A Guide to Ferrocart Vibrators


The vibrator at the bottom left is the later crimped can 150 cycle Utah type; the others are the older 100 cycle types with soldered cans and use either an Oak or Utah mechanism.

Manufactured by (or for) Radio Corporation/Electronic Industries in Melbourne, these were found in their products such as Astor and Air Chief branded radios.



                                                                                       Ferrocart Vibrator Application Notes

The following text is taken from "Engineering Data on Vibrators" by Ferrocart A/Asia Pty. Ltd. The text appears to be a condensed version of the Mallory "Fundamentals of Vibrator Power Supply Design" handbook. This might be suggestive of Ferrocart  manufacturing the Mallory design, but the internal mechanisms are actually made by Utah or Oak. In any case, it is known that other manufacturers used the Mallory literature; for example the Radiotron Designer's Handbook authored by AWA; a company who manufactured the competing Oak vibrator.

1. GENERAL
 Although vibrators of one form or another have been in use for many years in telephone exchanges, and other similar environments, it was not until the development of automobile radio receiving sets that compact and relatively inexpensive vibrators were produced capable of withstanding the wide fluctuations of battery voltage and mechanical jarring found in a modern automobile.  Not only do modern vibrators operate in any physical position, function over a wide range of conditions and give long life, but they are quiet both mechanically and electrically. All of the synchronous and all of the non-synchronous vibrators are identical in construction except that a different driving coil is used for each voltage, and different numbers and arrangements of prong bases are used.

2. NON-SYNCHRONOUS VIBRATORS
This type of vibrator is also called "Single" or "Valve Type" since it has a reed vibrating so as to make alternate contact with a single contact on either side, and hence requires a separate rectifier to produce direct current for high potential supplies used in battery operated radio receivers. They are intended for use with a full-wave or center tapped primary winding of a step-up transformer.  The reed is energized by means of a small electromagnetic coil which acts on a magnetic armature mounted on the free end of the reed.  The coil is connected electrically between the reed and the fixed contact which closes when the reed is attracted by the coil. Thus when the starting switch is closed, the vibrator coil is in series with one half of the transformer primary winding. The resistance of the vibrator coil is high compared to that of the primary winding, so that no appreciable effect is produced at this instant in the primary winding.  However, the vibrator coil attracts the reed armature, closing the initial or "starting" contact thereby short-circuiting the coil.  This creates a direct path for battery current to flow through the primary winding. The momentum of the reed keeps the initial contact closed for a time, and then the elasticity of the reed causes it to swing back, open the initial contact and close the second or "rebound" contact.

3. WAVE-FORM
When the primary winding of the transformer is connected directly to the battery, a counter electromotive force is induced in all of the transformer windings, which is in opposition to the battery potential, in
the primary winding. The induced potential remains practically constant as long as the contacts remain closed.  When the contacts open, the induced potential in the transformer windings starts to reverse. However, the rate of reversal is controlled by a condenser usually connected in shunt to the high potential secondary winding, sometimes called a "buffer" condenser. This condenser usually is given such a value that the induced potential in the primary winding has reversed but has not yet equalled the battery potential by the time the alternate contacts close.  Since the direction of current flow around the transformer core is reversed when the alternate contacts close, the counter electromotive force during the second half-cycle will have a polarity opposite to the first.  The result is that the wave-form of potential in all windings consists of a series of flat-topped half-cycles of alternate polarity.  Each flat-topped wave is connected to the following one by a sloping line terminating in an abrupt
voltage change just as the contacts close. The slope of the wave between flat-tops, that is, while both sets of contacts are open, is controlled by the size of the buffer condenser.

4. RECTIFICATION
When current of this wave-form is rectified by a full-wave rectifier of any type, a series of current impulses is obtained, each having the characteristic flat-topped wave shape, but all of the same polarity. This
is passed through a smoothing filter consisting of an iron cored reactor or choke, with a filter condenser, usually electrolytic, connected across the circuit at both input and output ends of the filter reactor. The output current and voltage from the smoothing filter is quite steady and contains negligible ripple if the reactor and condensers are of proper values. In the case of automobile radio receivers, the ground or common electrical point of the receiver is connected to the low potential battery, and is negative with respect to the high potential required for the anodes.  If a hot cathode rectifier is to be used, the cathode must be at a potential several hundred volts positive with respect to the battery, which is the best available source for cathode heating current. To meet this problem, overseas engineers introduced the first indirectly heated cathode rectifier for automotive use, in which there was sufficient insulation between heater and cathode to permit a potential difference of several hundred volts between them.  Thus the heater is operated from the battery, but the cathode operates at full positive "B" potential. From this original rectifier developed the present 84 or 6Z4 type.

