Please click your language flag. Bitte Sprachflagge klicken.

The Philips Red valves series, part II

Ernst Erb Otmar Jung 
Please click the blue info button to read more about this page.
Papers » Tubes and valves (history, technique etc.) » The Philips Red valves series, part II
Jacob Roschy
Jacob Roschy
D  Articles: 1568
Schem.: 17
Pict.: 68
21.May.05 09:32

Count of Thanks: 31
Reply  |  You aren't logged in. (Guest)   1

Continued from part I

Use of this article (entire thread, any parts) is not permitted for commercial purposes!

1938 additions

Triple diode with common cathode.
Released May 1938.

This valve was developed specifically for the three-diode circuit invented by Philips. A goal was to eliminate distortion and other side effects which arose with delayed automatic gain control. The first diode served as detector, the second as (undelayed) AGC rectifier and the third only to provide the AGC delay voltage. This was done by giving this diode a positive bias current, e.g. 5 µA, instead of the usual negative bias. A high-value resistor from the HT line supplied this current. A second resistor was connected to the negative AGC bias voltage.

As long as the negative current through this resistor stayed under the 5 µA (in this example), the diode voltage remained at 0 V. Only with larger negative currents did the diode no longer conduct and the voltage then go negative [= the principle of the clamp-diode] which gave the delay.

This circuit was used only in few top-of-the-range sets of the time. Afterwards, this technique went out of fashion, and no newer valves were developed for this purpose.

Double diode + variable-µ pentode, supersedes the EBF1.
Released July 1938.
gm 1.8 mA/V.

This is the first double diode variable-µ pentode with "sliding," rather than fixed, screen grid voltage in the Red series, replacing the earlier EBF1 in the same way as the EF9 superseded the EF5 (refer there). The Telefunken EBF11 has very similar characteristics to the EBF2.

Triode-heptode for frequency changing
. Supersedes the EK2,
released June 1938.

gc 0.75, Ih = 0.95 A !!

Since the AK1, it had been Philips's tradition to supply octodes as frequency changer valves. However, neither the AK1 nor the improved EK2 could match the performance on short waves which was possible with the triode-hexodes. In Germany there had been the ACH1 triode-hexode since 1934, in the UK the FC2 and AC/FC, and in France, Tungsram released the 6TH8 in 1937.

As one solution, with the EK3, Philips, tried to manufacture an alleged octode with the quality of a triode-hexode, which was, however, substantially more complicated than a normal triode-hexode!

A dwindling demand for the octodes drove Philips, besides the pseudo-octode EK3, to a further act of desperation, the triode-heptode ECH2!

The technology of this valve was, however, ridiculous compared to the others: its heater current is, at 0.95 A, nearly 5 times as high as the (very soon to be contemporary) Telefunken triode-hexode ECH11 and even higher than that the EL3N output valve. Also the physical size of the ECH2 was enormous for a signal- stage valve: the original Philips version had the large cylindrical, tapered-top envelope like the old EL3, while the Tungsram version here in the picture has quite a corpulent domed envelope, clearly even "fatter" than the EL3N output valve.

After the 1934 release of the last signal-stage valves with 4-watt heaters, and from 1936 those with only 1.26 W, the ECH2 now needed a full 6W.

This valve had probably rather an alibi function, so that the embarrassing question of customers, "When will Philips finally also supply triode-hexodes?", could now be answered with: "We now even have a triode-heptode!"

The acceptance of the ECH2 was only small, and Philips also seemed to have few hopes for it, because in important countries such as France it never appeared. Only the later triode-heptodes ECH4, ECH21 and ECH81 were highly successful million-sellers.

Variable-µ, low-noise HF amplifier valve, particularly for RF amplifier stages.
Released May 1938.
gm 1.8mA/V.

Despite "F" in the designation this is not actually a true pentode. For RF-(only) amplification one wishes in principle a valve with an amplification as high as possible, which calls for the use of pentodes. However, these produce an unwanted noise (partition noise), which results from the current distribution between screen grid and anode. The EF8 is constructed like a pentode, however yet another grid with same pitch as those of the screen grid is fitted before the screen grid, so that its turns lie exactly before those of the screen grid. If this auxiliary grid is on 0 V, then it diverts most electrons from the screen grid and the screen grid current is very small in relation to the anode current and the noise is correspondingly reduced.

The EF8 was used only in RF stages by top-of-the-range sets of this time, afterwards no newer valve of this kind was developed. The Telefunken EF13 serves the same purpose as the EF8, however with a different electrode structure.

variable-µ pentode, replaces the EF5.
Released May 1938.
gm 2.2 mA/V.

The EF9 is the first variable-µ pentode with sliding screen grid voltage in the Red series, replacing the EF5 which had been released two years earlier. Instead of a "stiff" screen grid supply voltage, which is produced e.g. by a potential divider (screen grid pre- and parallel resistors), only one screen grid supply resistor is required in this case, which produces the correct screen grid voltage of 100 V when there is no additional bias due to AVC action. Now if the grid bias is increased, the screen grid voltage rises as the screen grid current drops. This valve construction makes a larger slope possible whilst having lower anode current in the uncontrolled condition. gm =2.2 with Ia= 6 mA. The Telefunken EF11 has very similar data to the EF9.

