Philips High Impedance Speakers

The use of high impedance speakers in certain Philips sets in the late 1950s was an unusual and brave step. It is exactly the sort of odd-ball (although generally well-intentioned) engineering that collectors have come to expect from this company!

It certainly gives repairers something else to think about, and I feel it is appropriate to give the subject its own page.


High Quality Sets

To get us started, this email from Jan Willem Wegener gives a good deal of background information:

I just came across your (very interesting, by the way) repair story of a Philips B7X14A65 �Reverbeo� receiver where some questions about the use of high-impedance speakers were raised. I have a small number of Philips receivers with these speakers, including the �Reverbeo� model, so I can provide you with some background information.

Philips experimented a lot with so-called transformerless outputs, in order to arrive at a better, i.e. more competitive, quality-cost ratio. Their idea was to use an EL84 and an EL86 tube in a push-pull circuit vere akin to the NPN-NPN asymmetrical transistor output stages. With an HT of 320 V both tubes were in series for DC. Patents have been described and circuits have been published in the scientific literature.

The whole idea was put into practice in about 1958, with some beautiful �Bi-ampli� receivers (�Bi-ampli� stood for two output tubes, whether in single channel push-pull or in two separate channels for either fake or real stereo). It was stopped soon after, however, due to the mere fact that the extra costs for a second output tube and the increased HT didn�t compensate for the costs of an output transformer!

The positive experience with 800 Ohm speakers was subsequently being used for circuits with transformers having a ratio of about 2.5 between primary and secondary. Good transmission properties can be much more easily achieved at low coil ratios, hence this compromise between �classic� low impedance output stages and truly ironless stages.

Don Bouwhuis sent the following information and observations in relation to my Philips Reverbeo repair:

It is with some interest that I read your report on restoring the Philips Reverbeo, in which you commented that the speakers were of the 800 ohms type. You considered this strange. As it happens there were interesting reasons for Philips to do so.

For a long time Philips produced most of its speakers as 400 and 800 ohm types. The reason was that Philips tried to get rid of the output transformers, which caused losses and were expensive to produce. With the advent of the U valves, for which no mains transformer was needed, the output voltage of the end valves was also suitably lower, so, with a suitable circuit topology the loudspeaker could be driven directly by the push-pull output stage.

Of course the impedance had to be quite high, but they succeeded in winding the required coils on a mass production basis. The first experimental types had even 1500 ohms, with a centre tap and a small series was even produced as part of an existing radio type. As in the original design the output transformer was missing, Philips called these designs "direct energy transfer" and their first Hi-fi products all featured 400 and 800 ohm designs. In those days the sound difference between a traditional design with transformer and one without appeared to be an audibly better low frequency response.

For whatever reason Philips also reasoned that separate amplification of low and high frequencies was necessary for hi-quality sound and fitted special radios with two amps and called it Bi-Ampli. The Reverbeo was such a radio, and it was the top of the line of the so-called "plano" series, radios with the speakers on either side, and being decidedly flatter than regular boxes. They were also intended to provide stereo sound from the first stereo gramophone records. At that point in time, FM stereo was only a vague, and mostly unlikely possibility.

The only thing I never understood was why eventually the signal way to the loudspeakers could not be made entirely "ironless", as they called it in the Netherlands. My parents had a plano radio which I liked to hook up to my stereo, and I never found out exactly what they did, and what they were there for as the idea was to get rid of them entirely in the first place. Maybe they ran the signal from a single ended stage and wanted to get rid of the DC, I don't know.

Nevertheless one would at least expect a much smaller transformer than the regular ones, which is what you found. I remember that I myself found this output iron always suspiciously small, so there must be something funny with them. Weren't they 1:1?

As Jan mentioned above, the speakers were not 1:1, since this would not give the correct match to an EL84 valve. Jan suggests anout 2.5:1, which seems reasonable.

Gerard Tel has more information about the transformerless output stage on this page of his website. He also has a few sets using the circuit in his collection.


Low Cost Sets

At the other end of the scale, Philips used high impedance speakers connected directly in the anode load circuit of the output valve in a few low cost sets such as the Philips B2G25U, the Cossor CR1202U and the Stella ST113U.

These sets use a UCL82 output valve. The DC resistance of the speaker is about 700 Ohms, which similar to the primary of an output transformer, so the DC operating conditions of the valve are reasonably OK.

However the AC impedance of the speaker is around 800 Ohms whereas the UCL82 is rated for an anode impedance of 3.9k (at 100V) to 5.6k (at 200V), so 800 Ohms is a long way out. In theory this should result in significant amounts of harmonic distortion.

Another consideration is that the speaker has to carry the full DC anode current of the valve - typically 40mA. This causes significant cone displacement, and in my experience the speaker should be connected so that the cone is displaced outwards by this current.

Despite all this however, the set actually sounds better than the equivalent models (Philips B2G81U, Cossor CR1201U, Stella ST112U etc) that use a conventional speaker and output transformer. This may be due to the use of a poor quality output transformer used, but it does show that the output transformer has a large effect on the sound quality, and justifies Philips' attempts to eliminate it from the signal chain.


Reliability

A major concern is the reliability of these speakers. To achieve this sort of impedance there is thousands of turns of extremely thin wire, which is not going to be particularly durable, particularly since it is being shaken around by its own vibrations. At the low end of the market we have the 40mA standing DC current too. And then the effects of damp due to poor storage....

It is generally claimed that reliability was a real problem with them back in the late 1950s and early 1960s. I have only encountered two sets with these speakers so far (the Philips Reverbeo and the Stella ST113U), both of which had good speakers. However it would be prudent to check the speakers in any of these sets before doing anything else, since direct replacement is probably impossible.

If the speaker is faulty, the easiest solution would probably be to fit normal conventional speaker, and add an output transformer somewhere. A speaker with the output transformer attached would be a tidy solution. For a single-valve output stage the connections are straightforward, but more thought may be needed if the set has the two-valve circuit shown above.


If you have anything to add to this sujbect, please contact me. Thanks.




This website, including all text and images not otherwise credited, is copyright © 1997 - 2006 Paul Stenning.
No part of this website may be reproduced in any form without prior written permission from Paul Stenning.
All details are believed to be accurate, but no liability can be accepted for any errors.
The types of equipment discussed on this website may contain high voltages and/or operate at high temperatures.
Appropriate precautions must always be taken to minimise the risk of accidents.

Last updated 14th April 2006.