There is a rule in tank development, especially in wartime: immediately start designing the replacement for the tank you just built. The Americans had to follow this rule in WWII. The situation with light tanks was fine, but medium tank production was just spinning up. Thanks to new model coming out annually, the Americans caught up to world leaders in medium tanks by 1942. The next step forward was the T20, which was radically different from its predecessors.
Revolution instead of evolution
The situation in Europe had a direct impact on American medium tank production. The Americans didn't expect France to lose the war so quickly, and the armistice signed on June 22nd, 1940, came as a surprise. Despite their neutrality, the USA understood that they will be drawn into the war sooner or later. The only question was: how much peace time do they have to prepare for war? The French campaign showed that the battlefield has changed, for instance 37 mm guns were no longer strong enough as protection of tanks increased.
The only way forward was to develop the Medium Tank M2 further. The Americans had no choice but to continue development of an existing design. Creating a new tank from scratch would cost time, and there was no way to replenish this precious resource. The result was the Medium Tank M3, a tank of compromises. The Ordnance Department knew very well that a 75 mm gun in the hull was not an optimal solution. The issue was that there was no turret of appropriate size available, nor a turret ring of required diameter, and so the strange hybrid tank was the only option. The vehicle was quite unusual, but could be put into mass production in the summer of 1941, and experience showed that the tank was not too bad. The Medium Tank T6, this time with a 76 mm gun in the turret, entered trials in September of 1941. This tank was put into mass production in February of 1942 under the designation Medium Tank M4A1.
One of the first concepts for the future Medium Tank T20.
The Medium Tank M4 was not just one of the most mass produced tanks in history, but also a dead end for the Medium Tank T5 chassis. The Ordnance Department understood very well that it was not possible to continue developing it forever. The Medium Tank T5 weighed 13.6 tons, the Medium Tank M3 weighed 27.9 tons, the M4 already weighed over 30 tons. Engineering experience showed that reinforcing the suspension can only take you so far. The layout of the Medium Tank M4 also had quite a few drawbacks. The front transmission meant that the drive shaft made the tank taller and took up precious space in the fighting compartment.
The logical idea of a replacement came up as soon as the Medium Tank M4A1 was put into production. Gladeon Barnes, Chief of Research and Engineering in the Ordnance Department, was one of the promoters of the new development. Barnes had his own idea of what a tank should look like, and it was quite different from what Harry Knox had in mind. Barnes wanted the transmission to be located in the rear of the hull. This made the hull longer, but resulted in better front armour and made servicing the transmission simpler. This solution also reduced the height of the hull. The first new generation tank built according to Barnes' concept was the Heavy Tank T1E2 (Heavy Tank M6). It turned out way too tall anyway, as the only appropriate engine was the air cooled Wright G-200 Model 781C9GC1, which was not particularly compact.
Various options for armament were considered. This drawing shows the 76 mm M7 gun, the same one as used on the GMC M10 tank destroyer.
The first work on replacing the Medium Tank M4 was done by the spring of 1942. The first steps were quite evolutionary. The desire to reused as much as possible from the existing tank hung above the designers like the Sword of Damocles. There were few alternatives to Harry Knox's VVSS suspension, and so the first drafts were reminiscent of the Medium Tank M4 with a rear transmission.
However, the Tank-automotive Center founded in Detroit developed a concept for a new tank that only kept the overall layout of the chassis by April of 1942. The Center, headed by Joseph Colby, raised Barnes' concept to a whole new level. In part, this was possible thanks to a new engine. The Medium Tank M3 already had the GM 6046 engine. This engine was still weaker than the Continental R975 C4, so a search for an alternative continued. Ford had the answer. The famous automobile manufacturer had an air wing, which developed a V-12 engine. This engine did not succeed in aircraft, but came in quite handy for tanks. The shortened 8 cylinder version was given the name Ford GAA. The 18 L water cooled engine had a nominal power of 450 hp, a little lower than the R975. The maximum power was 500 hp, close to the Soviet V-2 diesel engine. Production of the M4A3, a Medium Tank M4 with this engine, began in June of 1942. This engine also allowed the potential replacement for the M4 to have a much lower hull.