5. SYNCHRONOUS VIBRATORS
Another method of rectifying the output of the vibrator transformer is to add a second set of contact points on the reed to engage with a second set of fixed contacts. Such vibrators are called "synchronous" since primary and secondary contacts operate in synchronism, also "double" since there are two complete sets of full wave contacts, and "tubeless" since no rectifier tube is required.  The secondary contacts are adjusted to close after and open before the corresponding primary contacts, to prevent destructive arcing. This results in an advantage over the non-synchronous type of vibrator, in that the primary contacts open and close at times when the transformer is disconnected from its load.  The transformer in a no-load or idle condition draws the relatively small exciting or magnetizing current from the battery, so that the primary contacts operate at moments when they are carrying very little current.  This prevents appreciable arcing at the primary contacts. On the other hand, the secondary contacts are not required to open or close with a large difference of potential across them, since the input condenser of the smoothing filter retains nearly its full charge during the interval that it is disconnected from the transformer secondary winding, and the prior closing of the primary contacts produces the full no-load potential of the secondary winding before the secondary contacts are brought together. As soon as the secondary contacts close, the secondary voltage drops from the no-load to the full-load value, which is not much lower if the transformer is designed to have good voltage regulation.
When the secondary contacts reopen, the secondary voltage rises again to its no-load value. Thus the secondary contacts operate at times when very little difference of potential across them exists.  By the time the primary contacts open the secondary contacts have separated far enough to prevent a spark from occurring.

Synchronous vibrators therefore have several advantages over non- synchronous vibrators: They eliminate separate rectifiers while costing no more than non-synchronous vibrators which they equal in external
dimensions; they are more efficient, since they eliminate the power required to heat rectifier cathodes, and also the space potential drop inside the electronic rectifier; and they will handle relatively large amounts of output power with less deterioration than non-synchronous vibrators.
 

6. SPLIT REED SYNCHRONOUS VIBRATORS
Radio receivers with output tubes having directly heated filaments present a special problem in connection with the grid bias for the output stage. Unless a bias battery is used there is no way to obtain a potential more negative than the negative end of the filaments using an ordinary synchronous vibrator, since the moving contacts of the secondary circuit are electrically common with the primary reed contacts, which in turn are connected to one side of the battery. To meet this difficulty, the split-reed synchronous vibrator was developed.  It differs from the normal synchronous vibrator in that the reed is divided longitudinally, each section carrying a set of contacts electrically insulated from those of the other section.  The armature is mounted on the free ends of the two reed sections by means of small insulators, while the fixed ends of both reeds are insulated from one another and from the frame.   The circuit is arranged so that the secondary reed is negative, and is returned to common or ground through a resistor which is by-passed by a condenser.  The potential created across this resistor by the "B" current flowing through it is then used for grid bias.  The design and operation of split-reed vibrators is otherwise the same as ordinary synchronous vibrators.
 

Elimination of Vibrator  Interference.

1. GENERAL
The introduction of a vibrator into a radio receiving set for the purpose of obtaining a high voltage B supply from a lower direct current supply such as a storage battery, at once raises problems concerning the inter- ference such a vibrator causes due to interrupting a direct current at a constant rate. These problems are entirely apart from such questions as mechanical vibration transmitted directly from the moving elements of the vibrator to the radio set. The mechanical cushioning of present-day vibrators is such that this is not an important factor. Electrical interference from the vibrator may occur due to the following
kinds of action:-
(1) DIRECT PICK-UP from the vibrator circuit by unshielded coils,
exposed grid leads or the antenna lead itself.
(2) ANODE MODULATION of any of the high frequency amplifier
or detector tubes, due to improper filtering of the anode supply
voltage.
(8) HEATER MODULATION of any of the high frequency amplifier
or detector tubes, due to improper filtering of the direct current
connections to the heaters.
(4) CHASSIS-COUPLED VOLTAGE PICK-UP in any of the high
frequency circuits, usually grid circuits, due to the chassis base
acting as a common path for currents of signal frequencies, and
the interfering currents from the vibrator circuit.

