Other Meter Manufacturers

These meters were made by companies that were unable to compete with GE, Westinghouse, Sangamo, and Duncan.

Early Models

Samuel Gardiner's Lamp-Hour Meter

(1872)
This was one of the earliest meters and the first known model to be patented. It was used with the earliest DC arc-lamp systems and only measured the time energy was supplied to the load as all the lamps connected to this meter were controlled by one switch. The internal workings were nothing more than a clock that was started and stopped by a electromagnet that was connected in series with the load. With the introduction of Edison's light bulb, subdividing lighting circuits became practical and this meter became obsolete, replaced by meters such as Edison's chemical meter and the Thomson Recording Wattmeter. 
Photo: Westinghouse Handbook of Watthour Meters

J.B. Fuller's Lamp-Hour Meter

(1878)
This meter was the first known AC meter to be patented, and like the Gardiner DC lamp-hour meter that preceded it, it only measured the time energy was supplied to the load. The meter was simply a clock operated by the alternating fields of the two coils. Lamp-hour meters were soon abandoned as customers started adding other electrical appliances and the need arose for meters that meters that measured energy, like Shallenberger's ampere-hour meter. One of Jim B. Fuller's assistants was James J. Wood, who would soon become one of the key people at the Ft. Wayne Electric Co.
Photo: Westinghouse Handbook on Watthour Meters

Edison Chemical Meter

(1878 to late 1880s)
Edison developed the first meter that measured the amount of electricity instead of how long the circuit was energized. This meter was connected across a shunt in the load and consisted of several jars with zinc plates and a chemical solution. One set of jars was intended for the main reading while the second set was operated off a smaller shunt and was intended for comparison purposes (a primitive check on the meter's accuracy). The monthly reading was made by removing the plates from the jars and weighing them with a laboratory balance. The change in the plates' weight between readings was a measure of electricity consumption. This meter was very inconvenient to use, and in a couple cases mishandling of the plates resulted in large billing errors. Also, as there was no ready way to indicate the usage to the customer, this also made it hard for them to trust its accuracy. These disadvantages and the fact that this meter itself had high internal losses made it unpopular enough that these meters were rapidly replaced by more reliable meters in the late 1880's, including the Thomson Recording Wattmeter.
Photo: Handbook for Electrical Metermen

Thomson Walking Beam Meter

(1888 to 1889)
This meter was developed by Elihu Thomson and was so complex and delicate it is almost inconceivable it was developed by the same person who developed the robust and very successful Thomson Recording Wattmeter! A heating element was connected into the circuit and warmed the lower bottle closest to it. As the alcohol warmed, it evaporated and flowed into the opposite upper bottle. This would upset the balance, shifting the other lower bottle closer to the heater and emptying the upper bottle that was just filled. As the meter rocked, an escapement drove the register which was marked off in lamp-hours. 
Photo:
Schenectady Museum Archives

Siemens & Halske Duncan Integrating Wattmeter

(1899-1900)
After R.T. McDonald died in 1899, Duncan left Fort Wayne for a brief stint at Westinghouse before coming back to Ft. Wayne. Shortly after his return, he and several other employees were lured away by C.S. Knight (who was a manager at Ft. Wayne himself before being hired by S & H). Once at S & H, Duncan refined his watthour meter into a slimmer, top-connected version. The light load adjustment was changed to use a copper ring with an iron core. In early 1900, shortly after Duncan's arrival, there were widespread strikes in the Chicago area, and Siemens & Halske's plant was one of the affected companies. This put the company in such a position they had to sell out to GE, but the strikes were soon settled and existing orders were filled. Late that year, GE closed the S & H plant and moved the tooling and many employees to Fort Wayne. However, Duncan felt the time was right to leave for Lafayette, IN to establish the Duncan Electric Mfg. Co. 
Photo by Mark Hertzler

Diamond Meter Co.