Trials of the 76 mm Gun T1, August 1942. This gun was later accepted into service under the index M1.
In addition to a lower hull, the new tank needed more powerful armament. The M34 gun mount used on the M4 tank was initially designed to carry more powerful armament than the 75 mm M3 gun. This gun was already available: the 76 mm Gun T9 with the ballistics of the 3" AA gun was developed for the Heavy Tank T1. Some consider this gun the result of Colby's trip to North Africa where the Germans used new Pz.Kpfw.III Ausf.L and Pz.Kpfw.IV Ausf.F2 tanks. This theory is quite suspect, as both tanks were first used in late May of 1942, by which point the requirements for the tank known as the Medium Tank T20 were already finished and a full scale mockup was built. The Americans likely foresaw new German tanks, especially as there were already several mounts for 76 mm guns.
The Ford GAA engine was chosen for the new tank. This engine was already in use on the Medium Tank M4A3.
The final requirements were as follows. The mass was on the same level as the Medium Tank M4, but the silhouette of the tank was lower. The crew composition was the same, but the driver's controls were duplicated. An exhaust fan was installed in the driver's compartment roof. The armour and armament were upgraded. Specialists from General Motors joined the project, and thanks to them the engine compartment was laid out in a completely new way. The idea of combining transmission elements into a single unit that could be removed through the front armour was first tested on the Light Tank T7. The new tank developed this concept further: the engine, gearbox, and other transmission elements were all a part of one assembly. The transmission group included a planetary gearbox and torque converter (the latter was also tested on the Medium Tank M7).
A full sized model of the Medium Tank T20 that was presented in late May of 1942.
A full sized model of the Medium Tank T20 was presented in late May of 1942. The new vehicle impressed the high command. The thickness of the front armour was increased to 64 mm while remaining at the same angle as on the Medium Tank M4. The height of the turret roof was 2438 mm (t o compare, the M4A3's roof was at 2642 mm). The turret front was 89 mm thick. A 76 mm gun was installed. The tank looked quite promising. The ASF and armoured forces authorized one prototype of the Medium Tank T20 to be built. The Ordnance Department gave orders to build the tank on May 25th, 1942. Work on the new generation medium tank went from theory into practice, even though many changes awaited the T20.
1942 was not just a year when the USA launched its most numerous tanks in production, but also the year when a crisis of ideas took hold. This began even earlier, during the development of the Heavy Tank T1 in 1940-1941. There was no agreement on what engine or transmission was to be used on the tank, and so work split into several directions. One of the options was the GM Hydra-Matic automatic gearbox, which was never delivered. A whole spectrum of gearboxes was developed for the Heavy Tank T1: hydraulic, mechanical, even electric. This delayed the work on the tank for so long that by the time the army made its choice it was clear that this tank was no longer needed.
Experimental Medium Tank T20, June 1943.
The requirements for the Medium Tank T20 began to wander in the summer of 1942. This affected not only the engine compartment, but also the armament. The 76 mm gun remained the highest priority, but alternatives were being explored. On August 26th, 1942, the Ordnance Committee recommended building three experimental tanks: Medium Tanks T20, T20E1, and T20E2. The work was approved on October 12th. The T20 kept its 76 mm gun, which was later accepted into service as the M1. The T20E2 got a different 76 mm gun, the 3" M7 that was used in the GMC M10 tank destroyer. This was not the first time this gun came up, as the draft of the T20 already had this gun in the gun mount from the GMC T35 (GMC M10 prototype).
Thanks to a new layout, the tank was much lower than the Medium Tank M4.
The Medium Tank T20E1 was the most interesting. It received the 75 mm M3 gun with the same M34 mount as on the Medium Tank M4. The difference was an autoloader, the first of its kind that was actually built in metal. The ammunition was loaded from racks on the fighting compartment floor. The rate of fire reached 20 RPM. The tank carried 64 rounds of ammunition split between two 32 round racks. One carried HE shells, the other AP. The use of an autoloader reduced the ammunition capacity (the T20 and T20E2 carried 68 76 mm rounds), but the benefits outweighed the drawbacks, especially considering that the 75 mm M3 gun was still modern at the time.
