GENEVA
The Mercedes 230SL made its public debut at the Geneva Auto Show in March 1963, although Mercedes-Benz PR officials also rented a small French racing circuit just across the border to ensure that journalists would have the chance to drive as well as see the new car.
It was immediately clear that despite its weight, the 230SL was a great step forward from the 190SL in straight-line performance. Although the factory’s 124 mph (200 km/h) claimed top speed was slightly optimistic, the 230SL was capable of 120 mph (193 km/h) or better. The 0-60 mph (0-97 km/h) sprint took around 10 seconds with manual transmission, a bit longer with automatic. (Uhlenhaut insisted the automatic SL was just as quick as the manual, but that claim was not supported by independent testers.) Either way, the 230SL was about 3 seconds quicker to 60 mph (97 km/h) than the 190SL and might have been quicker still with different gearing; both available transmissions suffered awkwardly spaced ratios that exposed the engine’s shortage of low-end torque.
Nonetheless, the 230SL was no match for the now-discontinued 300SL or for the Sting Ray and E-type, either of which could simply walk away from the Mercedes. In fact, the 230SL was only slightly quicker than the big 300SEL sedan, which had a commodious back seat and air suspension. There was certainly nothing wrong with the 230SL’s performance, which was brisk by European standards, but it suggested that Mercedes was deliberately opting not to vie with high-end sports car rivals.
If the 230SL was not among the first rank in acceleration or top speed, its roadholding was another matter. The W113’s suspension was softer and less sophisticated than an E-type’s, and its ultimate limits were not extraordinary, but it could be driven hard over unfamiliar real-world roads with considerable assurance. With relatively low spring rates, lots of sprung weight, and a rigid body structure, the roadster also provided a remarkably civilized ride, particularly given its short wheelbase. In contrast, many small sports cars of the time handled little better and suffered a punishing ride on anything but a completely smooth road.
Another point in the 230SL’s dynamic favor was the optional power steering. Its availability raised eyebrows with critics — power steering was rarely offered on sports cars in those days, and even Corvette buyers seldom ordered it — but the 230SL’s power steering put the manual setup to shame, offering a faster ratio and less effort with no noticeable sacrifice in precision or feel.
THE MERCEDES AUTOMATIC TRANSMISSION
If the Mercedes 230SL fell short of its mission as a relaxed and unflappable cruiser, the culprit was the optional automatic transmission, announced at launch but not actually catalogued until later in the year.
Like most elements of the W113, the transmission was shared with the W111 and W112, where it had become available in late 1961. Designed by Hans-Joachim Förster, it had four forward speeds and used a two-element fluid coupling rather than a torque converter. The Mercedes automatic was frequently compared to GM’s early single-coupling Hydra-Matic, but the similarities were superficial. The Mercedes unit weighed about half as much as the older GM automatic, used only two planetary gearsets to Hydra-Matic’s three, and drove its fluid coupling impeller directly rather than routing power through the front planetary gearset.
The design nonetheless flew in the face of contemporary American thinking on automatic transmissions, which increasingly favored three speeds supplemented by a torque converter for added flexibility. Mercedes rejected that approach, presumably in the interests of maximum efficiency; the engineers claimed the lightweight automatic consumed only 8 hp (6 kW). Shift quality was firm but not lightning-fast, probably for the sake of band/clutch longevity. The staggered shift gate (which would be much imitated in the years to come) allowed full manual control of second and third gears, either of which could be held all the way to redline.
If you treated the automatic as a sort of semi-manual gearbox that could shift for itself when required, it worked reasonably well. Where it performed less adroitly was in gentler driving, where it tended to shift with the same sort of jerk GM had spent millions of dollars engineering out of the Hydra-Matic. This was compounded by the widely spaced ratios, which made for big changes in engine speed with each shift.
An added annoyance was the fact that the automatic would only start in first if you floored the throttle or moved the shifter to ‘2’ — which was tempting, since few Mercedes engines of the time had enough torque to make a second-gear start anything other than lethargic. Even with a full-throttle, first-gear launch, the transmission would automatically upshift to second at 28 mph (45 km/h), dropping engine speed about 2,000 rpm in the process. (Admittedly, the four-speed manual transmission was not a great improvement, with an even shorter first gear (4.42:1), widely spaced ratios, long throws, and less-than-satisfying shift action.)