Diamond/Scheeffer Type C

(ca. 1895)
Very little data exists on this model other than these meters were similar to GE's Form C Thomson Induction Wattmeter or Fort Wayne's Duncan watthour meter. 
Photo: 1897 Electric Appliance Co. catalog

Diamond/Scheeffer Type D

This model was an induction type like the previous Type C. The motor was of a radically different design. The stator was formed out of a pair of yokes - the lower one carrying the windings and the upper one serving as a return path for both sets of coils. The voltage coils were wound on the limbs of the lower yoke and the current coil was positioned above and between the voltage coil sections. The light-load adjusters were positioned at both ends of the disk in the gap between the two yokes. The full-load adjustment was two iron pieces on the end of the brake magnet. As with the early GE Thomson Recording Wattmeters, the register was non-direct reading, a multiplier being marked on the register to arrive at the correct reading (one register ratio was used for all except the smallest capacities). The current terminals are at the top and the voltage coil tap is at the bottom right corner of the base.
Photo: Unknown collection

Diamond/Scheeffer Type E (Early)

This model was similar to the Type D except the position of the motor and terminal arrangement are turned upside down. Also, a lag coil was added and was no different than lag coils used on other meters. The two light-load adjusters on the Type D were replaced with a single strip that could be moved sideways. The terminals were also redesigned so they could be more easily accessed without having to take off the cover. Starting with this model, the registers included ratio gearing so that the meter could be read directly without using a multiplier.
Photo: Phil Shelley

Diamond/Scheeffer Type E (Second Version)

This second version of the Type E seems to just be a simplification of the meter's design and to eliminate the 'shelf' in the cover. The light-load adjustment vane was eliminated and the lower yoke of the stator was modified so that it could be tilted to either side to act as the light-load adjustment. The thin sheet brass cover previously used was replaced with one made of cast iron(!).
Photo: Scan by Bill Ostrander

Diamond/Scheeffer Type F (DC)

Unlike the previous models, the first version of the Type F was made mainly for use on DC circuits (but could be used on some AC circuits). This was a typical commutator-type meter but with the internal components arranged so that they could fit the same round enclosure used with the previous Type D and E induction-type meters. The light-load adjusting coil was of the type where it could be slid in or out of the current coil, as was done on the Thomson Recording Wattmeter.
Photo: Scan by Bill Ostrander

Diamond/Scheeffer Type F (DC/AC)

Like the earlier version of this model, the later version of the Type F was a commutator-type meter. However, the motor was radically different. The current coil was split into 4 sections, all connected in series. These were arranged so that there was a horizontal gap between the upper and lower pairs of the current coil sections. The armature was made up of three flat coils positioned on a three-pronged disk of mica and which rotated through the gap in the current coil sections. No picture of this meter yet.

Diamond/Scheeffer Type H

This model was the last one produced by this company and apparently the only one that G.A. Scheeffer did not design. This was a commutator-type meter that could be used on either AC or DC circuits. Unlike many other commutator-type meters, the resistance was wound in a disk shape and was accessible from the back of the meter. Also, this model was available with either the regular pointer-type register or with a small cyclometer register.
Photo: Unknown collection (found on Ebay) 

Columbia Meter Co.

Columbia Type D Meter (DC)

(1910s)
The Type D meter has several radical differences from other DC meters. First, the armature has only 3 coils instead of the 8 normally used. Secondly, the current (and light-load compensation) coils are mounted at a 45-degree angle, and the light-load adjustment is a tapped resistor to vary the voltage fed to the armature. 
Photo: Electrical Meterman's Handbook

Columbia Type C1 (Single Phase)

(1910s)
This model had a number of unique construction features, possibly as a way to avoid patents owned by the 'big 4' companies. The stator on this meter was arranged with 2 potential coils in parallel and set near opposite ends of the disk and the current coils set in corresponding positions underneath the disk. Instead of the usual closed winding on the voltage coil(s), the lag load adjustment involved varying the effective length of the current coil with a slider on a U-shaped extension of one current coil. The light load adjustment was an adjustable iron punching on top of each voltage coil (to distort the shape of the field through the coil). Replacing the jewel bearing on this meter was as simple as sliding a retaining spring out of the way. 
Photo: Electrical Meterman's Handbook

Columbia Type C3 (Polyphase)

(1910s)
This was apparently a polyphase model, but we have ZERO information on it beyond that it was built in the more common voltage and current ratings being produced by the other manufacturers.

