Jump to content
Sign in to follow this  
Tony S

Mercury Arc Rectifier #1

Recommended Posts

Tony S

Mercury Arc Rectifier (MAR)






The mercury arc rectifier was invented by Peter Cooper Hewitt in 1902 as a development of the hot cathode thermionic valve. It was further developed throughout the 1920s and 1930s. Before the advent of solid-state devices, mercury-arc rectifiers were the most efficient of rectifiers. They were far more efficient than rotary converters as they had no mechanical losses other than the cooling fans. They do have a fixed volt drop across the rectifier just as a modern diode does so really in a comparison it should be disregarded.



The Manx Electric Railway still used them until a few years ago at the Laxey station.




Some good news regarding the Laxy  rectifier station, the Manx government have stept in to save it.



Mercury arc rectifiers were used until the 1980s for large industrial DC supplies including trams, electric railways and anywhere where large variable speed drives were used.





The anodes are graphite moulded to a tungsten lead out terminal.

The cathode, the pool of mercury with a tungsten lead out.

Ignition electrode, a tungsten wire with a magnetic armature crimped around it.

Ignition choke reduces the current to the ignition electrode.

Ignition coil pulls the ignition electrode down in to the cathode pool, in doing so it shorts itself out allowing the electrode to spring back.

Exciter anodes, again graphite moulded to tungsten lead out terminals.

Exciter current coil, once excitation current is established the ignition electrode is taken out of circuit to stop it being burnt out.


Where things look odd in the drawing is a red phase wire going to a blue anode. The anode is only taking the negative half of the red phase AC current. Up to the bulbs anode connection everything is 3Ph AC.




The transformer:


A strange beast to say the least. The star point forms the positive of the DC circuit, the star point is actually an secondary interphase transformer/choke. Its function is to allow anode phase overlap easing the stress on the transformer and bulb. Without it only one leg of the transformer is in use at one time, electrically it’s OK but it knocks the stuffing out of the windings and core.




A, B, C primary windings

a1, b1, c1, a2, b2, c2 are the secondary windings feeding the anodes.

Ex1, ExN, Ex2 exciter transformer

Ig, + ignition transformer



Now if we put it all together as a working unit





The one type and the one that you will most likely find in transport museums now is the glass bulb rectifier. The rectifier consists of an evacuated glass bulb with a pool of liquid mercury sitting in the bottom as the cathode. The large glass envelope condenses the mercury that is evaporated as the rectifier operates.


The envelope has two or more arms with graphite rods as anodes. (My drawings only show two arms, but the connections are for six). The number depends on the application. If direct current is to be produced from a single-phase, then two anodes are used, each connected to the outer ends of a centre-tapped transformer. With three-phase current three or six anodes are used to provide a smoother output current. Six-phase (see the transformer drawing) operation can improve the efficiency of the transformer as well as providing smoother DC current. During operation, the arc transfers to the anodes at the highest negative potential.


The shape of the arms and envelope is intended to prevent an arc from forming between the anodes; this is "backfire" and is a critical factor in the design of the rectifiers. If this happens then the fuse for the affected arm blows, without the fuse the anodes would just blast away at each other until they destroyed themselves and the bulb. 9 times out of 10 the rectifier would just carry on working until someone happened to notice an inactive arm. This would cause havoc in electroplating so phase failure relays would be fitted.



Glass envelope rectifiers can produce hundreds of KW. A 6-phase rectifier rated 250 amperes has a glass envelope approximately 30 inches high by 18 inches outside diameter. These rectifiers will contain several pounds of liquid mercury. The large size of the envelope is to dissipate heat and condense the vapour so as to return it the cathode pool. Every part of the bulb slopes downwards to allow the mercury to flow back to the pool.



For larger rectifiers, a metal tank with ceramic insulators for the electrodes is used, with a vacuum pump system to combat slight leakage of air into the tank. The entire tank is live and so has to be mounted on insulating posts.





Both glass and metal envelope rectifiers can have control grids inserted between the anode and cathode. This allows the rectifier to be controlled, for example to delay the instant at which the arc transfers to the anode on the alternating current waveform, thereby giving control of the mean output voltage produced by the rectifier. (I’ll cover this in “the mercury arc thyratron” later)


The temperature of the envelope must be carefully controlled. The units I worked on had a series wound motor driving a fan, the supply coming from a shunt in the cathode lead. As the load increased so did the speed of the fan.

It’s quite funny to watch a MAR working. A motor on the plant would start, the envelope would glow brighter and the fan would rev up.






Operation of the rectifier relies on an arc discharge between electrodes in the envelope containing the vapour. A pool of liquid mercury acts as the cathode.

The mercury emits electrons freely, whereas the carbon anodes emit very few electrons even when heated, thus rectifying action occurs.

Once an arc is formed, electrons are emitted from the surface of the pool, causing ionisation of mercury vapour along the path towards the anodes. The mercury ions are attracted towards the cathode, and the resulting ionic bombardment of the pool maintains the temperature of the “emission spot”, so long as a current of a few amperes continues. On low load you can see a circle traced on the surface of the pool caused by the emission spot traveling around the surface.


The mercury ions emit light at characteristic wavelengths. At the low pressure within a rectifier, the light appears pale blue-violet and contains much ultraviolet light. I don’t suppose I should have spent so much time looking at them working, but it was like watching a living thing, absolutely captivating.


The cathode is connected to the DC load, which in turn is connected to the centre tap of the transformer. Each anode arm is connected to a leg of the transformer. As the voltage on each anode goes negative, it will begin to conduct through the mercury vapour to the cathode. As the anodes of each phase are fed from opposite ends of the transformer winding, one will be positive, and the other negative, and thus a current will always be maintained from one or more positive anodes to the cathode.



Starting a mercury-arc rectifier is by a brief arc within the rectifier bulb between the cathode pool and the ignition electrode. The ignition electrode is brought into contact with the pool and allows a low AC current to pass current through an inductor. Once current flows it shorts the ignition coil out allowing the electrode to lift and create an arc, if the exciter electrodes fails to fire the coil energises again and the cycle starts again. This little arrangement is known as the “tickler” due to its high-speed vibration. Once current is flowing in the exciter circuit the current relay opens the ignition circuit. Occasionally in cold weather and low load the emission spot will go out, with no exciter current the current relay drops out energising the ignition relay. By this the rectifier is self-restarting. Although the ignition electrode is carrying AC when in contact with the cathode it becomes DC the moment the arc is formed.

There is another way to get the excitation arc started but it’s not recommended, shake the bulb until some mercury splashes in to one of the excitation anodes. I’ve seen it done once, the transformer was not happy.





The ignition coil pulls the start electrode “the tickler” in to the mercury pool. This short circuits the ignition coil allowing the electrode out of the mercury pool causing a spark and forming an emission spot. The process is repeated until sufficient vapour is created to allow the excitation anodes to maintain their own emission pool. Constant excitation is needed to maintain the emission spot while the main anodes are idle.





Three wire supply


If a three wire centre neutral supply is required then it’s simplicity itself, connect the +ve of one rectifier set to the -ve of another. It does away with the need for mechanical balancer sets. The centre wire would usually be earthed to give a +ve→N→-ve   supply that doesn’t vary with imbalanced load.











For higher power three wire systems a “ganged” set of rectifiers is used. A main rectifier supplies the +ve and -ve, two smaller sets balance the load on the centre neutral.





Making a two leg MAR, the glasswork is amazing.





© Tony S

View full knowledgebase

Share this post

Link to post
Share on other sites
Sign in to follow this  


Important Information

Terms of UseWe have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.