This article was edited 25.Jun.05 17:27 by Jacob Roschy .

Jacob Roschy
Jacob Roschy
D  Articles: 1568
Schem.: 17
Pict.: 68
21.May.05 09:40

Count of Thanks: 60
Reply  |  You aren't logged in. (Guest)   2

Magic eye + variable-µ AF- pentode;
Multiple valve for tuning indication and (at the same time) AF- (pre-)amplification,
released July 1938.

The EFM1 was derived to a large extent from the EFM11 developed by Telefunken. Even the large envelope diameter of 37 mm, needed for the "Stahlroehren" (steel valves) base was maintained, although this was unnecessary with the side-contact base.

With the EFM1, the AF- preamplification was combined with the tuning indication such that the pentode section was used as a variable-µ valve with sliding screen grid voltage (see EF9). The screen grid is thus connected to the deflection rods of the display section. The control grid receives the AF- signal and also AVC voltage. Thus the AF- signal is controlled depending on the RF- signal strength, while the display is simultaneously controlled by the sliding screen grid voltage.

With the EFM1, it was possible to save a valve, or the number of valves did not need to be increased despite having tuning indication.

This gives these valve complement options:

EK2 + EBF2 + EFM1 + EL3 or

EK2 + EF9 + EFM1 + EBL1.

Because of the rather high distortion factor and the poor range of the display section, this valve combination did not work satisfactorily and was soon replaced by double range indicator eyes, the EM4 or EM11.

Four-Beam-Octode for frequency conversion, replaces EK2, r
eleased May 1938;
conversion gain 0.65, heater current = 0.6 A !

On frequency converter valves at the time (1938) one could chose between two versions: Octodes or Triode-Hexodes.

Octodes have a rather simpler electrode structure but have the drawback of frequency shifting on higher frequencies in the shortwave range caused by AGC action, which is pretty nasty due to the continual fading in and out of shortwave stations.

Triode-hexodes are free of this setback, but their construction is rather complicated due to the two separate triode and hexode sections.

Strictly speaking, this valve is also not an octode, but a triode-heptode with a circular horizontal arrangement of electrodes: around the common parts cathode and grid 1 (oscillator grid,) the system is divided into four sectors.

Of it in each case two opposite sectors, separated from each other by electron guiding systems form a triode to the left and right and two further, to the top and bottom, form a heptode.

The construction, from a monstrous tangle of wires and plates, is clearly larger than those of the EL3 output valve, with commensurate envelope size, and the heater current is, at 0.6 A, three times that of the normal octode EK2.

With the EK3, Philips could appear to protect their octode tradition, and probably did not need to pay Telefunken any royalties for the triode-hexode patent, but this valve was so expensive to produce that it probably had to be subsidised, so that anyone would buy it at all.

18-W output pentode, with double the slope of the EL5, supersedes it.
Released July 1938.
Pa=18W, 72 mA, -7 V, gm 14.5 mA/V; Ih=1.2A.

Since the AL3 and EL3 had been available, output valves of normal output power needing relatively small input signals (4.2 V) .

Wishing to also have a more powerful output valve with similar sensitivity, the EL6 was developed simultaneously and identically to the steel-based Telefunken EL12. Needing an input signal of 4.8V, the EL6 could be driven with normal circuits, which were intended for the EL3, whereas the older EL5 needed 8.2 V.

In class A operation, the EL6 gave an output power of 8.2 W.

In a push-pull circuit, for which it was not really intended, it only gives 14.5 W, clearly less than the EL5 (19.5W).

The EL6 was further developed to the 4699, which could supply more power output by operating from a higher supply voltage.

Double output pentode for car radios for push-pull operation.
Released July 1938.
2 x 4.5 Wa, -20 V, gm 1.8 mA/V; Ih= 0.45 A.

The ELL1 is the first push-pull output valve with two small output pentode structures in one envelope and was meant for car radios. With half the heater current of the EL3 and a smaller quiescent current (30 mA) a rather larger power output (5.4 W) could being obtained than with the EL3. This was, however, with substantially smaller slope, so a grid signal of 2x19V became necessary, which with to the technology of that time had to be supplied by a push-pull driver transformer.

A direct descendant of the ELL1 did not exist. Only many years later did the ELL80 appear, which however was meant more for domestic apparatus (stereo).

The comparable competitive Telefunken type was the push-pull output triode EDD11, which like the 6N7, was operated with grid current (class AB2) and above all operated at a substantially lower quiescent current (7 mA.)

1882 and 1883

Philips used the 18xx series for various different vacuum rectifier valve types, e.g. 1802 = RGN354 (Telefunken), 1805 = DW4/350 (Mullard) = RGN1064 (Telefunken).

At first sight the affiliation of the types 1882 and 1883 to the Red series is not obvious.

