PART III

The Process as used at the Recalcitrant Press

The requirements for electrodeposition of type matrices are conceptually simple... a piece of type to be duplicated, a holder to position it, a tank of electrolyte and a power supply; but the variations within these requirements are almost unlimited. In this section I will comment on the methods that have worked for me, methods shared by a number of generous friends, and a few that have occurred to me in the process of solving practical problems. Electrotyping is still part art and part science so far as the occasional matrix maker is concerned. Only by actually depositing and finishing matrices will the reader learn how to apply this advice.

The first step in the process, as in the preceding description, is to clean and prepare the type to be duplicated. Since most of the type private founders are likely to be duplicating has been used, cleaning is necessary. I soak ink-encrusted type in household ammonia containing a mild detergent. I can usually find a bottle of "Parson's Sudsy Ammonia" in my wife's kitchen cabinet. After an hour's soak, a gentle scrub with an old tooth brush will usually remove all traces of ink; if not, an overnight soak may be necessary. A ten power magnifier is a necessity for this and other inspections in the course of matrix making. Use it to make sure the type is REALLY clean, especially in the shallow lines of ornaments. As before - any imperfection in the original will be faithfully reproduced.

The first necessary piece of apparatus is an accurate fixture for mounting the type. The Fusible Casting Mould served the purpose at OUP, but few private founders will have pivotal casters, and so will normally wish to deposit other than foundry matrices. The most common matrix among my acquaintances is the Thompson style or (U.S.) Monotype Display matrix. My own caster is a Monotype-Thompson, so this is the pattern I follow.

The critical dimensions of this style matrix are the head bearing and "left" side bearing (looking at the matrix with the nick side of the type down). For Thompson matrices, the side bearing is 8 points in all cases, and the head bearing is 24 points for type under 30 points body size, and 18 points for body sizes 30 points and above. The thickness of the matrix is about .094", and the depth of drive .050". The overall dimensions for most sizes are 3/4 inches wide by 1 1/8 inches high. For matrices over 42 points in set, a wider matrix is required, but normally should not be over 1 inch (the maximum casting width of a Monotype Thompson is 56 points).

Monotype display matrices vary in head bearing from size to size. the reason for this is that these matrices were designed to be cast on the Monotype Sorts Caster (aka "Orphan Annie" from its serial numbers which started with OA). Since the sorts caster used a centering pin, as did the composition caster, the molds were designed to center the face over the mold opening. Thus the head bearings varied as follows:

* 42 and 48 point matrices are for Thompson only. Display caster casts up to 36 point only. In the larger body sizes or in ornaments over 42 points in set width, the matrix is made wider so that adequate side bearing is maintained on the right side.

Generally there are two methods of depositing matrices. One is Starr's method (Appendix 1) where a copper or brass matrix blank is used. The blank has a hole routed in it with sloping sides (about 45 to 60 degrees). The fixture holds the type centrally in this hole, and the rest of the blank is masked off with wax so that copper is deposited only in the "eye". This method is economical of copper anodes and reduces the necessary current, considerations in large scale operations. Also, with careful adjustment of the depth of the type in the base, surfacing of the finished matrix is minimized. The disadvantage of blanks is that they must be prepared, requiring preliminary machining operations.

Andy Soule' uses this method, mounting the type in type metal holders which he casts in a special mold. He routs the blanks on his lathe in a milling fixture, using a 1/8" milling cutter that has cutting lips in its shank which is 1/4" diameter. Andy also has had made an ingenious finishing fixture that surfaces the matrix, trims it to size and bevels the corners... But this degree of sophistication was beyond my mechanical capabilities when I bagan matrix making.

The second method, and the one I use, is to deposit the complete matrix on a prepared base in a manner similar to that used at the OUP. The principal difference from Oxford's method is that the base is a machined, reusable part in which the type is mounted, rather than a type metal casting. This method was introduced to me by Paul Duensing and Andrew Dunker and for me has been the simplest way of doing the job. The basic advantage of this method is that no prepared blank is necessary, but the final machining of the matrix is potentially more exacting. So there is still no "free lunch"!