The Medium Tank T20 looks larger than the M4, but the length of the hull was about the same.
The suspension deserves a separate mention. Initially, the tank was supposed to have the same suspension as the M4 with horizontal volute springs. This system was not ideal and the tanks only got heavier. Harry Knox foresaw these issues and developed a new variant of his suspension. The springs were now horizontal, and the bogey was changed. Knox didn't come up with anything revolutionary, he essentially took the Horstmann suspension that dated back to the late 1920s and upgraded it. This didn't stop Knox from patenting his design (although all of Knox's patents went to the government anyway).
This tank uses an early variant of the HVSS suspension.
Changing the suspension was not enough. The military saw the benefits of a torsion bar suspension, which Kolby got to see up close in North Africa when he examined a Pz.Kpfw.III tank. It was difficult to dethrone Knox, as his benefactor, General Christmas, sat at the very top of the Ordnance Department and dictated the design of American tanks. Nevertheless, Kolby and Barnes formed a serious opposition. Various options were developed in Detroit, as a result of which it became clear that a torsion bar suspension was the optimal way forward. Armoured force commanders also insisted on a torsion bar suspension starting in September of 1942. Like any new idea, the torsion bar suspension still needed to be worked out, and so on October 2nd, 1942, the Ordnance Department authorized work on yet another tank. This tank got the index Medium Tank T20E3 on February 12th, 1943.
The Ford GAN engine and elements of the transmission, including the gearbox, were combined into one power pack.
The contract for Medium Tank T20 prototypes was given to the Fisher Body factory, one of the largest automotive body manufacturers and a subsidiary of General Motors. This factory was the main producer of M4A2 tanks as of the spring of 1942 and also took part in various experimental works due to its proximity to the Tank-automotive Center. Various "optimizations" introduced a significant delay into development of the tanks and resulted in cancellation of half of them. The T20E1 and T20E3 remained on paper, and the only T20 was finished in May of 1943, a year after work was approved.
A turret with an autoloader mechanism was planned for the Medium Tank T20E1. It was built, but installed on the T22E1 tank instead.
Despite the significant delay, the T20 was still among the best tanks of its class. It was only surpassed by the Panther Ausf.D at the moment of completion, and the German tank was significantly larger and 50% heavier. The new American tank weighed only 29,827 kg, half a ton less than the Medium Tank M4A3. The chassis of the T20 turned out to be not just lower, but also shorter than the M4A3: 5740 mm vs 5906 mm. The T20 was wider (2920 mm vs 2620 mm), but the widening mostly took place due to bringing the suspension bogeys further out. There were no longer sponsons that hung over the tracks, and the hull was narrower. This was the secret of why the new tank was lighter than its predecessor. The hull was welded together from many cast components. This was especially noticeable in the front, where the upper part of the driver's compartment, machine gun mounting, and idler carriers were cast. This was not unusual for American tank building. The upper front hull of the Medium Tank M4 also had many cast parts welded into it. An exhaust fan was installed between the driver and assistant driver's hatches. The cowling noticeably protruded above the driver's compartment roof.
Medium Tank T20E3, the second tank built as a part of the T20 program.
The engine and transmission compartment was laid out very originally. The roof was built as several panels that could flip up to service various components easily. As planned, the engine, gearbox, and differential were combined into one unit. The engine was redone to reduce its height. The new variant received the index Ford GAN. The tank used the Alisson Torqmatic planetary gearbox with a torque converter. A similar layout was used in another vehicle developed by the Tank-automotive center: the GMC T70. As initially envisioned, the tank had dual controls. The driver and assistant driver's seats could lift up during travel, which improved visibility. As for the running gear, with the exception of the suspension it was very similar to that of the Medium Tank M4. The drive sprockets, road wheels, idlers, track links were all taken from the production tank.
The fenders changed as did the running gear.
Unlike the chassis, the fighting compartment was the result of evolution. The turret was based on that of the Medium Tank M4, but its shape changed. The turret bustle was enlarged to fit the bigger gun, and some changes were also introduced in the front. The commander's hatch, sights, and parts of the gun mount, and turret ring were borrowed from the Medium Tank M4. This made the two turrets interchangeable. The turntable remained, but it had openings, unlike the M4. The loader also got a two flap hatch.
