MSP/Oak Vibrator Information

The most popular vibrator in Australia was the MSP/Oak/AWA unit. It was exclusively used by numerous manufacturers including Ferris, Philips, Ferguson, Aegis, and of course, AWA. In fact, it was used by most manufacturers except for Astor, Air Chief, and other Electronics Industries brands, who used their Ferrocart branded vibrator. The other common Australian vibrator type was Van Ruyten, but these are high power 50 c/s units used for 240V AC inverters, and appear to have been made by Electronic Industries. Aside from these, foreign brands are sometimes seen, particularly in imported items.


Later style label and box design.

Oak Vibrators in Australia.
AWA was manufacturing vibrators in Australia, under licence to Oak in the U.S. Initially, the vibrators were branded AWA, but later branded MSP. "Manufacturers Special Products" was the component manufacturing division of AWA which came into being after WW2. It was a way to sell AWA components to competing manufacturers, as MSP was not a brand name associated with consumer products. It can be imagined, for example, that an AWA vibrator inside a Philips radio would look awkward from a commercial point of view.  The AWA manufactured version was based on the 1934 design, and unlike in the U.S., did not change until the cessation of vibrator manufacturing around 1973. Oak types in the U.S. had several can styles, many of which were crimped closed. Fortunately, AWA kept the can design to the original type, with the mechanism secured inside by a spring clip around the base. This makes subsequent servicing, if ever required, very simple.

AWA's first car radio (model 74) in 1934 used an RCA vibrator, which was the same type used in the RCA M34 car radio. The next AWA car radio model, the 153 from 1935, used a motor-generator. A return to vibrator power supplies was coincident with the 401 in 1937.
However, it is not until the model 404 of 1938 that the service manual acknowledges that the Oak vibrator was available as a replacement. The original vibrator for the 404 was an Electronic type (Electronic Laboratories). This is a shunt drive type and the manual describes rewiring the vibrator socket to take the Oak replacement. From this, it seems that the Oak vibrator was standard in Australia by 1939.


Prior to MSP coming into existance, original production was labelled AWA.

Apart from types V5105 and V5124, most Australian type numbers are not used by Oak in the U.S. and vice versa. AWA also did local development of the design. At the time, there was little demand for 12V vibrators in the U.S., and so it fell on AWA to provide 12V types for the multitude of English cars in Australia. Likewise, the 4 volt types for domestic sets were not common in the U.S.  A patent taken out by AWA relates to the development of the split-reed type. A method of driving 6 volt dual-interrupter vibrators off either 6 or 12V without changing the vibrator or transformer is another AWA patent.

24 and 32V types were not manufactured until after WW2. Earlier 32V radios are shown with a 12V V5258 synchronous vibrator operated with a 100 ohm resistor in series with the drive coil. By 1950, a 24V type was available, and was being used in at least one Radio & Hobbies project on 32V, again with a drive coil dropping resistor. It appears that 32V types became available shortly after.


At left is the first generation of MSP labelling, and at right is the early style box.



Patented in 1934 by Oak.
The Oak vibrator was first patented in the U.S. in 1934. Its main improvement over other vibrators was the series driven coil, having its own contact independent of the transformer switching contacts.

The design was patented by Oak in 1934. Adjacent photo shows internal parts of a non synchronous vibrator. This particular type is a V4012.

Good Reliability
The "series" or "separate" drive design means the vibrator always starts, even if the main switching contacts are worn or the supply voltage is a bit low. This is because the driver contact is only switching the small coil current, and the contact itself is a non tarnishing precious metal. One test showed no visible wear of this contact after 1500 hours of operation.
With the more common shunt drive type, the vibrator will not start if the power contacts have worn or are out of adjustment.


From notes in Radio Engineering, April 1935. Their prophecy of longevity turned out to be correct.