2. DIRECT PICK-UP
In order to eliminate direct pick-up all high frequency coils should be enclosed in individual shields.  Grid leads should be kept as short as possible. The antenna lead should be shielded over its entire length from the point where it enters the receiver to the antenna coil itself. An effort should be made to make the mechanical design of the receiver such that all the power supply components are grouped together and should be kept as far away from the high frequency input of the receiver as possible.

3. ANODE MODULATION
Anode modulation is easy to detect and comparatively simple to cure. The simplest method of detecting this form of interference is to connect a resistance load of such a value that the power supply is operating under normal load, then supply the anode voltages to the receiver from batteries; if there is still interference, with the power supply operating under these conditions, it is evident that interference is occurring in another portion of the circuit. However, if the interference is reduced when the receiver is operated from batteries, then the high frequency choke reactor in the B output circuit, if used, is either too small, it has too high distributed capacitance, or the associated radio-frequency by-pass condenser is too small. Generally it need not be larger than 0.05 to 0.1 M.F.  The axis of the high-frequency reactor should be changed to make sure it is not coupling to either the iron-cored choke reactor or the vibrator transformer. On tube type circuits the r.f. by-pass condenser is seldom required.

4. HEATER MODULATION
Heater modulation is usually detected by operating the power supply from a separate battery.  When the power supply is obtained from a separate battery, a shielded cable should be used, grounded to the chassis, to prevent radiation of interference from this cable which might entirely mask the heater modulation interference.

It must be kept in mind that if any change is made which reduces the power of an interfering noise or signal by one-half the apparent reduction will be slightly more than detectible by the ear.  This corresponds to a change of 3 decibels in loudness, while an actual change of approximately 10 decibels is necessary to give the impression of a 50 per cent. reduction in loudness. Thus if the interference is coming equally from two sources, elimination of either one will not seem to help much, but if both sources are eliminated simultaneously, the interference ceases entirely.  The use of an output meter on the audio output is suggested, as changes of noise of much less than one half are easily detected, especially if the interference is relatively steady.

It has been found that receivers having high sensitivity may require two h.f. reactors between the battery or d.c. power supply, and the heaters. The use of the chassis as a common connection for all of the heaters is not recommended due to the chance of voltage pick-up in the chassis. This may not show up on model receivers, but in production, the resistance of the grounding may vary slightly, and cause large changes in the amount of interference caused. The heater circuit should be grounded to the chassis at only one point.  The usual method is to wire all heaters together, grounding one of them to the chassis. The heater to be grounded should be found by experimenting to find the best point, as this will vary with different designs. Care should be taken that there are no radiating loops formed by the heater circuit which might couple to some portion of the high frequency amplifier.

5. CHASSIS-COUPLED VOLTAGE PICK-UP
Voltage pick-up due to improper grounding of the power supply and high-frequency amplifier elements is the most common source of interference and also the most difficult to locate.  The simplest method of locating the source of interference is to short the grids of the tubes, starting with the output tube and determine in which stage of the amplifier the noise is originating. A common source of trouble is found
in receivers using automatic volume control. In such receivers the tuned circuits are completed through condensers by-passing the grid return to ground.  When these condensers are grounded directly to the chassis, a voltage which is developed across the common impedance between the point where the condenser is grounded and the wiping contact of the variable condenser is picked up and applied to the grid of the tube. In order to eliminate this interference the by-pass condenser should return directly to the wiper of the section of the variable condenser tuning that particular coil.  The condenser wiper should be bonded to the chassis through a piece of heavy flexible copper braiding. As a rule, it is desirable to ground the variable condenser at only one point on the chassis.

6. LOCATING INTERFERENCE IN A COMPLETED RECEIVER
In order to check for interference on a completed receiver, the antenna lead-in should be grounded through a .0002 M.F. condenser.  If the interference appears with the lead-in short-circuited in this manner, but does not appear with it open, it indicates improper grounding of the primary circuit of the antenna coil.  In some cases, this type of interference can be eliminated by returning the ground end of the antenna
coil primary to the condenser wiper.  Sometimes it will be found that there is less interference when the Automatic Volume Control condenser or the primary of the antenna coil is grounded to some point on the chassis rather than on the condenser wiper. This is due to an out-of-phase voltage being picked up and balancing out the interference, or neutralizing it.  As a rule, this method of eliminating interference leads to erratic receivers in production, as small changes in the impedance of the current paths will change the balancing-out effect a great deal. In some cases, interference has been located in the grid circuit of the first audio frequency tube, due to the ground return of the volume control being at a point remote from the tube's cathode circuit.  Where diode detection is used, it has been found that often a hum voltage is induced in the last high-frequency transformer through coupling with the power transformer. The grid lead of the first audio tube will pick up considerable interference if it is long and unshielded, or if it runs close to the power supply or heater wiring.