Mercedes would eventually adopt a three-speed torque converter automatic with a Simpson gearset like the Chrysler TorqueFlite or GM’s new Turbo Hydra-Matic, but not until the early 1970s, and not on the W113. In the meantime, the automatic represented a minor but sometimes irritating lapse in the 230SL’s otherwise polished manners.
A NEW (THREE-POINTED) STAR
Initial reviews of the Mercedes 230SL were generally positive. It probably helped that contemporary critics were not overly attached to either of the W113’s predecessors. The press had always been at best lukewarm about the 190SL, and there was a general acknowledgment that whatever its pedigree, the 300SL was past its sell-by date.
Except for the Pagoda hardtop, which provoked a few strong reactions both pro and con, many early reviewers politely sidestepped judgment of the 230SL’s aesthetic virtues while praising its well-trimmed interior and high standard of fit and finish. Similarly evaded was the matter of the SL’s middling acceleration, usually deflected with praise of the car’s roadholding and ability to maintain high average speeds without fuss.
Despite its softer ride, the Mercedes W113 actually handled better than did the old 300SL, thanks in large part to a much wider track (4 inches/100 mm wider in front, 2 inches/50 mm wider in back) and (comparatively) fat radial tires. The main shortcomings of the W113’s suspension were the need for frequent chassis lubrication (reduced somewhat on the 280SL) and excessive nosedive under braking, exacerbated by an overzealous vacuum booster. (author photo)
Mostly, the press recognized the 230SL as a well-engineered, quality product: It felt like money, which was appropriate because it was not at all cheap. In West Germany, list price was 20,600 DM (equivalent to $5,150) in roadster form, around 15% more than a 190SL. A fully loaded 230SL with both tops, automatic, power steering, and other options could top 25,000 DM (about $6,250). Admittedly, the 230SL was less expensive than a 220SE cabriolet and substantially cheaper than the old 300SL roadster, but that was hardly a bargain price. It was a similar story in the U.S., where the 230SL’s base price was around $6,500 POE, about $1,000 more than an E-type or fuel-injected Sting Ray.
Perhaps hoping to bolster the 230SL’s sporting credentials, Karl Kling, the manager of the very successful Mercedes rally team, added the roadster to the team roster. In late August, a single 230SL, driven by 1962 European Rally Champion Eugen Böhringer and co-driver Klaus Kaiser, entered the 1963 Spa-Sofia-Liège Rally and won outright. Of course, the rally car was not quite stock, using a bored-out engine with about 165 PS (121 kW), a beefed-up suspension, and 15-inch wheels, but the win was a testament to the inherent ruggedness of the W113 platform.
Mercedes trumpeted the victory in its advertising, but the 230SL was never as important to the rally team’s efforts as the 220SE and 300SE sedans, which garnered the lion’s share of the team’s eye-opening array of triumphs. The 230SL’s rally career was also short-lived. Several 230SLs ran in the 1964 Spa-Sofia-Liège Rally, but this time the victory went to an Austin-Healey 3000 driven by Rauno Aaltonen and Tony Ambrose. Mercedes withdrew from rally competition at the end of the 1964 season, but the 230SL had one final factory-backed outing at Greece’s 1965 Acropolis Rally, although a navigational error consigned driver Dieter Glemser and co-driver Martin Braungart to a fifth-place finish.
All this made little difference to buyers, who responded strongly to the 230SL from the start. The 230SL had its limitations, but, as we noted at the outset, there were almost no direct alternatives. There were precious few luxury cars in that era that could match the SL’s road manners, and even those that approached them, like the Jaguar Mark X and Mercedes 300SE, were far less sporting in flavor. There were cheaper sports cars that could outperform the 230SL, but none was as comfortable or as easy to drive. (Just clambering into the low-slung E-type, with its wide sills, was a test of coordination for middle-aged owners.) As a result, average annual 230SL sales were at least 50% better than the 190SL had ever done. Demand exceeded Daimler-Benz’s comparatively limited capacity, leading to lengthy waiting lists.
In Germany, the Mercedes W113 was offered as a single model with the Pagoda hardtop as an option, initially priced at 1,100 DM (about $275). If you ordered the hardtop and didn’t plan to remove it often (or didn’t have anywhere to store it), you could delete the soft top mechanism for a credit (initially 750 DM, around $187). In some markets, the three variations (soft top, hardtop, both tops) were catalogued as separate models, typically dubbed roadster, coupe, and coupe-convertible, respectively. A 250SL with the soft top deleted could be ordered with a small rear bench seat, a combination sometimes called the California roadster or California coupe.