It is because of the prevalence of the Oak design in Australia, that reliability problems which seem so common in the U.S. where the shunt drive type is standard, did not exist here. To help explain this, it must also be pointed out that in the era of vibrator powered equipment, that buffer capacitors were usually paper types with relatively poor reliability. This has more than likely contributed to the common  "vibrators are unreliable" myth that exists. With a poorly performing buffer circuit, the power contacts will wear quickly, and then the vibrator will not start. In the same circumstances, the power contacts of a series drive vibrator will also wear, but the reed will still vibrate as normal. The effect is that the output voltage of the power supply is still present, but gradually drops due to the descreasing duty cycle as the contact gap increases. So, under less than ideal conditions the vibrator is still "working". Of course, with the correct loading and operating conditions which avoid contact wear, the shunt drive type will also have good reliability, provided the vibrator is of reputable design and manufacture. Obviously a non-starting vibrator is more of an inconvenience than one which simply has reduced output, and in this regard the series drive is ultimately more reliable.
Out of my hundred odd items of vibrator power equipment it has never been necessary to actually replace a vibrator (shunt or series drive); and this is over more than 35 years.
So, fear not fellow Aussies; do not take what you read in overseas magazines and vintage radio forums as gospel, regarding vibrator reliability. Our series drive MSP/Oak units do not have those issues, instead lasting as long as any other electronic component. And, if you're into restoring Astor/Air Chief car radios, it's worth looking out for the MSP replacements for the original Ferrocart types. These are discussed further on.
Of course, it's important to note that unsuitable circuit design, loading, or defective buffer capacitors will ruin even the best vibrator. With the correct component selection and attention to power rating, an Oak vibrator is completely reliable with an extremely long life - so long that it's really a non issue.
Unfortunately, vibrators often fall victim to the incessant 'tweaker' who is drawn to its moving parts like moths to a light. Oblivious to the precision construction and careful setting up in the factory , they twist the contacts this way and that, attempting to repair a defective unit. The teeth marks of long nose pliers can often be seen...
In actual fact, randomly "adjusting" a vibrator is one of the worst things one can do. One needs an oscilloscope to do it properly, for the timing and duty cycle cannot be accurately set otherwise. Unfortunately, some enthusiasts cannot get past the "it's just a buzzer" mentality.

Run into a resistive load, the vibrator primary contacts are adjusted thus. A tweaker who jumps in with a pair of pliers and no instruments cannot hope to duplicate this. Add secondary contacts and the mess can be imagined! Yet, they're the first to complain about the reliability of vibrators...

Operation at Higher Voltage.
Another advantage of the separate connection to the driving reed is that it is possible to operate the vibrator off a higher voltage simply by connecting a resistor in series with the driving coil. For example, a V5105 which has a 6V coil can be run off 12V with a 22 to 27R 5W resistor in series. One model of Philips car radio, NT364V,  had this facility built in, so that the radio could be used in a 6 or 12V car simply by changing links under a panel. The power supply shown below operates a 12V vibrator from a 32V supply using a 100R 3W resistor in series with the drive coil.


32V vibrator power supply for an AWA domestic radio. When this was designed, 12V was the highest voltage available for the vibrator. Note R1 to drop the coil voltage. Once 32V vibrators became available, R1 was deleted. Incidentally, this power supply was issued with a modification to move the buffer capacitance to the primary to eliminate contact arcing.

The Bland shaver inverter is another example; in this case a 6V vibrator is used on 12V.

Construction of the Oak vibrator.
The design of the driving coil is unique to Oak and was one of their patents. It will be noted on most schematic diagrams that the Oak vibrator has a secondary winding on the driving coil which is short circuited. This is wound with chromoxide resistance wire and considerably lowers the Q of the driving coil.  The windings are bifilar wound. The purpose of this is to eliminate sparking at the driver contact. It does this because the lowered Q slows down the rate at which the magnetic flux collapses when the contact opens. The usual inductive kick back is prevented. Other manufacturers of series drive vibrators use more conventional spark suppression circuits with a resistor, a capacitor, or a combination of both.
The reed is made of Swedish spring steel. The power contacts are made of Oakalite tungsten, and the drive contact is palladium silver alloy.

Servicing.
As with most vibrators taken out of storage after many years, the insulating film that builds up on the contacts needs to be cleaned off. This is usually the cause of a vibrator not providing output, whether it starts or not. The driving coil contact seldom is affected this way, because it is made of a precious metal and does not tarnish. It is also normally closed, and thus nothing gets in between the contacts when not in use. Mallory mentions that the film is tungsten oxide which forms on power contacts. However, it seems to be more prevalent in vibrators where sponge rubber is used inside the can. In this case it appears to be a decomposition by-product (sulphur is one possibility). The film can be physically cleaned off the contacts by means of 600 grade sandpaper or a fine blade such as an X-acto knife. However, care must be taken not to upset the contact spacing. For this reason, I prefer to electrically clean the contacts where the film build up is fairly minor. Current limited high voltage will break through the insulation and burn it off.