7. COMPONENTS
Although the general construction of vibrator operated receivers follows the lines of a.c. sets, there are certain additional considerations with regard to some of the components having to do with the vibrator circuit.

8. VIBRATOR
Practically all vibrators now supplied to the industry have their own individual shields or metal housings. The shielding housing is not essential where the entire vibrator is enclosed within a shield together with the transformer and other components recommended to be so shielded. The vibrator housing will nearly always require grounding, however, especially if the housing projects into the unshielded space of the receiver. There are several ways in which the housing may be grounded. One most common way is to make a connection inside the vibrator, between the housing and the prong connected to the reed, which in turn is generally connected to the grounded side of the storage battery or d.c. source. Another method is to omit the internal strap, and ground the housing by means of a clamp surrounding the vibrator socket, having 6 or 8 spring fingers which grip the lower part of the housing firmly. Such vibrator ground clamps can also be obtained with bent or "formed" ears which fit into an annular groove at the lower edge of the housing, thereby preventing the vibrator from working loose from the socket, even if mounted in a position other than vertical. Another method less often used, is to connect the housing to an otherwise insulated prong of the vibrator base plug, grounding the corresponding socket jack as desired for best results.

9. PRIMARY RESISTORS
For 6 volt operation, it is generally found that improved operation is obtained if a resistor of from 50 to 100 ohms is connected from the reed of the vibrator to each stationary contact, the leads being as short
as possible. The rating should be from half to 1 watt. For operation on other voltages, the resistance will vary approximately as the square of the voltage.

10. HIGH-FREQUENCY FILTER BETWEEN  VIBRATOR CIRCUIT AND BATTERY
In stationary radio receivers containing vibrators, it is necessary to place a filter between the d.c. supply and the vibrator circuits to prevent interference from coupling to the signal circuits via the d.c. supply. In
automobile receivers it is also necessary to prevent interference from the ignition system of the car from entering the radio receiver. It has been found that it is seldom necessary to use suppressor devices on the spark system of an automobile, if certain filter elements are added to the receiver, which are designed to operate at very high frequencies. From one to three air cored choke reactors are used in the battery lead to the vibrator circuit, having from 30 to 100 turns of sufficiently heavy wire to carry the current.  One form of choke which has been used satisfactorily in many sets consists of 74 turns of No. 16 A.W. Gauge wire (0.05 inch, 1.29 mm. diameter) wound with 4 layers insulated with paper, on a mandrel having a diameter of 5/16 inch (7.94 mm.). Single layer chokes are also used. When multilayer chokes are used, it is usually best to connect the inner end toward the d.c. supply; the outer end, toward the vibrator. To prevent interference from the vibrator, coadensers of approximately 0.5 M.F. are connected to ground from both sides of the choke nearestthe vibrator, if more than one is used. These must have very low power factor at high radio frequencies, and must have short leads, of low
resistance material.  The ground return of these condensers should be as short as possible and soldered directly to the chassis.  The ground connection to the vibrator reed should be soldered to the same point as these condensers. To prevent spark interference from the automobile motor, low-capacity condensers called spark plates are used, generally connected between ground and the ends of the air cored choke nearest the battery, if more than one is used. These condensers have a capacitance of from 10 to 100 mmf., usually between 20 and 50 mnif. One type of spark plate consists of a
steel plate having an area of several square inches riveted to the radio case by insulated rivets, and insulated from the case by either mica or a good grade of insulating or fish paper, to give the desired capacitance. Spark plates are not required on non-automobile sets.

11. HIGH-FREQUENCY FILTER IN HIGH-VOLTAGE OUTPUT CIRCUIT
To keep vibrator interference from reaching the anode supply circuit an air-cored high frequency choke reactor is placed between the iron-cored choke reactor and the cathode of the rectifier tube, or center tap of the secondary winding of the transformer in synchronous vibrator circuits. It has an inductance of from 0.5 to 5 millihenrys, and should be of "universal" or self-supporting construction, having low distributed capacitance.  It should be physically small to restrict its external field.