That popularity allowed the 230SL roadster to remain in production for four years with almost no cosmetic changes, although there were various mechanical refinements, including wider wheels (from 5.5J to 6J), a relocated spare tire, and a larger fuel tank (expanded from 17.2 gallons/65 liters to 21.7 gallons/82 liters).
One of the most intriguing was the addition in March 1966 of an optional five-speed gearbox. Made by ZF, it had a taller first gear than the Mercedes four-speed and an overdrive fifth gear for more relaxed cruising. (There was also a close-ratio gearset with a direct top gear, but we don’t know if any were installed in W113s.) While the five-speed sounded ideal on paper, its shift quality was nothing special, and it was both hard to obtain and quite expensive: 1,200 DM (around $300) in Germany, a hefty $464 in the U.S. It was generally available only by special order, and is very rare today.
MERCEDES 250SL
In August 1965, about two years after the introduction of the 230SL, the W111 and W112 sedans were replaced by the new W108 and W109 series. The W111 and W112 two-door models, which the new sedans strongly resembled, continued with few changes, but all now had larger four-wheel vented disc brakes and a new 2.5-liter (152 cu. in.) six.
Called M129 in fuel-injected form, the new engine had a substantially new block with seven main bearings for greater smoothness. It combined the 82mm (3.23-inch) bore of the 230SL’s 2,306 cc six with a longer 78.8mm (3.10-inch) stroke for a total displacement of 2,496 cc (152 cu. in.). The cylinder head was similar to the 230SL’s, albeit with bigger valves, and Bosch multiport injection was again specified. Peak output was the same 150 PS DIN (110 kW, 170 bhp SAE) as the 230SL’s 2,306 cc engine, albeit with 159 lb-ft of torque (216 N-m, 174 lb-ft SAE). An oil-to-water oil cooler and a viscous-coupling fan (which became optional on the 230SL in 1966) were also included.
The two-door Mercedes W111 models received the M129 engine and four-wheel discs brakes for the 1966 model year; this is a European 250SE cabriolet. We have to admit we’re not sure why the W113 lagged so far behind, particularly since we assume production rationalization was a major motivation for the introduction of the 250SL. (author photo)
It was more than a year before Mercedes began installing the new engines and brakes in the SL. The revised car, now called 250SL, debuted in Europe in late February 1967. Except for its engine and brakes, the 250SL was very similar to the late 230SL — a point Mercedes advertising took pains to mention — but the four-speed manual transmission had slightly taller ratios and the axle ratio was lowered from 3.75 to 3.92:1. A taller 3.69 axle was now available by special order.
In Germany, the bigger engine and four-wheel disc brakes brought the 250SL’s base price to 21,600 DM (equivalent to about $5,400). U.S. cars cost about $300 more than the last 230SL. A fully loaded U.S. 250SL now topped $8,000 — not quite in the realm of exotics like Aston Martin, but in the same range as the new FWD Cadillac Eldorado.
The Mercedes 250SL was not the fastest sports car on the road in 1967, but did have top-flight brakes: vented discs all around with diameters of 10.7 inches (270 mm) in front and 11.0 inches (280 mm) in back; 430 sq. in. (2,775 sq. cm) of swept area; a pressure-limiting valve to reduce the chances of locking the rear brakes; and integral drum-type parking brakes. (Photo: “Mercedes-Benz 250 SL W 113 2012-07-15 14-55-26” © 2012 Berthold Werner; resized 2013 by Aaron Severson and used under a Creative Commons Attribution-ShareAlike 3.0 Unported license)
The 230SL remained in production through early 1967 and was still on sale for some time after that, particularly in the U.S., where the 250SL wasn’t introduced until April and wasn’t available in substantial numbers until summer. Midway through the year, the 250SL received an interior makeover and other minor changes to comply with the new U.S. federal safety standards scheduled to take effect January 1.
Otherwise, the 250SL was not easily distinguished from its predecessor unless you were close enough to read the emblems. The extra 190 cc (12 cu. in.) still didn’t make the six particularly muscular at low speeds, but its additional torque meant that a 250SL with automatic was now about as quick as a 230SL with manual transmission, a modest but useful improvement. The M129 was also smoother than was the M127, although some owners complained that the bigger six was prone to fouling its spark plugs in slow traffic.