Another sensible aspect of design is that service access for the Australian Oak vibrator is easy. Unsolder the earth tag and remove the spring clip. The mechanism can then be slid out. In some instances, it will be found the base of the can has expanded in diameter, with the phenolic base a looser fit than normal. This is a result of the constant pressure from the spring clip against the soft zinc can. The zinc creeps over time under pressure.
Usually, it is not necessary to open the vibrator unless it has been damaged, or after a long period the contacts in a synchronous or dual interrupter type need to be retimed.
If the driving coil does not function it's either: 1) the wire broken off just near the coil, 2) the adjustment screw for the contact is a bit out, 3) dry joint in pin or at bottom of vibrator frame. In a couple of rare instances broken or cracked reeds have been found, and in one, the foam lining of the can had been replaced with a type which decomposed rapidly, leaving a sticky substance between the contacts.

To electrically clean the contacts, the following test jig can be assembled:

A battery of appropriate voltage is used to get the reed vibrating. While a bench power supply can be used, there is a risk of inadvertent damage if it should somehow come into contact with the high voltage. High voltage is obtained from a 240V isolating transformer and current limited to 400mA by a 100W incandescent light bulb. This is applied to each contact in turn until the contaminants burn off. Obviously, once this happens the bulb lights up. Because of the vibrator frequency not being related to the mains supply, the bulb will flicker once the contacts make. This is quite normal.
It goes without saying that no part of the circuit, or the vibrator should be touched while the mains is connected. The can of the Australian made Oak vibrator is usually connected to the reed.  Although the transformer provides isolation from the mains, there is still 240V between its secondary terminals.
Disclaimer: There are restorers who don't appreciate the dangers of a direct mains connection and will run this circuit without a transformer. No reponsibility is taken for undesirable results that occur.

Check all contacts
It's important to see that all contacts are functioning, because if the vibrator works in half wave, those contacts which are working will be damaged. Simply observing the transformer waveform will confirm this, but the output voltage is also revealing. A potential problem is someone who gets a vintage radio, turns it on and find that it appears to work. If the vibrator is working in half wave, the radio will indeed be working, deluding the owner into thinking all is well. However, damage will soon set in, and if the contacts are badly burned and the spring temper damaged, the long term reliability could be poor despite efforts to repair the vibrator.

Adjustment of the Oak vibrator.
Discussion with a retired AWA employee who worked in the vibrator division of MSP brought forth some useful information. The vibrators were adjusted after assembly using a test panel fitted with dwell meters and also a CRT. The vibrators were tested with a resistive load. Apparently, this test panel was "ancient" and pre-war. The CRT was assumed to be a 5BP1. The adjustment was done only via the side (fixed) contacts. As it happened, the precision of manufacture was that not much adjustment was needed anyway. A standard type of relay adjusting tool was used; i.e. a rod with a slot cut into the end. Each primary contact was set up for 40% duty cycle with a 5% tolerance. When operating in a full wave circuit, the duty cycle is therefore 80%.
Because the Oak vibrator operates at 100c/s, this means each contact is set for 4ms closing time. With a calibrated oscilloscope, the adjustment is therefore quite easy. While I could have designed my vibrator test panel with dwell meters, it's more accurate to use a CRO, which these days are inexpensive and commonplace anyway. Further detail on adjustment is given here.



MSP / Oak Vibrator Data


 

Non Synchronous types.
The basic vibrator, most commonly used in car radios and small inverters. Where DC output is required, a valve or other rectifier is required. Most commonly found are the V5105 and V5123 types. These were used in every AWA  car radio, of the valve rectifier type, up until the last model in 1965. Note that with all the vibrators types listed (not just the non - synchronous types), the contact current rating goes down as input voltage increases. This is because arcing is more prone to occur with an increase in voltage.