In either method, accuracy in the mounting of the type is of primary importance. Any sloppiness at this stage multiplies the problems of finishing and justifying the matrix. It is critical that the type carrier establish head and side bearings with close tolerances. If possible, these dimensions should be held to +.002 -.000 inches. The positive allowance is so that metal can be removed in the justification process. It's difficult to put more metal on a matrix once it is out of the tank! Also, the face of the type must be accurately parallel to the surface of the carrier to avoid a skewed face. This is taken care of by making two surfaces of the slot or hole in which the type is clamped accurately square with the depositing surface.

Appendix 4 gives details of a depositing case designed by Andrew Dunker. Most of my matrices have been deposited in a similar case made by Andrew, and loaned to me by Paul Duensing. Andrew has commented with characteristic modesty, "It is probably the design that only a tool maker would think of using, but it has proved to be quite useful in practice." Indeed it is another of those devices that serve so well because of the quality of its design.

Copper is deposited directly on the surface of the type carrier. The side holes direct additional current to the area between the face of the type and the side of the case, an area that otherwise would receive less copper than the more exposed areas.

The dimensions of the carrier determine the critical head and side bearings, outside dimensions and depth of drive. About .005" overdrive is designed to take care of type that may vary slightly in height to paper. This much overdrive is difficult to remove by filing, so machining is necessary. If accurate measurement of the distance the type face protrudes from the base can be made, and overdrive adjusted to .002", hand finishing would be more practical. Don Turner removes .004", but he is a better filer than I am!

For type of 24 points or less a carrier accommodating a 24 point quad body is convenient, and simplifies the selection of spacing. The type to be duplicated is mounted in the carrier using spaces and quads to fill in the space around the type. If the slot for the type is accurately machined the type plus spacing will fit snugly.

Larger size (up to 48 point) carriers may be prepared for Monotype/ Thompson matrices. but the layout of the screw holes will be slightly different from those shown in the appendix. Forty eight point carriers should be made about 1 3/8" wide or wider to give adequate support to the head end of the carrier.

If the type carrier is made of brass or bronze, the face and sides should be treated with a parting agent. A recommended solution (1) is 7.6 g./l (1 oz./gal.) sodium sulfide in distilled water. The carrier should be soaked in this solution for a few minutes to allow the sodium sulfide to react with the metal to form copper sulfide, in the case of a bronze carrier. Alternatively the carrier may be tinned with 50/50 solder. On carriers that I have made, I heat the bronze with a propane torch, coat it with solder and wipe off the excess with an old towel before it cools. The lead/tin alloy will not accept an adherent deposit of copper, and so serves as a parting agent. Andrew Dunker has suggested making the type carrier of 18-8 stainless steel. No parting agent would be required, but stainless steel is more difficult to obtain as well as to machine.

After mounting, the type should be snug against the side plate and head end, with its feet firmly on the base plate. This assures accurate side and head bearings, and avoids excess overdrive. The same hazard of copper growing under kerned characters exists with this method as with OUP's. The remedy is to warm the type and mold a little wax under the kern. If both the type and the carrier are warm when assembled, any excess wax will be squeezed out.

I find it convenient to warm the type and carrier under a heat (infra-red) lamp while mounting the type. The temperature can be controlled by the distance from the lamp, and the wax kept soft but not at the melting point. A small soldering copper with a lamp dimmer for heat control will melt wax where needed, and act as a brush to "paint" wax on the type carrier.

After the type is mounted, carefully trim away the wax that squeezes out around the kern. A copper space (1/2 point) or a small stick sharpened to a flat point works well. Try to preserve the slope of the beard of the type when trimming off the excess wax, and be sure none is left on the top of the carrier. Now brush the wax with graphite or bronze powder, using a small soft brush. This makes the surface of the wax conductive and allows the copper to bridge an otherwise non-conductive gap.

Coat the sides and bottom of the carrier with a thin layer of wax to avoid copper building up in these areas. The electrolyte will creep into the smallest space and leave a film of copper. Also be sure to fill all screw counter bores with wax, including those in the interior parts of the carrier. After the Plexiglas side and end plates are assembled, wax all joints and screw holes, and inspect the face of the carrier and type once more. If any excess wax is found, remove it carefully. The wax I use is common household paraffin, the kind sold by grocers for sealing jams and jellies, mixed about half and half with bee's wax.

Cleaning the type was mentioned in part I, and in general, the cleaner the type the better. However, really clean type presents problems of its own: the copper tends to bond to the type, making the matrix difficult to separate from the type. A parting agent should remedy this, although I have not yet tried one. Sodium sulfide as recommended for the carrier should be effective.