This tank had metallic track links, similar to those used on the GCM T70.
After a QA run by the Ordnance Department, the tank with registration number U.S.A. 30104302 was sent to the General Motors proving grounds in Milford, Michigan. The tank drove for 480 km in total, and this was not an easy distance to cover. On one hand, it was faster than the Medium Tank M4A3. The top speed was 56 kph, a record among medium tanks of the time. On the other hand, reliability was in question. There was a whole spectrum of defects found in the transmission. Various breakdowns, oil leaks, overheating, and breakdown of the brakes were recorded. The tank was simply not ready for trials. It took 385 hours to correct these defects, after which the tank drove for another 100 km. The tank arrived at the Aberdeen Proving Grounds only in February of 1944 when it was no longer of any interest to the Ordnance Department.
Trials showed that the torsion bar suspension did not work as well as the HVSS.
The second prototype, the Medium Tank T20E3, was ready by July 1st, 1943. This tank with registration number U.S.A. 30104303 turned out noticeably heavier. It weighed 30,617 kg. Such was the cost of a torsion bar suspension. Experience from studying the T-34 and KV-1 was used when designing the running gear, as well as prior work on the T70. The number of road wheels increased to 6 per side and the design was different than the T70's wheels. The track links were similar, but they were now 457 mm wide. The idler and return rollers were replaced. The 1st, 2nd, 5th, and 6th wheels received telescoping shock absorbers. There was also a system that linked the idler and rear suspension arm together. This allowed the track to remain in constant tension.
The transmission issues did not disappear. Both tanks suffered from the same defects.
The fate of the Medium Tank T20E3 was in many ways similar to that of the T20. Factory trials showed the same transmission and engine cooling issues. There were also additional growing pains with the suspension. The shock absorbers broke, the bolts tore, and many other problems cropped up. As a result, the Medium Tank T20E3 also ended up in repairs from which it emerged only in the early spring of 1944. On March 15th, 1944, the tank arrived at the Aberdeen Proving Grounds where it was used for suspension testing. The military also lost interest in this vehicle by then.
New gearbox, old problems
As mentioned above, the Americans were in search of an optimal transmission layout in 1941-42. The story of the Heavy Tank T1 taught them that it would be nice to have a simpler and more reliable alternative. As a result, by the fall of 1942 the Medium Tank T20 split into several projects that had not just different guns, but different transmissions.
The first Medium Tank T22 pilot at the Detroit Tank Arsenal proving grounds, May 26th, 1943.
The first split took place on September 3rd, 1942. The Ordnance Committee recommended the development of a medium tank analogous to the T20, but with a mechanical gearbox. The name of the tank was T22, not T21 as one could have guessed. The Americans already reserved that index for the Light Tank T21, which was a T20 with reduced armour thickness and a 75 mm gun instead of a 76 mm to bring the weight down to 20 tons. The tank would be lighter and faster. As it often happened, the tank's mass kept climbing during development. The American military could not resist boosting the tank's armour and bringing back the 76 mm gun. As a result the tank's mass rose way past 20 tons and there was no longer any sense in building it. The Light Tank T21 program was closed in July of 1943.
Assembly at a different plant had its impact on the insides and outside of the tank.
The Medium Tank T22 evolved along a different path. Like the T20, it was split into multiple variants. The T22 was given the 76 mm M1 gun, the T22E1 had a 75 mm gun and an autoloader, then the T22E2 had the 3" M7 gun. These projects were approved on September 17th, 1942, although the T22E2 quickly vanished from the list, since the M7 gun was much bulkier than the M1. The T22 kept its 75 mm gun with an autoloader mechanism.
The rear of the hull was altered.
The difference lay not only in the transmission, but also the manufacturer. Unlike the T20, the tank with a mechanical gearbox was built at Chrysler, or rather the Detroit Tank Arsenal that it owned. This had an effect on the look of the tanks. Both vehicles were built by late May of 1943 as T20s, meaning that they had 76 mm M1 guns. The prototypes were given registration numbers U.S.A. 30104304 and U.S.A. 30104305.
