Valves designed for use in (UK and European) u.h.f. television sets ought to be a good place to start looking. However, be aware that valves made by different manufacturers, and even valves from different batches, can exhibit wildly differing characteristics when operated at very low h.t. voltages.

Valve Codes

So what are these 'valve codes' I've mentioned? You might have noticed that the markings on most valves rub off at the slightest touch. But there are often other markings, near the pins of many miniature valves, which are far more durable. These are manufacturing codes. They identify the valve type and where and when it was made. You'll find them on most European valves, particularly those sold by Philips and Mullard. They're often found on valves sold by other manufacturers too, simply because no matter whose name is on the box, most valves were made in the same factories. Take a look at a miniature valve and you'll usually find both a three-character code and a four-character code. The photograph on the right shows the codes on a Mullard ECC81.

The first two characters of the three-character code indicate the valve's type: VF = ECC81.
The third character is the revision number, which would be incremented when the valve underwent a design change of some sort: 3 = revision 3 of the ECC81 design. (Yes, it is a '3' although it's difficult to tell from the photograph.)

The first character of the four-character code indicates the factory in which the valve was made: B = Mullard's factory at Blackburn in the UK.
The next three characters, in order, are:
The last digit of the production year: 4 = (probably) 1964.
The production month (A = January, B = February, etc.): F = June.
The week number in the month the valve was made: 3 = third week (of June).

I found the type code on my ECH83 was YD. Yet that's the code for an ECH81. Similarly, the G3 code on my EBF83 is actually the code for an EBF89. Interestingly though, both the ECH83 and EBF83 do have their own codes: NJ and MJ respectively.

There's a copy of the factory codes listing on the Internet at: http://www.triodeel.com/philipsfactorycodes.jpg. The list of valve type codes is at: http://www.triodeel.com/images/philipstubecodes.pdf, which also includes (albeit at a lower resolution) the factory codes list.

Return to the contents page