In their distribution area in Western Europe, particularly in France, they were used exclusively in conjunction with the red series, therefore one must add them here.

So far, only rectifier valves with 4-V heaters (AZ1...4) or 6.3 V (EZ1...4) were intended for the Red series.

In European countries with free valve markets however, there were also many rectifier valves with 5-volt filaments and heaters following American practice.

To make it easier for the set- and transformer-makers, the types 1882 and 1883 rectifier valves with 5-V filament or heater (see below) were now added to the red series, so one could now universally use the same mains transformers as for American valve sets (Octal).


Full-wave rectifier valve, directly heated, replaces AZ1;
released June 1938.

2 x 350 V 125 mA / 2 x 400 V 110 mA; filament 5 V 2 A.

Apart from the side-contact base, it corresponds to the American types 5Y3G and 80. The historical chance to designate this valve 1880 instead of 1882 following its origin with the 80 was missed. Another type with the designation 1880 was never seen. Since also the 5Y3G was derived from the 80, now there were three equivalent rectifier valves 80, 5Y3G and 1882 to chose from, with American 4-pin (UX4), octal or side-contact base respectively.

The 1882 could not become generally accepted against the successful indirectly heated 1883 and was discontinued after 1945.


Full-wave rectifier valve, semi-indirectly heated, side-contact counterpart to the octal 5Y3GB.
Released June 1938.

2 x 350 V 125 mA / 2 x 400 V 110 mA; heater 5 V 1.6 A semi-indirect.

As already described with EZ3 and EZ4, directly heated rectifier valves cause the problem, in which they conduct earlier than the other, indirectly heated valves in the receiver, giving harmful overvoltages in the meantime.

As with the existing EZ3 and EZ4 6.3-volt valves, the 1883 solves this problem in a rectifier valve with 5-volt heater. In order to make the 1883 interchangeable with the 1882, it was designed to have the same internal resistance as the 1882 despite indirect heating.

For this, the electrode structure had the same anode cylinders as in the directly heated 1882 and simply had indirectly heated cathodes fitted.

The heating was however only "semi-indirect", because the cathode is connected to one end of a heater, in this case on the left side of the base (inconsistently, because the cathode normally on the right-hand side.) This saved having good heater/cathode insulation. Later the construction of the 1883 was changed, as was the equivalent and identically-constructed octal valve, the 5Y3GB, by arranging a particularly formed plate for each half wave around and to opposite sides of a single common flat cathode. Besides this, both the 1883 and the 5Y3GB were supplied in a number of different versions, which differed in the electrode structure or envelope.

Although both valves 1882 and 1883 came out together in June 1938, because of their advantages the 1883 was much the more widely accepted. The 1882 is not found in equipment manufactured in France after 1945.

In German data books the indicated designations RGN1882 / RGN1883 were chosen exclusively by Telefunken and have no general validity. Against some 100 or 1000 replacement valves, which Telefunken supplied under these designations, millions of others also exist with the original designations 1882 and 1883.

This article was edited 25.Jun.05 17:30 by Jacob Roschy .

Jacob Roschy
Jacob Roschy
D  Articles: 1568
Schem.: 17
Pict.: 68
21.May.05 09:43

Count of Thanks: 68
Reply  |  You aren't logged in. (Guest)   3

Triode-hexode for frequency changing,
released approx. Summer 1939.
gc 0.65; If = 0.2A.

Only few months after Philips brought out their gigantic triode-heptode ECH2 and the hardly less gigantic and highly-complicated EK3 "four-beam" pseudo-octode, a new generation of FC valves, with the ECH11 from Telefunken in Germany was developed, which outperformed all previous FC valves.

Philips could only be pitied with its monster FC valves ECH2 and EK3 compared with this new generation of valves, and was obliged to adopt the electrode structure design of the ECH11 in the form of the ECH3 into the Red series. From now on, the triode-hexode ECH3 was exclusively used as the FC valve in the Red series.

Magic eye with dual sensitivity, replaces EM1 and C/EM2,
released July 1939.

All past tuning indicator "eyes" lacked the ability to indicate clearly the entire signal strength range from very weak to very strong. If they were supplied with the full (detector) diode voltage, they were already fully closed with medium-strength signals and strong signals resulted in no further change in the display. If the control voltage was decreased so that it closed fully only with the strongest signals, there was hardly any deflection with weaker signals.

The problem was solved with the construction of the EM4. In contrast to the EM1, which had four deflection rods, which operated in parallel to give the same deflection, the EM4 had only two deflection bars but with different deflection sensitivity. This was achieved by having an amplifier section under the fluorescent screen, which has a cathode common with the fluorescent screen, a grid with different slope and two differently-sized anode plates. These two anodes form two triode systems with different amplification due to the different grid turns. These anodes are connected to the deflection rods. The sensitive system now enables the indication of weak signals, while the other shows nearly no deflection, whereas the less sensitive system still shows clear changes in deflection on strong signals if the previous one is already overloaded.

Continued at part III

This article was edited 25.Jun.05 17:31 by Jacob Roschy .