A by-pass condenser of from 0.0005 to 0.1 M.F. capacitance may be required connected between ground and the side of the air-cored choke nearer the interference. In tube type circuits, this condenser is seldom required.

In some cases, the high-frequency filter is placed on the other side of the smoothing filter, that is, between the iron-cored choke and the tube anodes.

12. SMOOTHING FILTER
The rectified high-voltage direct current is smoothed out by means of an iron-cored choke reactor shunted by electrolytic condensers much as in a.c. radio receivers. The input filter condenser may be from 4 M.F. up, and the output filter condenser from 6 M.F. up to as high as 30 M.F. if exceptionally good filtering is required. The choke usually has a resistance (d.c.) of from 200 to 500 ohms, with an inductance of from 5 to 30 henrys.
 

13. SPARK FILTERS
Besides the spark plates and high-frequency chokes in the battery leads, interference filters or condensers are required on all other leads from or to the receiver.

In the antenna lead-in, a small high-frequency reactor as small as 20 to 40 turns, 1/4 to 1/2 inch in diameter (6 to 12 mm.) is used, with by-pass condensers.  In many cases, the lead-in is of shielded wire, which acts as a by-pass condenser. The other side of the antenna choke is by-passed by a small spark plate of from 5 to 20 mmf., usually with mica insulation. Any other leads, such as to dial lamps, external controls, etc., usually require spark plates to prevent bringing in interference from the spark system.

CIRCUITS AND COMPONENTS






 
 


The PM237 and PM238 are the most common types as these were used in the Astor/Air Chief car radios installed in Holdens and Fords of the 1950's. Why they show voltage output is anyone's guess...it's the transformer that determines that, not the vibrator.



Types of Ferrocart Vibrator.
Essentially, there are four types that will be encountered.


Old type - soldered can.
The split reed synchronous types, such as PM126, use Oak mechanism which has been modified so that the driving coil operates in shunt drive. As explained here, Oak vibrators normally use a series driven coil. In the Ferrocart adaptation, the drive coil contact is present on the reed, but the adjusting screw which forms the other part of the contact is omitted. Because of the UX 6 pin base, there are not enough pins to allow for a separate drive coil connection. (MSP/Oak uses a UX 7 pin base).


The earlier type of Ferrocart is easily opened by desoldering the base. This type is actually quite reliable once the insulating film is cleaned off the contacts.


The split reed synchronous Ferrocarts use an Oak mechanism modified for shunt drive.

For the non synchronous and non split reed synchronous types the mechanism was changed to a Utah type. This by default is shunt drive. Evidentally, Utah did not make split reed mechanisms, hence the choice of Oak where this was required. Why Ferrocart didn't use the Oak mechanisms for all types is not known, given the superiority of this type. Apparently, split reed vibrators were not popular in the U.S., and Mallory literature explains there was a move away from them, due to the extra work in their construction. Oak didn't seem to have a problem with this, and as such they were very common in Australian equipment. (Maybe Oak had a patent on the split reed design, and like the series drive coil patent, Mallory and others didn't want to pay royalties).


The ordinary synchronous and non synchronous types use a Utah mechanism operating at 100 cycles. This type seems to have good reliability.

Soldered can - octal base military types.


"R.C.V" undoubtedly stands for Radio Corporation Victoria. Oak series drive mechanism is used.

This type is octal based and was used for military applications. As such it also carries a Department of Defence number. Internally, it contains an Oak mechanism, but unlike other Ferrocarts, retains the separate contact for the driving coil. As can be seen from the photo, type PM957 is a 12V split reed synchronous type. The acoustic insulation for the can is paper based. Apart from the octal base and the fact the coil is series driven, the construction is the same as the civilian types.
The use of an octal base is interesting and one might be curious as to why it was never popular for vibrator use. On the contrary, it provides far better grip in the socket, and octal sockets remained standard right up until the end of the valve era. Vibrator development began in the pre octal era which explains the common UX bases used, but there is no reason there couldn't have been a gradual change to the octal standard. After all, the likes of Delco introduced their own unusual 3 and 5 pin bases.
It should also be pointed out that the extra pins of the octal base allow for the separate driving coil connection. As the civilian split reed vibrators (eg. PM126) have only a 6 pin base, they must use the shunt drive method.