After properly preparing and mounting the type, copper is electrodeposited to form the matrix. The necessary apparatus is a tank, electrolyte and a direct current supply. Most references to electroplating or electrotyping relate to commercial operations which require large volumes of electrolyte and describe large lead lined tanks. The private founder will rarely need more than one or two gallons of electrolyte, so small tanks are in order. I began with two quarts of solution in a gallon jar. The tank must be made of a material that will not be attacked by the sulfuric acid in the electrolyte. Glass is satisfactory, although fragile. For a time I used a Pyrex loaf pan liberated from my wife's kitchen, but it was a bit too shallow (only 3" or so). Oxford's tank is rubber lined, and most commercial tanks are wood or metal lined with asphaltum or rubber.

The most satisfactory containers that I have found are the clear plastic tanks from the large lead-acid cells used in Telephone Central Office power plants. If you know a person that works for the Telephone Company, they might be able to direct you to the junk dealer that buys the discarded batteries from the company. You should note that in the United States, battery electrolyte is classified as a hazardous waste, and strict regulations are imposed on its disposal. Try to get empty cell cases if you can (it is the lead plates that the junk dealer is interested in, so this is usually possible). My tank is approximately 5 by 9)1/2 inches by 11 inches deep. With one gallon of electrolyte, the solution is about five inches deep. This tank gives adequate space for at least eight depositing cases, more than enough for my needs.

The electrolyte is a solution of copper sulfate and sulfuric acid in distilled water. I use no addition agents, and really have not found them necessary or desirable, although Paul Duensing and Andy Soule' use a bit of molasses as a leveling agent, and both are pleased with the result. Apparently the solution composition can vary over a moderate range without ill effects, but as in any technique, reasonable precision gives more predictable results.

The concentration I use, given by Phil Nuernberger(2) is 178 g./l. (23.8 oz./gal.) copper sulfate and 29 g./l. (3.9 oz./gal) sulfuric acid. These are measured by the specific gravity of the solution(3). I begin with distilled water in the quantity needed for the tank, and add the number of grams of copper sulfate indicated above. For example, for two gallons of electrolyte about 48 ounces of copper sulfate will be required. I add the copper sulfate to the water and stir until it is completely dissolved. Then I measure the specific gravity of the solution, and adjust it to 1.103 by the addition of water (if too high) or copper sulfate.

Concentrated sulfuric acid is extremely corrosive. When handling even small quantities, I wear splash proof goggles, rubber gloves and protective clothing (no short sleeves).

NEVER POUR WATER IN ACID

Always pour the acid slowly and carefully into the copper sulfate solution. I work near a source of running water, and immediately flood any spills with water and neutralize spills with baking soda (sodium bicarbonate)

but on skin use water only.

Now I add sulfuric acid to bring the specific gravity to 1.124. The amount needed can be calculated as follows: The specific gravity of technical grade sulfuric acid is about 1.840. This means that one fluid ounce ( a volume measure) of the acid weighs 1.84 ounces. In our example, for two gallons of solution we needed about 8 ounces of acid. Dividing by 1.84, we find that roughly 4.5 fluid ounces are required. In this case, add less acid than you calculate, measure the specific gravity, then add more acid to bring the solution to the proper gravity. Note that in both cases slightly more chemicals than calculated are required because their addition increases the total volume of solution. When you put the solution in the tank, mark the solution level on the side . All that is necessary to maintain the solution to its approximate original concentration is to add distilled water to this level.

The whole subject of solution control brings up a curious belief of electrotypers before modern control methods evolved. It was apparently common for electrotypers to believe that the electrolyte accumulated excess free acid.(4) The usual remedy for this was to discard a part of the solution, then add water and both acid and copper sulfate to bring the solution back to standard. As a matter of fact, it is the copper sulfate that accumulates in the solution, not the acid.

In a typical electrodepositing solution the anode efficiency is greater than that of the cathode. This means that more copper is dissolved from the anode than is deposited on the cathode, causing the accumulation of copper ions in the solution. These combine with sulfate ions from the sulfuric acid to form copper sulfate. This results in the accumulation of crystals of copper sulfate in the tank and on the suspension rods or wires - in fact on anything in contact with the electrolyte.

For the small operation typical of a private type foundry, occasional filtering of the solution to remove anode sludge will also remove the excess crystallized copper sulfate, and maintenance of the original solution volume will keep the concentration in tolerable range.