Synchronous types.
These are fitted with secondary contacts for self rectification. Commonly used where it is undesirable to use a valve rectifier, such as where dictated by space or power consumption constraints. Some car radios did use synchronous types, but mostly they were used for accumulator powered domestic radios. Note that the timing of the primary and secondary contacts is not identical. The secondary contacts open and close slightly later than those of the primary. The 4 volt type might cause some curiosity. It had two popular applications. One was for portable camera flash units where a small two cell lead acid battery was used, but more commonly they were used for 6 volt accumulator powered radios. Here, the vibrator is connected across the 4 volt section, and the valve heaters across the remaining cell (2V). The idea was that by supplying the directly heated filaments from a separate cell, it was easier to keep vibrator interference out of the system.

 

Synchronous Split Reed types.
These are mostly used in domestic accumulator powered radios, because with directly heated battery valves an external bias supply is often required. As the valves filaments are also the cathode, it is not possible to use cathode bias in the same way as indirectly heated types.
It is also possible to configure a split reed vibrator so that it works as a full wave switch into a non centre tapped transformer primary. The Bland shaver inverter takes this approach. It is worth keeping this in mind where it is necessary to convert a 6V car radio to run on 12V. By changing the vibrator to a split reed type, it is possible to keep the original 6-0-6V transformer on 12V by not using the centre tap.

6-12V conversion with a Split Reed Vibrator.
Vintage cars having their 6V electrical system converted to 12V are a fact of life, as much as the practice is undesirable. When it comes to the radio, several options are available. A 6 volt radio can be simply run from the new 12V supply by means of a resistor. The resistor is selected so that with the battery 12V (i.e. not charging), the radio receives 6V as measured on an analog meter. Why analog? Because the current draw is not steady DC and the peculiar waveform is more likely to give a less accurate result with a digital meter. In practice the resistor value is somewhere around 2.5 ohms and dissipates a fair bit of heat; 30W is not uncommon. The exact resistor of course depends on the particular radio and has to be selected on test.
The disadvantage of this scheme is the wasted power and that the nominal 6V changes as the radio warms up. This may or may not be a problem in the particular situation.
The more efficient method is to convert the radio to 12V. For the valve heaters and dial lamp, these can be replaced with their 12V equivalent; e.g. 12X4, 12AQ5, 12BA6, etc. If this is not convenient, the existing valves can be rewired in a series parallel circuit. See here for further details.
The next thing is to deal with the vibrator and transformer. If possible, the transformer primary could be rewound with twice as many turns using thinner wire. Under the assumption that the vibrator is an MSP type, a 27R 5W resistor can be connected in series with the drive coil. Simply installing a series resistor to power either the transformer alone, or transformer and vibrator is not recommended. The problem is that until the valves warm up, there will be no load on the inverter, causing excessive voltage in the transformer secondary winding which can cause insulation failure, damage to the buffer capacitor and arcing at the rectifier socket. Where the whole radio is fed from a resistor this is less problematic as the valve heaters provide a constant portion of the load.
However, there is a much more ingenious way around the problem. It allows the original transformer to be used without modification.


How to use a split reed vibrator to enable a 6V vibrator transformer to operate on 12V. Note that the driving coil is not shown for simplicity.

Here, a split reed vibrator is connected so it works as a DPDT reversing switch, applying +12V then -12V to the full primary. In effect, the primary receives the same peak to peak voltage that it did on 6V.
The centre tap is no longer used. The advantage is of course the regulation and efficiency is as per original, and the transformer needs no modification. The vibrator and its socket do have to be replaced however, but as these items have the same appearance as the originals, the radio will still look original. The Bland shaver inverter uses this technique.
Of course, if the radio used a synchronous vibrator for rectification in its original 6V form, then either a solid state or valve rectifier will have to be fitted, but this is likely to be preferable to replacing the transformer.
 

Dual Interrupter types.
These are commonly used in 240VAC inverters or high power amplifiers, radio transceivers, etc. They are constructed the same as synchronous types but with the contact timing adjusted so both sets open and close at the same time. The contacts can be connected in connected in parallel to increase current rating, or more effectively, be used to switch two separate transformer primaries, or two separate transformers with the secondaries in parallel.  These latter two configurations provide better current sharing between the contacts. In reality, it is impossible to ensure both pairs of contacts open and close at exactly the same time over the life of the vibrator, so the current rating is not actually doubled as might be imagined. In the U.S., the term "Duplex" has been given to this type of vibrator.