Serviceability for these early types is good as the base can be simply desoldered and removed from the can. These early models operate at 100 cycles and were used in pre 1950's radios. Acoustic insulation is by a paper based material, or the later sponge rubber.

Later type - crimped can.
Continuing on with the Utah mechanism, the later type of Ferrocart uses crimped can construction. These operate at 150 cycles. It can be seen the reed weight has been reduced in size upon close examination. Typical types include PM237 (6V) and PM238 (12V). These were made from the early 1950's onwards. To open these requires carefully turning back the crimp which is easily done with a pair of side cutters to lever up the edge. Unfortunately it is impossible to restore the smooth appearance if the edge is folded back after servicing. As the base is a tight fit at the bottom of the can, it is easiest just to leave the can uncrimped. Most times the vibrator socket in the radio is fitted with a spring clip which will hold the can in anyway.


This later type is opened by turning back the crimp. Unfortunately, the disfigurement is permanent.

A feature of all Ferrocart vibrators is that the operating frequency is stated on the label. Note that while the later non synchronous crimped can types have only four pins, the position of the pins is such that a six pin socket is required (UX6). Being shunt drive, only three of the pins are actually used.


 This photo is from a Utah advertisement in 1941. The origins are clearly seen. The photo on the right shows the insides of a genuine U.S. made Utah vibrator. This particular example is stamped with a Ferrocart type number.



Vibrator does not start..
Most times a Ferrocart vibrator will not start whether N.O.S. or already in a radio. The problem here is the shunt drive design used by Ferrocart relies on the transformer switching contacts to also switch the driving coil. The exception of course is the octal based military types with their series drive Oak mechanisms and the Oak/MSP article should be read in relation to these.

Basic shunt drive vibrator circuit.

It can be seen that when power is applied, the driving coil is energised by the current flowing through the upper half of the primary. Thus, the contact swings to the left and the coil is shorted out. With the coil no longer magnetised, the contact swings back to the other contact. The process repeats, and hence the reed vibrates.
Obviously, the condition of these contacts determines how well the reed vibrates. Since they are also switching the comparitively high transformer current, their rate of wear can be higher than if the coil had its own switching contact handling only a couple of hundred milliamps. If the contacts in this kind of vibrator are out of adjustment, damaged, or dirty, the vibrator will run erratically, if it starts at all.
As mentioned elsewhere, over a period of disuse an insulating film builds up on the contacts. The usual result is that most Ferrocart (and other shunt drive) vibrators fail to start in recently acquired vintage equipment. As the contacts are normally open, it's easy for oxide or film to build up. In the series drive vibrator, the driving contact is normally closed and uses non tarnishing silver contacts - which it can do because of the driving coil's low current.

Contact cleaning the electronic way.
For those with a shunt drive vibrator that won't start because of insulating film on the contacts, it's often possible to clean the contacts without opening the can.
The procedure is to burn off the insulation with high voltage.


Test jig for cleaning the film off shunt drive vibrator contacts.

A 30V power supply, current limited to 500mA, is used to generate back EMF across the driving coil. This eventually breaks down the film and the vibrator will start.
When the 30V is applied to the coil, the contacts close, but because of the film they don't actually connect. But when the 30V connection is broken, the magnetic flux collapses generating a high voltage which is able to puncture the physically connected, but electrically insulated, contacts.
It can take a while for the process to work, and in this regard a high voltage switching transistor could be used to do the job. A square wave oscillator operating at the resonant frequency of the vibrator could drive the transistor. It took around 10 minutes to start the last vibrator I fixed by making and breaking the connection by hand.
The non driven contact(s) also needs to be cleaned, and here a 240V supply, current limited to 400mA by means of a 100W bulb does the job. It will burn through the film almost immediately. Of course, the reed has to be vibrating to do this, and the 30V 500mA power supply can be used for the purpose. Current limiting is essential because the power supply is shorted out each time the driving coil contacts make connection.
The contacts must be cleaned in turn, otherwise once the insulation burns off one one contact, any others on that side of the reed will not be exposed to the high voltage during contact closure.
In some cases, the driving coil can be forced to vibrate by connecting it to the current limited 240V source. This depends on the vibrator and its operating frequency.  It won't vibrate properly unless the reed frequency is exactly a multiple of the mains frequency. The same 100W bulb can be used to drive the coil.
It is most important to use the isolating transformer, otherwise the power supply and vibrator will float at mains voltage. This is a lethal shock hazard as well as being a risk to the internal circuitry of the power supply.
30V might seem damaging to the coil, but it must be remembered that in a shunt drive vibrator the coil operates at twice the supply voltage because of the transformer action.
So, a 6V vibrator actually has its coil operating at 12V, etc. As the 30V is applied only for short periods, the coil will not get hot enough to be damaged. It is important not to leave the 30V connected for more than a second or so if the vibrator does not start.