For the operator who prefers more systematic methods, solution testing will allow precise adjustment of the acid/copper sulfate content. Assuming that no addition agents are used, i.e. the bath contains only water, sulfuric acid and copper sulfate, the specific gravity of the solution is independent of the relative concentrations of acid and salt.(5) Thus if the specific gravity is known, we also know the total content in grams per liter (or ounces per gallon) of acid and salt. The amount of acid can be determined by titration, and then a simple subtraction yields the copper sulfate content. Appendix 3 gives details of the procedure and apparatus required.

After this determination, a suitable amount of solution can be discarded and distilled water added to make the copper sulfate content correct. The addition of acid to bring the electrolyte back to its original specific gravity completes the correction. Alternatively, lead anodes can be operated in the solution to 'plate out' part of the copper sulfate. All that is required is to connect a lead anode in parallel with the copper one, although operating with a second power supply and a separate lead anode and copper cathode allows better control. Again, chemical testing is necessary to control the solution makeup.

This method appeals to my conservative nature, as copper is recovered from the solution and waste minimized. However, lead anodes do present one significant problem: oxygen is formed on the anodes, and is evolved as almost microscopic bubbles. These create a very fine mist of electrolyte that coats the nearby tank areas, and is potentially hazardous if inhaled by the operator.

Now you need power supply. You will not require a dynamo unless you are going to go commercial, so a supply providing a few amps at one to three volts will do. Appendix 5 gives the details of a three amp. supply I designed; you may wish to find parts of about the same values, or design your own. If you design the supply, the minimum current for one matrix is about 50 milliamps at .20 volts. The supply need not be capable of supplying more than one volt, the usual depositing voltage being between .4 and .5 volts. At .4 volts, the current per Thompson style matrix should be about 100 ma. with the above electrolyte.

Meters for measuring both tank voltage and current are convenient, although it is possible to work by voltage alone. I arranged the voltmeter on my supply for two ranges - five volts and one volt. Normally the meter is on the five volt range, so that if the tank circuit goes open, the meter is not overloaded. During normal operation a push of the button changes the range to 1 (or .5) volts full scale.

Fortunately for the occasional matrix maker, acid copper electroplating does not require precise controls. Satisfactory deposits can be had over a range of current densities as well as solution strengths. In general, the lower the current density, the less tendency there is for nodule growth and treeing, i.e. the deposit is smoother. But low current density means slow growth, so a compromise value is sought. The basic objective is to grow a deposit of adequate thickness before treeing becomes objectionable.

Most texts give current ranges for electrotyping, in which a moderately thin (.010) shell is deposited in as short a time as possible. To be able to use high current densities strong agitation is usually recommended. The problems of electrodeposition, with its very thick deposits, are usually different. Since a very thick deposit is required, treeing and growth of nodules is a significant problem. Moderate current densities and still electrolyte baths are required. The National Bureau of Standards has published technical papers on electrotyping (see bibliography) that are helpful to matrix makers.

Circular 387 is the most comprehensive and tables from it are included in appendix 3. For still baths this document recommends current densities of 2 to 3 amps per square decimeter (a./dm.)(2). This translates to about 109 to 163 ma. for a 1)1/8" by 3/4" matrix. In my tank these currents are achieved at .38 and .50 volts respectively. The lower voltage / current yields a slightly smoother deposit. Still lower voltages seem not to improve the deposit substantially, and result in longer depositing times. Voltages much over .5 volts lead to rather wild growth, and are best avoided, although the more adventurous may like to experiment using leveling agents such as glue or molasses. If you have minded the details of mounting the type and waxing the fixture, depositing should go smoothly. I deposit at a constant voltage of .4 to .5 volts, depending usually on my travel schedule. The higher voltage deposits an adequate thickness in about 6 days, the lower in 8.

One of the more important considerations in electro-depositing is the connection to the power supply. With the low voltage used, any appreciable resistance will reduce the current to unacceptable levels. I use 22 gauge plastic insulated wire, and make the lead long enough to go directly to the binding post of the power supply. A plastic clothes pin clamps the wire to the top edge of the tank, suspending the case just below the surface of the solution.

For anodes, I use pieces of scrap copper tubing, or sometimes electro-refine my own anodes using chips from matrix finishing. Almost any piece of copper that is relatively clean will serve (avoid contamination of the electrolyte from organic substances on scrap anodes). Generally it is best to have the anode and cathode areas approximately equal, but this is not critical.