Nonetheless, the 250SL was well received by critics. Like the 230SL, it didn’t really excel in any single area, but it offered a unique blend of virtues wrapped in an aura of quiet but unmistakable affluence that provoked period automotive writers — and no doubt many buyers and observers — to fits of covetous eloquence.
Aaron,
Thanks for the mention as a source. Glad I could assist! As a former owner of a 190 Heckflosse, I really enjoyed the write-up!
Keep up the excellent work!
EB
Thanks for another good read.
In the second paragraph of the 250SL section is this line:
“…combined the 82mm (3.23-inch) stroke of the 230SL’s six with a longer 78.8mm (3.10-inch) stroke for a total displacement of 2,496 cc (152 cu. in.).”
You’ve got two numbers listed as the stroke. Am I correct in guessing the 82mm number is the bore of the engine?
Oops! Yes, that should have been "the 82mm (3.23-inch) bore."
Thanks for the correction — I’ve amended the text.
Thorough and judicious as always, you capture perfectly the virtues and drawbacks of the113 with a fine sense of historical context and the competition.
I may be wrong, but I believe the dark blue 230SL you show has replacement US headlight frames from a 280, with the much larger orange turn signal lens, instead of the narrow orange strip at the bottom–a not uncommon thing to see, as are replacement tail lights with orange blinkers, instead of the full red lenses. Not nit-picking here–in the eternal restorers’ debate between aesthetics and accuracy, I am an aesthete. I have a 67 pseudo-250SL: originally a 230 (as VIN indicates), but a replacement block from previous owner from a 250SE. He also put on a 250SL trunk badge, which is one of many hard-to-find-so-absurdly-expensive parts for 113’s. I’ve further mucked it up for purists by installing European headlight units on an originally US car, but they express the clean original Bracq design much better to my mind.
Market values for the 250SL (real, not pseudo, like mine) remain oddly less than one would think, given their short production run, and what at least some consider its combination of superior power to the 230 and lighter weight and less squishy suspension than the 280. The latter seems to have been modified with more of an eye to the US market, particularly in the increasing prevalence of automatics over manual gearboxes. With the 107 firmly aimed at American tastes, the manual all but vanished from US-spec models, and Mercedes certainly cashed in. There’s an interesting (arguably saddening) history to be written about the evolution of Mercedes in response to the US market, from the 70s to the present, as they’ve recalculated designs according to customer tastes, changing technology, and (perhaps most dramatically) the rise of Lexus in the 90s.
For all the drawbacks of the long throws, the engine noise from gearing ratios at high speed, and so forth–about which you’re spot-on–the manual in my 113 is so much fun to drive, and the exhaust note is wonderful (the sound of engineering, not “sound engineers” fiddling with the muffler or (laughably, to me) electronic tweaks to the stereo system, as in some new sports cars. The 113 as a whole can be a money-pit far beyond the 107, given the complexity of welded body work around the grill, hood, and fenders (none bolted on, as with 107’s); aluminum hood, trunk lid, doors, and use of metal instead of plastic in so many places (grill screen, dash, door sills, and on and on); complex, difficult-to-fit wood and other trim on the dash; literally countless grease fittings (and I do mean “literally” in the dictionary sense; even experts on SL forums seem to discover new ones); and antiquated rust-proofing on a body that seems designed to collect water rather than drain it in certain spots. The mechanical fuel injection needs adjustment beyond the skills of all but specialists, including MB dealers, and front-end body work is both expensive and, unless very well executed, will give itself away immediately in panel gaps around that lovely hood design. A 107 is a much saner and cheaper collector choice. But with a fairly well-sorted solid “driver,” by no means fully restored, I’ve never regretted getting my 113, which feels like a machine, not an appliance, yet is utterly easy to drive and rides better than many cars designed decades later. And the shape and proportions are to my eye almost perfect; next to it, the far-more-modern 107 sacrifices character for sleekness. In Dallas, I get happy waves from people who would never notice a 107 (which are all over the place) or a late-model SL that cost 5x what mine did to buy.
Apologies for self-indulgent response. And thanks again for your work!
Thanks, Rick! I think you’re right about the lights and have amended the caption. I was so distracted dealing with the paste-up (typing HTML img tags by hand — don’t ask) that I missed that.