View from above with maintenance hatches open.
The design of the hull and running gear was nearly identical to that of the T20, but this vehicle can still be distinguished from its predecessor. The tank received unique fenders with stampings for rigidity. Second, the tank had different toolboxes on the fenders. Finally, the tools stored on the engine deck were different, and the deck itself was visually distinct.
The power pack on the Medium Tank T22 used elements from the M4A3 tank. Trials showed that it was more reliable than the T20, but there were other issues.
The biggest differences were on the inside. Chrysler didn't reinvent the wheel and essentially adapted transmission elements from the Medium Tank M4A3. This meant that the Ford GAN engine was coupled together with a 5-speed synchronized gearbox and a differential. The engine compartment was big enough for this combination, but thanks to the enlarged fenders the full length of the tank increased to 6102 mm. The mass rose with the size, coming up to 31,434 kg.
The T22 turret was put through initial trials in a test rig.
One tank of the two was sent to the Aberdeen Proving Grounds. The second was sent to the factory proving grounds where it drove for 1440 km. The mechanical transmission was more reliable, but the "family curse" didn't go away. The issues with the engine cooling system were the same as on the Medium Tank T20. The top speed also dropped significantly to 40 kph. This meant that the tank had no advantages in mobility over the Medium Tank M4A3.
The fighting compartment of the T22E1. Quite cramped, but at least there was no loader.
The engine on the first prototype broke, which was the rule rather than the exception. The second prototype behaved similarly at the Aberdeen Proving Grounds. The issues were more severe, as neither the engine nor the gearbox passed trials. The whole power pack would have to be replaced. Due to poor trials results the T20 and T22 looked like a bad option and a worse one. There were also a fair number of complaints about the fighting compartment. The new turret made the fighting compartment larger, but not large enough. There were also complaints about observation, plus the T79 gun mount was deemed inadequate. There was no point in a thorough correction of the tank's defects.
The T22E1 after conversion. The tank was sent to gunnery trials in this form.
The alternative armament was being developed in parallel with the chassis. After the 3" M7 gun left the stage, the 75 mm M3 with an autoloader remained as the only alternative. As this gun could fit into a smaller fighting compartment, a decision was made to develop a new smaller turret. The Medium Tank M4 turret was converted by the United Shoe Machinery Corporation. There were some differences, but not many. For starters, the commander's cupola was moved to the left side of the roof, and a single flap hatch was added for the gunner on the right. Thanks to the autoloader, the amount of crewmen in the turret was reduced to two.
Since the gun was smaller than the 76 mm option, the turret on this tank was also smaller.
The first tests of the 75 mm gun took place in November of 1942. The turret was supposed to go on the Medium Tank T20E1, but since that tank was cancelled it was installed only on the Medium Tank T22E1. The turret was tested on its own in the spring-summer of 1943, largely to tune the autoloader mechanism. Finally, the turret was installed on the experimental tank renamed T22E1 in August of 1943.
The final variant of the Medium Tank T22E1. A counterweight is installed on the barrel to ensure proper function of the stabilizer.
The tank was sent to the Aberdeen Proving Grounds in November of 1943. The interest towards the 75 mm gun was waning, as it was not powerful enough to combat the latest German heavy tanks. The autoloader was still interesting, but the system proved insufficiently reliable and could not be improved. In February of 1944 the Ordnance Committee ordered that work on the T22 be cancelled and the T22E1 specifically be frozen.
Unlike the T20 and T22 turrets, the commander was located to the left and the gunner was to the right.
Despite this sad engine, the Medium Tanks T20 and T22 were a large step forward for the American tank industry. Unsatisfactory results in trials led to some not entirely correct conclusions. There was another option: the T23. The Medium Tank T23 entered development alongside the T22, and this project was considered a high priority. Work on the T23 started well, but later it turned out that the success was partial. The failure of the T23 deserves a separate telling, but it ended up one of the reasons why the American tank program slipped and the Americans fell behind by the end of WW2.