As deposition progresses, bits of copper and sludge from anode impurities will fall to the bottom of the tank. Periodic filtering of the solution will remove sludge, or the anode may be bagged. Dynel (a synthetic fabric) is recommended by some texts, but most of the polyester fabrics should serve. Experiments will show which are suitable for use in the electrolyte. I have never found the need for anode bags.

Check the case occasionally during deposition to see that all goes well. Any pinhole in the wax will plate through, and you will find a "flower" growing on the case. This does no real harm, but robs the area we wish to plate of copper, extending the depositing time. Flowers may be waxed over to stop their growth, but try to keep the face of the deposit covered with electrolyte while waxing. If the case is out of the tank for any length of time, some oxidation will occur. Let the case soak in the tank for about 10 to 15 minutes, or give the matrix about two or three minutes of deplating (if it is sufficiently thick to be sure you don't deplate the carrier).

When the deposit is thick enough, remove the case from the solution and wash it thoroughly under running water. This minimizes corrosion, which will occur rather quickly after the case is out of the solution. Now separate the matrix from the base.

Using a Dunker style depositing case (detailed on sheet 1 and sheet 2, linked here), I remove the Plexiglas sides and the side bearing plate, then lay a matrix blank or similar tool against the side of the base and strike the matrix a (not too) sharp blow to detach it from the base and type at the same time. Usually the matrix has a bit of flash around the sides, and a raised ridge formed by the bevel in the opening for the type. A few strokes with the file squares the edges, then the matrix is laid flat on the file to dress the face flat. An initial measurement may then be made with a depth gauge(6). If all has gone well, the matrix should measure about .055" deep. It is possible to rub down the face of the matrix to get the proper depth of drive, but a small metal lathe will do the job quicker. I designed a fixture that screws on the nosepiece of my lathe and finishes two matrices at a time. It is illustrated below.

The side clamp plates have four screws, two to push out the lower edge of the plate and two to clamp the plate to the matrix. This tends to clamp the matrix against the face of the as well as the center abutment. The routed slot and recess in the areas where the matrix might have high spots also helps assure that the matrix lays flat on the fixture. A similar fixture I designed will accommodate four matrices, and mounts on the lathe face plate. Doubtless anyone that pursues matrix making to this point will have his own ideas for a finishing fixture.

My method of finishing is to mount the matrix with its face toward the plate and turn the back surface until it is just smooth, without regard for total thickness. I then turn the matrix over and take a light cut across the face of the matrix. After a depth measurement I take further cuts to bring the depth to .0505"+. For close control the compound of the lathe is set to 30 degrees, giving .0005" feed for each .001" on the compound. The matrix is then turned over again and brought to final thickness. Remember that the thickness of the matrix affects the centering pin travel in the composition caster, and that the Thompson compresses a rather strong spring to seat the matrix on the mold. Make the thickness .094" as closely as possible.

If you plan to use this style matrix in a Monotype Composition or Sorts Caster (Orphan Annie), two corner bevels are required. They are not needed if you cast only on a Thompson.

Return to index

(1) William Blum and George B. Hogablum, Principles of Electroplating and Electroforming (Electrotyping). New York: McGraw Hill, 1924, p.142. Also recommended are solutions of sodium bisulfide, sodium chromate or sodium bichromate, the last two at dilutions of 4 g./l (1/2 oz./gal.)

(2) Phil T. Nuernberger. Electrolytic Matrices. Kalamazoo: The Private Press & Typefoundry of Paul Hayden Duensing, 1966, p.4. This is the "slow tank" concentration. (2) on next page.

(3) H. D. Holler & E. L. Peffer, The Relation Between Composition and Density of Aqueous Solutions of Copper Sulfate and Sulfuric Acid. Bulletin of the National Bureau of Standards, Vol. 13, p.273, 1916. Tables from this document are included in Appx. 3

(4) Nuernberger, p.8, and Blum & Hogablum, p.193

(5) Regulation of Electrotyping Solutions. Bureau of Standards Circular 13, 1915.

(6) I use a "Geneva" photoengraver's depth gauge, manufactured by the Chicago Dial Indicator Co., P.O. Box 1068, Des Plaines, Illinois, 60016. (Address correct, 1981.)