This discussion is locked.
You cannot post a reply to this discussion. If you have a question start a new discussion

Early issues with rectification off AC mains

In 1949 a lighting dimmer was invented that used 3 phase, phase-controlled, half-wave rectification by means of 3 thyratrons. Its output was pulsed DC and was very effective in dimming incandescent lamps. It however was a prodigious generator of harmonics with damaging neutral currents, and had a poor power factor even at full output.

Nowadays such a product would be banned by supply authorities and product regulation, but things were much laxer then. However problems caused by large DC loads such as mercury arc rectifiers on AC supplies, would have been experienced by many practitioners in the 1950s. Can anyone point me to UK standards or custom and practice from that era that might have been known and gave guidance on the problems caused by rectifying AC loads?

  • I have some experience with mercury arc rectifiers, having restored a pair of 1930’s rectifiers with a colleague, and which are now on display operating at Kempton Steam Museum on steaming weekends ( http://www.kemptonsteam.org/history/arc-rectifiers/ ).  With appropriate configuration of the supply transformer Mercury arc rectifiers need not generate as much interference as you might think.

    My experience is with 6-phase rectifiers.  The 6-phase supply is normally derived from a 3-phase transformer with delta connected primary and two star connected secondarys in antiphase with each other.  Connecting the star points together gives a 6-phase supply to the 6 arms of the rectifier, the star points being connected to the rectifier cathode.  Each rectifier anode conducts in turn, carrying the full load current.  Unfortunately this results in a relatively high third harmonic content in the primary current.

    To overcome this the two star points are connected via a transformer similar to a centre tapped autotransformer, each star point connected to an end of the winding, the rectifier cathode to the midpoint.   The transformer is magnetized by the current flowing in the individual secondary windings, producing corresponding voltages which tend to equalise the voltage between the two groups of phases.   The effect is that the voltage at each of the anodes is the instantaneous average of the two secondary groups. The result is that at any time, the voltage at two of the anodes is similar (having been averaged) and thus two anodes conduct at any time, carrying half the load current. Thus the third harmonic is not generated.

  • David, thank you very much for the information, so prompt and in fact so near me. I must visit soon!


    I was aware from 1930s text books [eg "Mercury Arc Power Rectifiers, Theory and Practice", Marti and Winograd] about the technique of using an inter-phase transformer for a 6 phase system to shift the phase between the two groups of 3, and also add reactance to improve current sharing. My concern is about what jobbing electrical engineers post war (1945-55 era) would or should have understood about the challenges of rectification on the mains, and especially the accumulating triplen harmonics and DC bias. I can find learned texts for the consulting/professional engineers, but no standards or published guides (the 1950 IEE 12th edition is silent) about the pitfalls awaiting the naiive. May I ask where your expertise came from?
  • Former Community Member
    0 Former Community Member
    By 1955 television sets would have been growing in number, each having a half wave rectifier straight off AC mains. In this respect, this could have been a considerable problem. Manufacturers were not forced to have a transformer and two diodes. A bridge rectifier off the mains without an isolation transformer would have safety issues over and above the already dangerous "live chassis" designs used.
  • David, the next public weekend at Kempton is on the 22nd September.  The engine is well worth seeing for it’s massive scale.

    I am an electronics broadcast engineer by career, my expertise on mercury arcs came from a number of sources, mostly textbooks from the 1920’s, handbooks from the Hewitic Company (Hewit was the inventor of the mercury arc rectifier, and his company was a major manufacturer of rectifiers, based in Walton on Thames), various internet sources on specialised things such as ultra violet radiation, but above all the experience of completely dismantling and refurbishing our recifiers.  There appear to be very few people around now that know anything about them.  I have been contacted several times for information from around the country and abroad.

    Incidentally, one curious feature of the interphase transformer is that at low current (say >3 amps, there is not enough flux in the transformer to be effective, so the device works as a 6-phase rectifier (dc out 1.35 times the rms secondary voltage).  At higher currents the transformer is in operation, the device works as a double 3-phase rectifier (dc out 1.17 times the rms voltage).  So as the load current increases the output voltage suddenly falls.

    You might also try posting on the IET Wiring and the Regulations site, https://www.theiet.org/forums/forum/categories.cfm?catid=205&entercat=y .  There are many experts both on regulations, standards, mains-borne interference.

    John, I suspect that the effect of television sets in the 50’s may not have been particularly noticeable. Many would have been powered from non-polarised two-pin plugs, and the line and neutral connections may not have been particularly consistent, balancing out the conduction cycle.  From memory though, many were fitted with valve or selenium rectifiers, and I think these had a higher forward resistance, so spreading the current over a longer conduction period.  I can recall 60’s televisions with the chassis at half mains, and that might imply they were bridge rectifiers.

  • Hi Gents, sorry for any truncated replies as I was having problems with my new password.


    Standard textbooks from the 30's relate to 6 phase (with/without reactor( for large plating plants and possibly some drives applications. The more advanced work will analyse the systems and show low PF issues, but I think this was just considered a cost penalty.

    We do have a static machine set up as a phase advancer in the Discovery Museum, but no details of what it operated with.

    The growth of TV's as has been pointed out. would, with the advent of AC/DC (transformerless) sets from the early 50's cause poorer power factors with the 1/2 wave rectification being used in place of the full wave of transformer fed sets.

    This was made worse with the advent of colour sets that had regulated power supplies driven by 1/2 wave ,phase angle fired SCRs. It was noted in some of the trade press that this was not a good thing.

    At this time there was also the advent of phase angle light dimmers!!!

    Total of all these loads was probably only a small proportion of the full connected load.

    Again it was considered a cost penalty and cables were well rated, so not too much of a problem. There may be some references in guidance notes issued by local boards to large industrial users, but I have not seen any; certainly as has been pointed out there was nothing in the then current regs.

    On a slightly different angle, the effect was well known as by the 60's I think fluorescent light units were fitted with PFI caps.


    There was an incident in the 70's at a local factory where a 1MVA 11Kv/440 transformer blew up .

    It fed a spot welding shop where the labour force was keen to prolong jobs, and thought that by setting the cycles control on the spot welder to max, then adjusting the phase angle control to give a satisfactory weld they would have an easier time! The computed PF after the accident was pretty poor and the triplen effects were large.


    From memory the regs started to come in in the 80's when a whole sheaf of EEC rules started to appear that related to EMC.

    The climax of this appeared in Airbus Electrical directives for A400M in the 2000's. Not only were the alternators frequency wild(380 to 620Hz), but emc/ power quality standards demanded 18 phase rectification to feed the PFI correction circuits of the various electronic PSU's being designed.

    It was an interesting time.


    Ed