The normal Mk1 coach dynamo does not reach charging voltage until 25mph or so (30mph or more for WC type dynamos) has been reached. This is no use to the majority of preserved lines whose speed limit is 25mph.  However, even though the voltage output of the dynamo is only about 12volts at 12mph, the current capability is quite large. A possible solution to the problem is to design an electronic inverter which "kick-starts" the dynamo up to charging voltage at around 12mph.

The inverter is conveniently sited inside the existing MD Regulator housing.  There is a reasonable amount of space in the housing, and weight is of no concern, so a low frequency design is probably the cheapest and simplest. The (very large) prototype circuit diagram is shown here. You may have to resize the image to view it properly on some browsers. Use the ZOOM icon in the lower RHS corner to make the drawing intelligible.

How did it work?

The dynamo starts to generate volts when it turns. This is due to residual magnetism in the field pole pieces. As volts are generated, current flows through the field coils, thus perpetuating the effect. At about 12mph, the following is the case:
If the armature generates 12v, then the field has only 12v across it, so only 12v is generated.
However, if the dynamo has previously been charging and has slowed down to 12mph, the dynamo is still charging (just)!
So if the dynamo can be fooled into thinking it has been charging, it ought to be possible to get charge into the batteries at any speed above 12mph. Hence this circuit.

At a dynamo output of approximately 12v, the inverter starts up. This produces a DC voltage of 10-12v negative at the bottom end of the field coil. Thus the total voltage across the field coil rises to 24v. This immediately means that the armature voltage will also rise nearer to 24v. The inverter will then try to transform this higher supply into even more volts across the field coils, but the armature voltage will then exceed the battery voltage, and the MD regulator will pull in and connect the dynamo to the battery. The same relay also powers up the regulator coils, and this action shuts down the inverter. The inverter then stays dormant until the dynamo charging drops out because the speed has dropped to 12mph or less. It will then restart and try to reconnect the dynamo, but if the speed is insufficient, it will not connect. It will keep trying until the speed drops to 6mph or so. Then the inverter shuts down until it next sees 12v output from the dynamo.

The unit was installed in the MK1 brake of the regular service train and was tried out at the beginning of April 2000. It worked!!!! Some modifications and improvements were immediately obvious, but it made the dynamo charge at well under 20mph. Modifications have been implemented and have been tested on RMB 1866. This coach had a "Keighley" style solid-state regulator paired up with a WC dynamo. The design of the regulator is not compatible with a WC dynamo, so the dynamo was changed in between service trains one Sunday lunchtime in July 2000. Quite a feat to change out a dynamo in 45 minutes!
With the new dynamo fitted and the regulator repaired, the inverter worked well and the buffet coach now had a working charging dynamo. This has helped reduce the requirement for charging the coach every night between trains during the operating season as the refrigerator consumes about 20A and, combined with the lighting load, will easily flatten the batteries within a day.

These inverters were available from the GoToPCs Catalogue and over the years several improvements to improve reliability and robustness were incorporated. Unfortunately, due to the RoHS directive from the EU, semi-conductors became not quite so robust and these units became so difficult to make that production ceased. The biggest barrier to reliability is the enormous voltage spike which occurs when the dynamo pull-in relay operates or breaks - this spike is in the thousands of volts sometimes.