Columbia Type SA Meter (DC Switchboard)

(1910s)
This meter was made for use with shunts on high-capacity DC circuits but had an unique internal design. The rotor had 3 disks, the lowest serving as the regular brake disk which operated between 4 astatically arranged magnets. The upper two, surrounded the main armature coils which had a thin piece of steel going through each coil. The strip of steel extended from the underside of the top disk, down through an armature coil, and out along the top of the lower disk. This served to conduct the flux between the 4 series coils surrounding the upper disks and the armature coils. Also, the light-load adjustment was very similar to the tapped resistor used in the Type D. The lower bearing on this model, as with other high-capacity meters, was a cupped diamond. Some later examples of this model are marked as being made by the Roller-Smith Co., so there may be a connection between the two companies.
Photo: Electrical Meterman's Handbook

Stanley Instrument Co.

Stanley Model D

(1898 to 1903)
This model was an unique attempt to eliminate the jewel bearings that require periodic maintenance. By arranging a pair of magnets just so, a magnetic field was set up that was strong enough to float the disk in the gap between the magnets. To center it horizontally, the disk was mounted on a hollow steel shaft (the upper end had the pinion that drove the register) which had a fine piano wire strung through it. The full-load adjustment of this model was very similar to that of the GE IS switchboard meter (adjustable pieces on the ends of the magnets). Since there was technically no bearing system, no light-load adjustment was provided, and as was common with other early single phase meters of its time, no lag-load adjustment was provided. This model was made in only one capacity - 50 amperes - and instead of having the terminals in an enclosure, cables were run out of the meter's base to connect it to the circuit. This was one of a handful of models with the cover sealed at the factory as it was not intended to be adjusted in the field. 
Photo: Tommy Bolack collection

Stanley Model G

(1898 to 1903)
The Model G was a complete redesign of the D, and one main feature of the redesign was using separate magnets for the bearing and the disk brake. The two brake magnets were arranged with like poles touching, creating strong fields at their tips. A plunger was placed over the sleeve for the bearing assembly, and varying the position of this sleeve made for a sensitive full-load adjustment. The light-load adjustment used the typical adjustable punchings on the end of the voltage coil, and there was no lag adjustment. The main current terminals were on the sides (the right one was painted white to indicate line side, the left black for load side) and the shunt tap was on the bottom. This model was made in only 2 capacities - 15 and 25 amps, and like the previous Model D, the cover was factory-sealed. 
Photo: Northeast Utilities collection 

Stanley Model H (First Version)

(1898 to 1903)
The magnetic flotation system used on the previous models was not as successful as hoped, so it was abandoned and a regular jewel and pivot bearing was used. Otherwise, the overall design of the meter was similar to the Model G. This meter used a round cast-iron base with an internal partition that sat between the stator and the U-shaped brake magnet to protect the latter from demagnetization. Part of the shield was formed with projections that sat just below the ends of the brake magnet, and two large screws mounted in these projections (beneath the disk) acted as the full-load adjustment. The cover was cast from non-ferrous metal and fastened to the base with three sealable screws. The terminal box was located on the bottom with three terminals - two for the current coil and a third for the voltage coil tap. Since this meter was made in only one current capacity - 10 amperes - a three-dial register was used. No picture of this model yet. 

Stanley Model H (Improved Version)

(1898 to 1903)
As with the earlier version of this model, a regular jewel and pivot bearing was used, except the jewel mounting was rotated by way of a gear train coming off the left-hand register dial. Also, this is the only model made by Stanley Instrument Co. to incorporate a lag adjustment which was functionally similar to the lag coil used on other manufacturers' models. This meter used a round case with two terminal boxes: the lower one contained the current coil connections while the upper one was for the voltage coil tap. Another unique feature of this meter is the cover - it merely slides onto the base and a sealing wire is then used to seal it in place.
Photo: Unknown collection