There are also types V7706 and V7712, but data is not shown. As far has been ascertained without actually opening one, these are the same as the V66xx types as far as connections and the driver coils are concerned.
The circuit below, from Radio & Hobbies, April 1948, illustrates the use of a dual-interrupter vibrator. It is of a 240V AC 40W inverter, powered from either 6 or 12V. The transformer is a Ferguson VT146 which is designed for 6 or 12V operation by means of a tapped primary.
For 6V, the two primary windings are fed independently from each set of vibrator contacts. This provides more even current sharing. The 12V configuration simply has the two sets of contacts in parallel, but as the current is halved at this voltage, the current sharing is less critical.


This circuit from Radio & Hobbies, April 1948, illustrates how a dual-interrupter vibrator is used. Contacts may be paralleled, or drive separate primary windings.

Ferris car radios and inverters.
A type that seems to be unique to Ferris is the six pin based V4012 and V4006. These are simply the same as their 4 pin counterparts except for the base. Internally, the V4012 is the same as the V5123 and the V4006 is the same as the V5105. Why Ferris chose to use a 6 pin base is a mystery. Perhaps it's because the extra pins exert a firmer grip on the vibrator in the socket.

With the standardisation of pin connections, these types actually have the same pin connections as the synchronous and dual interrupter types, except there's no extra set of contacts. This means that one often finds dual interrupter or synchronous vibrators plugged in instead, even though the extra contacts are not being used. It's a worthwhile modification to simply link the other socket pins to bring the extra contacts into being in parallel. While this has limited merit with synchronous types due to the later closure of the secondary contacts, it means that although current rating won't be greatly increased, the secondary contact will take over if something goes wrong with the primary contact. No doubt radio servicemen found it convenient that any MSP six pin vibrator could be plugged into these sets, without having to do any socket rewiring.


The reliability of Ferris car radios was partly due to their use of the MSP Oak vibrator.

Astor car radios.
Type V4010 and V4016 are  interesting types. They also have a 6 pin base and are non synchronous. They are designed as a replacement for the shunt driven Ferrocarts used in Astor car radios. Four pin Ferrocart vibrators actually use a four pin Delco base. This is effectively a UX-6 base, with pins 2 and 5 physically omitted.
V4010 is 6 volt and replaces the Ferrocart PM237, and V4016 replaces the 12 volt PM238. These vibrators are immediately obvious because adjacent pairs of pins are strapped together on the underside. The straps are so connected that the three active pins of the V4010/V4016 match up with those of the Ferrocart type. However, some versions have the strapping internally.
Otherwise, they use the same components as the normal series drive non-synchronous types, except that the drive coil takes its feed from one of the power contacts. Also, the drive coil is of higher resistance, since the applied voltage is twice the supply voltage, due to the transformer action.
Unlike the shunt drive Ferrocarts which the V4010 and V4016 replace, the Oak design retains their separate drive contact, despite the use of shunt drive circuit. This means a considerably greater life expectancy can be obtained before any adjustments are required.

Separate drive contacts used with the shunt drive circuit.


Depending on the source of data, the internal schematic shows the V4010/V4016 as either a true shunt drive type, with no separate drive contact, or as has been confirmed with all the examples in my collection, a separate drive contact is used.

Types V4011 and V4017 are electrically the same, but without the strapped pairs of pins. It is not clear why the two variations exist, since they would appear to be completely equivalent, when plugged into a set designed for four pin Ferrocart vibrators. Indeed, I have never seen the V4011 or V4017. The likely answer appears that some of the "four pin" Ferrocart sockets were just that, and did not have holes for pins 2 and 5. Therefore a UX-6 base vibrator cannot be plugged in, short of removing pins 2 and 5. It appears V4011/V4017 are made without pins 2 and 5.


The three strapped pairs of pins immediately identifies the vibrator as a V4010 or V4016. Note that some examples have the pins strapped internally.

Although I have not done tests, it would appear necessary to check the buffer capacitor value when using an MSP to replace a Ferrocart in view of the PM237 and PM238 operating at 150c/s. On the lower frequency of 100c/s produced by the MSP, the buffer capacitor may need to be increased.

Aged types.
Prefixed by "A", these are painted black to enhance cool running. A typical type is AV5948. They are run for an ageing period prior to final adjustment. Mostly, they appear in mobile two way radio applications, such as the AWA Carphone.