Still won't start?
If the vibrator will still not start, the can will have to be opened for further examination. In one instance I found that one of the contacts had slipped out of position as the stack screws were being tightened during manufacture. It was glaringly obvious, and one has to wonder about the Ferrocart quality control, because this would have been evident when the vibrator was initially adjusted. Once repositioned, it started immediately and worked perfectly.


Note the angular wear in the contact. This vibrator had been assembled with the contact clearly in an incorrect position. Once restored, it started immediately.

Can lining.
Unlike Oak, Ferrocart lined the entire can. Where sponge rubber has been used, it can decompose and either turn into a brittle substance that can be heard rattling around inside the can, or it turns into a gooey mess. If the gooey mess gets into the contacts, no amount of electrically cleaning the contacts will work. The mechanism and contacts have to be physically cleaned.

Loose rivets.
For a company so prominent in vibrator manufacture, this came as a surprise when I discovered loose contact arm  rivets to be a weak point in design, because it has been observed in quite a number of units. I have yet to see loose rivets in other kinds of vibrator.

The weakness is that the contacts are riveted to the reed and don't appear to be as durable as those used in other mechanisms.. Over time it can be imagined that with all the vibration and contact hammering the rivets might loosen. And indeed they can, as I discovered when studying the reasons for the often poor starting of this type.
When a vibrator has this problem, it might appear to start and buzz, but it will be noted that the frequency is higher than normal, and the buzzing is also quieter. It will also be found that starting voltage is critical. At a certain voltage, which is just below normal operating voltage, the vibrator will actually appear to work normally but as the voltage is raised the frequency suddenly increases. It will also be noted that when this happens, the vibrator is operating in half wave with only the contacts for the driving coil functioning. The amplitude of reed vibration is not enough for the reed to swing over to the other contact.
If this is the situation, the can has to be opened and the rivets tightened. So far, the problem appears to be confined mainly to the later crimped can 150 cycle types.
A catch is for someone unfamiliar with proper vibrator operation is that the radio will still function, so it might be assumed all is OK when it is in fact not. This is one reason that it is important to check the waveform with a CRO.

A perhaps less desirable aspect of the Utah designs seems to be the mechanisms which operate at 150 cycles. This means a very small reed weight is used to get the higher frequency. As such, it has less inertia which results in less pressure is applied to the contacts when the reed swings over to them. The contact condition is therefore more critical, which is also compunded by the shunt drive system..

Conversely, the Oak, and early Utah types operate at 100 cycles. With a larger reed weight and lower frequency, starting is less problematic. In fact, these are usually quite easy to get going, and a simple contact clean after years of disuse is all that is required.

Once it's operating...
A waveform check is required to see that all contacts are working and that the contact condition is good. Electrically cleaning the contacts as decribed previously is only good for burning off oxide or insulating film. It will not repair physical contact damage. So, even though a vibrator starts reliably, it is still necessary to check the contact condition with a CRO. Dirty contacts will overheat because high resistance with current flowing causes heat. One set of contacts not working will damage the other set because the vibrator will be working in half wave, with the result that the buffer capacitance is no longer correct, as well as causing DC to saturate the transformer core.

Reliability.
My experience has been that the Ferrocart vibrators using the Utah mechanism which have been used in car radios have a very poor reliability. It appears that the contacts are simply unable to handle the current. They are certainly the smallest of any vibrator I have seen. Yet, Ferrocart vibrators used in domestic radios do not seem to be anything like as problematic presumably because they handle much lower current.

Statistically, it's unavoidable that shunt drive Ferrocart vibrators are going to need more work than those made by the competitor, MSP/Oak. Nevertheless, the early Ferrocarts with their modified Oak mechanisms are very good once restored. Even though the driving coil has been connected in shunt drive, all the other benefits of the Oak design are present. As has been observed over time, the Oak contacts are very durable and long lasting to the point of not being of concern (provided the operating conditions are correct), so in this situation the shunt drive modification is not really important.