Other Models

Willis Straight Line Meter

(1890s/1900s)
I have almost no data on either the company or this meter, but it seems Federal Electric was a supply house and that Willis Meter Co. is the actual manufacturer. This meter seems to be a commutator-type (as most DC meters are), but instead of using a "swinging coil" light-load adjustment as most DC meters do, this one apparently used tapped resistors (note the green leads around the register sides). Interestingly, the full-load adjustment is also accomplished with this tapped resistor arrangement (instead of just manipulating the position of the brake magnets). 
Photo by Brian Olearchik

Holcomb & Hoke Type A

(1910s)
Holcomb & Hoke Mfg. Co. appeared in Indianapolis shortly after Columbia disappeared from the scene. Only one watthour meter model was offered by this company, and I have no details of it as yet.
Photo: Detroit Edison collection

SEMCO Model 1

(1920s)
The only model offered by the Sewickley Electric Manufacturing Co. somewhat resembles the Sangamo Type H meter, but borrows heavily from other companies' models. The terminal box is similar to the one used in the Westinghouse B and OA meters, the disk assembly is similar to the GE I-14's, the register pickup assembly is similar to Duncan's M2 meter, and the retarding magnet is similar to the Westinghouse Round and Type A meters. The features unique to this model include a RED timing mark on the disk's edge, and an odd basic constant (5/24 as used on the Sangamo H instead of 1/3 or 0.6). Some of the glass covers used with this meter have 'SEMCO' prominently embossed in the cover. The metal covers resemble the ones used on the Duncan M2 meter. There are 3 styles of this model - an earlier version has the nameplate on the terminal cover and one example exists that apparently is a 'detachable' style as it has no terminal chamber but has a nipple with wires protruding.
Photo: Mike Andvik collection 

Stewart-Warner Model 746

(1930s)
This meter was conceived by some people within FDR's administration who thought the average price of $20 for a residential meter being asked by the existing meter manufacturers was too much, thinking $5 would be a reasonable price for meters being purchased by the rural electric coops that were just getting started at this time. Stewart-Warner (whose experience was mainly on various types of voltmeters, speedometers etc.) was the only company willing to try and make such a meter, and developed two models. The one shown cost a little more than $5, but was close enough to satisfy those who commissioned its design. The 746 shown is not complete; it mounted onto a breaker panel with the same finish, making for an integrated package when installed (never mind that the 'big 4' companies already had developed standardized mounting arrangements for meters). Stewart-Warner did produce an even more primitive model that met the $5 criteria and which was the 746-A; it has its own terminal chamber and a 3-digit cyclometer register. There are a number of features unique to the 746 - the first being the finish; it has a 'crinkle' finish over steel instead of aluminum alloy or lacquered cast iron bases as on other meters. The base and all internal support pieces are made of steel punchings (as done on the Westinghouse OB, for example). The register mechanism uses an absolute minimum of parts - there is only ONE gear between the disk and the register drums. The register drums are unique as the spindles that advance each drum are concealed between the drums themselves. Instead of the normal sliding or hook-type test link, the test link is a knurled nut on a screw that can be backed out to break contact with a spring. The most unique feature of this meter is the push button switch on the right side behind the cover interlock - this illuminates the light seen just to the right of the register. Despite efforts by FDR's administration to force rural coops to standardize on these meters, its poor performance in service and on the test bench and incompatibility with mountings made for meters made by the 'big 4' manufacturers doomed these two models. One coop in the southeast part of the US was reportedly so dissatisfied with this model, once all these were removed from service, they were promptly used as fill underneath the site of a new shop building. While a small number of 746 meters have survived, only a couple are known to exist with the fuse box attached. Even fewer 746-A meters have been confirmed still in existence.
Photo: David Dahle collection 

Roller-Smith Co. Type S3

(ca. 1948 to 1951)
Roller-Smith was an old and well-known company that specialized in making electrical instruments (and still exists today as TDE Alstom). They apparently acquired the design and tooling for the Stewart-Warner meter and reworked it into a socket version. The same accuracy problems that plagued the Stewart-Warner meter could have carried over into the S3, bringing about a quick end to its production. 
Photo by Bud Russell