Reverse polarity types.
Suffixed by "R"; for example AV5948R. These are synchronous types that have the secondary contact connections swapped over, so that the output polarity is reversed from normal. An application for this is where a radio might be transferred to a positive earth car from one which was negative earth, or vice versa. Instead of swapping over the transformer connections, a quicker alternative is to simply change the vibrator to the other type. They are also used where split reed synchronous power supplies are used in series. This is because the secondary centre tap of one transformer has to be positive (the normal situation), while that of the other has to be negative in order for the voltages to add.
On this note, the AV5948 is used with the AV5948R in one of the AWA Carphone power supplies to provide 150V on receive and 300V on transmit.


An "Aged" reverse polarity type, AV5948R. This is a 12V split-reed synchronous type.

AC types.
The final, rather obscure type, is one in which AC is applied to the driving coil. The vibrator functions as a switch, synchronised to the mains. So far, the only application I have seen such a type in is a mains operated ignition coil tester.


Type VAC5124 has a 6V AC driving coil.



Interesting MSP Data.
In the January 2002 issue of the HRSA Radio Waves magazine are tables of data for Ferrocart and Oak vibrators.
It is a curious thing as to where this data was obtained because there are differences to that shown above. For instance, the 12V vibrators are rated at 2.5A input current, instead of 4A. The frequency for all types is shown as 115 c/s instead of 100 c/s (as per the above data, and what I have measured).
The assumption is that this data may be early, and it was found that the 12v current ratings could be increased to 4A. There are recollections about revised ratings being issued. This doesn't answer the question about the difference in frequency which could actually be problematic because of buffer condenser values. The source of data is not known, but it is noted that the Radiotron Designer's Handbook mentions 115c/s being the standard frequency. However, the information here was largely sourced from Mallory, who did standardise on 115 c/s.
It is possible that whoever compiled this data assumed that because the RDH was an AWA publication, that AWA vibrators would therefore be 115c/s.
The data I have provided above is from genuine AWA literature that I physically own, and tallies up with my examination of the vibrators. Nevertheless, this other (older?) data is interesting.

Radio Parts Catalog 1968-1969.


Oak Vibrators in the USA and UK.

The English company, Wright & Weaire was another establishment outside the U.S. making Oak vibrators. However, unlike the Australian MSP version, the Wright & Weaire version does not externally look like the original Oak, and the numbering system is totally different. Instead of V5123, for example, W&W will use NS/12.  This is a more meaningful numbering system; "NS" being "Non Synchronous" and "12" being the operating voltage. The can is of a wider diameter, and the base is crimped in place. The can is aluminium instead of zinc. Looking at the cut away diagram above, we can see that the can contains much more rubber for acoustic insulation. The entire can is lined, whereas the original design is not insulated against noise except for at the top and bottom. Presumably the thicker zinc stops most of the noise transmission by itself, whereas thin aluminium is not so effective. It is unfortunate that the crimped can construction has been used, but internally these vibrators should be as good as any Oak design.


Oak vibrator of U.S. manufacture. At right is a Wright & Weaire vibrator type NS/12.


Side view of the vibrators. The W&W shows the Services number.

It was the opinion of some manufacturers to seal the base because they thought it would prevent the entry of air which supposedly caused contact oxidation when not in use. This was a fallacy, because in my large collection of open based and sealed vibrators there is absolutely no proof of this. After 40+ years of disuse, the contacts in the sealed units in most instances still need cleaning.
For most manufacturers, it was to discourage tampering with the adjustments - which few servicemen are skilled enough to do, and to make it obvious that a vibrator had been worked on, in case of returning one under guarantee.


Non-Australian Oak vibrators.  The first four on the left are U.S. made. The left most vibrator is the  type on which the Australian AWA/MSP version is based. The finned can type (4th from left) was also rebadged by Terado, and it seems, Raytheon and GE. The right-most vibrator is an English Wearite (Wright & Weaire) which uses the Oak mechanism.



Things that have been found...

Alarm bells rang when I saw this written on the side of a V6606 dual-interrupter vibrator:

It says " OK 12 Volt". Some clown had decided to use a 6V vibrator on 12V because it appeared to work. I dreaded what I'd find inside, and indeed the mechanism and sponge rubber was coated in a black substance.