The early 100 cycle Utah mechanism, also seems perfectly reliable once cleaned up. One early synchronous type I cleaned back in 1988 was still working perfectly in 2003 (used in my original Model T Ford radio), and required no further attention. However, this reliability is probably due to this type of vibrator seldom being used in car radios.

The later series of Ferrocart with the crimped can and 150 cycle Utah mechanism seems to be where more work is required. This is the type used in most Astor/Air Chief car radios.

It should be mentioned that MSP did make an Oak vibrator that is a direct plug in replacement for the 6 and 12V non synchronous Ferrocart types PM237 and PM238. These are types V4010 and V4016. As the Oak can is of smaller diameter, the earthing spring contacts often used with the socket will not secure the vibrator, so it needs to be seen that it cannot fall out.  Remember also, the buffer capacitor must be checked for correct value when changing between different types of vibrator. Not only does the frequency affect buffer capacitor value, but so does the duty cycle.

Is Shunt Drive bad?
The answer is definitely no. What has given it a bad reputation in Australia is that there has only been one brand sold here which was badly designed; i.e. Ferrocart. And, this is the one brand that the Australian restorer is going to encounter more than any other.
A box containing 48 vibrators was purchased from the U.S. It included brands such as Mallory, Radiart, GE, Oak, CDE, WIzard, James, Motorola, Delco, etc. All the vibrators were used and had obviously been removed from car radios or other equipment. Except for one vibrator with an open circuit in the driving coil circuit, and another with short circuits between the base pins, all of them could be made to start without opening the can. There were a few series drive types, but most were shunt drive as one would expect. One thing became very clear and that was the Mallory types were the easiest to get started, and in fact some started straight away without any 'prompting'.

The Mallory vibrator design notes claim that shunt drive was best, because not only is it cheaper to make, but they also claim that series driven vibrators have a problem with the drive contact putting an uneven bias on the reed. This may well be true in theory, but the undisputable fact is the Oak series drive mechanism has an unsurpassed reliability record.
Nevertheless, my experience has been that when properly designed, as with Mallory and their rebadges and clones, reliability is not an issue with shunt drive.
A good circuit design and correct operating conditions should not be wearing the contacts, and therefore there should not be any starting problems.

However, there is no doubt that, the series drive design as used by Oak will start even if the power contacts have been completely worn away. While this is a sign of durability, one must query the operating conditions which would cause such wear.

My experience with non Ferrocart shunt drive types has been good. The shunt drive Plessey types which appear to use a Mallory mechanism are very good, and to this date I still have not had any problems with a set of three type 614 that I bought back in 1982. I electrically cleaned the contacts way back then and they just keep going. And of course is the James vibrator discussed here. The plastic of the light it powers has gone yellow from lots of use, but the vibrator is still in excellent condition. A seemingly decrepit (and "got at") Mallory type 46 that I pulled out of my collection started straight away, and with a pretty good waveform. Closely looking at its construction, it does seem like a nicely made unit.

In conclusion, regarding Ferrocart vibrators, it can be said that the Oak series drive mechanism is the best of them all, with reliability being virtually a complete non issue. Where a shunt drive type is used, the modified Oak is the best, followed by the 100 cycle Utah type. The 150 cycle Utah needs the most work.

Not the vibrator's fault!
It must be pointed out that vibrators, aside from certain types of Ferrocart,  are inherently reliable components and should not fail any more than say a capacitor or valve. An advantage of course is that unless there has been physical damage to something such as by extreme overheating, they can be repaired. Where a problem may occur is with circuit design and operating conditions. Things to consider are:

Normally, if the vibrator is presented with a load that appears completely resistive, and within ratings, it should have complete reliability. The vibrator manufacturer has unfortunately no control over the design of the equipment it is used in.
It is also important to remember that vibrators in the present day are at least 45 years old, and in that time film has built up on the contacts, as well as the possibility of the sponge rubber insulation decomposing. So, it is no wonder that even a N.O.S. vibrator might require restoration. Vibrators may have been used in equipment that was poorly designed or faulty. One can hardly blame the vibrator if it has been damaged by such operation.

In Australia, car radios were a rarity, and were not standard in cars until the late 1970's. It took the popularity of Japanese cars, which were fitted with radios, to force Australian car manufacturers to change their conservative attitudes. However, 6V battery operated vibrator radios were quite common in rural areas.  Australian service literature does not seem to imply vibrators are any more unreliable than other components.
 

Homepage