Don't run a vibrator with incorrect coil voltage, although it might "seem OK".

Prior to taking the photo I had cleaned it up, but the black deposit is still evident around the sponge rubber.
This kind of component abuse is always appalling, in this case almost ruining a perfectly good vibrator. Thankfully, the driving coil appears to have survived, although something seems to have exuded from it particularly at the top of the bobbin.  It would appear that someone needed a 12V type but didn't have one, and discovered that a 6V type would work instead. Yes it will, but with excess driving coil current and the resultant overheating, the arcing at the driving coil contact, and of course the excess reed swing giving an extra hammering to the contacts. Who knows how long it would have taken to burn out the driving coil. If only a 27R resistor had been connected in series with it. It's rather reminiscent of people with vintage cars that feed 12V into 6V starter motors.
And yet, you can be sure whoever did this would be the first to complain about vibrators being unreliable or some such nonsense.

This V4016 shown below, refused to start. Upon opening it, the reed could be seen to be catching on the solenoid pole piece. The reed also was not centered at rest. In an attempt to correct this, it was found the reed was broken.


Reed catching on pole piece. It was found to be cracked.

This broken reed came from a fellow HRSA member. It was from a V4012. It should be pointed out these are the only two broken reeds I have seen in Oak vibrators.


Reed breakage is one fault that cannot be repaired, but fortunately is a rare occurrence.

Another vibrator that wouldn't start - a V5105. Someone in the past had replaced the foam rubber. Unfortunately, they had used a type of foam that disintegrates into a sticky corrosive mess. After a good clean up, this vibrator worked perfectly. The connecting wires to the base had also been replaced with ordinary rubber insulated type. This had gone brittle. The best kind of wire for this is ultra flexible instrument lead wire.


Don't use this kind of foam to replace the acoustic insulation!



Identifying Oak vibrators by coil characteristics.
It sometimes happens that a vibrator is obtained with a missing label. The first thing is to identify the voltage of the drive coil. Here follows the coil characteristics:
 
Volts Coil Resistance Inductance Q Examples of types
4 8.5R 640uH 0.162 V6804
6 12R 1.1mH 0.17 V5105, V5124, V6606, V7706, V5211, V4006
12 35R 1.1mH 0.09 V5123, V5258, V6612, V7712, V5958, V4012
6 (shunt circuit) 24R 1.2mH 0.11 V4010
12 (shunt circuit) 73R 1.5mH 0.06 V4016
24 131R 4.06mH .08 V6724, V6824, V6624, V6524
32 230R 3.64mH .054 V6632, V6532, V6732, V6832

The number of pins further narrows down the types. For example, 4 pins means a non-synchronous type such as V5105 or V5123. Six pins means synchronous, dual interruptor, or Ferris type non-synchronous. The Ferris types V4006 and V4012 can be identified in that pins 2 and 4 show no connection whether the vibrator is operating or not. Six pins could also mean it's a Ferrocart  replacement V4010 or V4016, and checking for continuity between adjacent pairs of pins will confirm this.

To differentiate between synchronous and dual interruptor types, the timing of each set of contacts has be examined on a CRO by using a set up like this.
Seven pin types are split reed synchronous, and once the voltage rating has been determined, the type can be further narrowed down. However, this does not determine if it is an "R" (reversed) suffix device. To do this requires the primary and secondary contact timing to be examined on a CRO, or simply by using the vibrator in a known device and seeing if the output voltage is reversed. Note that reversed types are in the minority and not generally encountered with domestic radio equipment.

Ferrocart replacement types V4010 and V4016, are immediately identifiable by their three pairs of strapped pins.

Of course, if one wishes to open the vibrator for examination, it can be narrowed down to a few types. However, it is a shame to break the stamped soldered seal on a prisitine N.O.S. type just to do this.

Old Type Numbers.
Initially, Oak vibrators were supplied using a different numbering system with two digits and two letters. This was later dropped, and the Vxxxx system was used exclusively.
 
Old Type Later Type
17XH V5105
18XH V5123
185XH V6524
33ZH V5278
35ZH V5124
36ZH V5258
365ZH V6724
63UH V6804
35UH V5211
66UH V5948


Homepage