I would be very curious to know how many European W113s of each generation had automatic. Based on the comparative U.S. and rest-of-world production figures and the various estimates I’ve seen for the percentage of cars with automatic, it can’t have only been American buyers opting for it. Still, the fact that the U.S. R107 was offered only with automatic is noteworthy.
For better or worse, the R107 is still too ubiquitous to really stand out. In this part of Los Angeles, it’s hard to go outside without seeing one and the longevity of the design means that it registers as an older car rather than an Old Car. (The C107 is considerably less common.)
On the fuel injection system, I suppose it’s only fair to point out that most mechanical injection systems of that time were similarly intolerant of shade-tree mechanics. On Triumph’s 2.5 PI engine, for instance, the factory service manual included stern warnings not to even try adjusting or resetting the metering unit’s diaphragm springs, which required very fine tolerances.
One would assume that since the R107 was offered only with automatic in the U.S., a majority of American W113 owners had ordered automatic. But there’s another wrinkle, which Road & Track mentioned in their first road test of the R107. Smog-certifying both manual and automatic R107’s for the U.S. would have been that much more expensive. Road & Track gave a figure (which I no longer remember) for the percentage of W113’s sold with automatic in the U.S. and said Daimler-Benz “had to go with the majority.”
As an indication of one demographic to which the W113 appealed, I read an article around the time the BMW 325 Cabriolet was introduced. The writer said that until the advent of the 325 Cabriolet, the W113 was really the only game in town for doctors’ wives.
To my eye the boxy look is fine on M-B sedans but not on the R107. The W113 is so much more graceful.
The separate emissions certification cost is a major reason why we no longer get a lot of cars with manual transmission even if they’re sold that way in other markets — or if we do, the manual actually costs more than the automatic. (That said, it’s interesting to note that some Japanese cars have been sold here with manual transmissions that weren’t offered in the home market.) In any case, the newer three-speed torque converter transmission was probably better suited to U.S. driving styles anyway, particularly with a V-8 engine.
I have a feeling the R107 was intended to look more “masculine.” I don’t recall offhand if the designers specifically said that (although they were certainly conscious that it was bigger and bulkier than the W113), but it would make sense from the results. I don’t know how the demographics of the R107 differed from those of the W113 overall, but at least in L.A., it seems the R107 had a fairly unisex appeal.
I think perceptions of the R107 end up being skewed somewhat by the model’s longevity. Its lifespan overlapped several generations of Mercedes-Benz sedans, each of which became progressively sleeker while the roadster stayed (at least visually) the same. Not that the R107 isn’t objectively bulkier and less graceful than the W113, but having viewed it against cars of the late ’80s as well as the ’70s, the effect is exaggerated.
This was a fascinating and quite comprehensive read. At the last Frankfurt Auto Show and more recently the Essen Techno Classica there were many W113s present. It is apparent their values are increasing, unlike the not particularly liked successor. A number of specialists here in Germany have taken to building cars that retain the 230-280SL styling but with modern Mercedes-Benz engines and transmissions. At an impressive price, of course!
i learned to drive on my dad’s ’71 280sl. only negatives were the automatic, 4.08 rear end (felt bad taking her on the interstate), heavy a/c that was never used, and the points constantly frying up.
love to do a resto-mod if the car would still sound the same. amazing exhaust note. perfect car for going to the golf course.
Thank you for this article. As I’ve only just discovered this website, I am very impressed with the thoroughness and depth that make this article a true gem. Very inspiring.
My grandmother bought a ’72 250C brand new, she kept it for 25 years, and I drove it a lot over that time. The comments on the Mercedes automatic transmission are so spot-on. If you drove it with just the right amount of partial throttle it shifted almost smoothly, anything else and it was jerking and slamming gears. Boy, I miss that car.
Did any of the early 1964 230 SL come with dual carbs instead of Bosch fuel injection.
Not from the factory, but since there were dual-carburetor versions of the 220 and 250 engines in the sedans, it seems entirely possible that someone might have done a swap for some reason.
Owner of 1970 250C …..motor swapped on 76′ to a 280 S/A…..yes yes stamped on the head is S/A ….definitely not a 8 ….block says 130 923 12 001641 …does anyone know this model ? Closest I could come was a 280S a/c or the 280 S/8
The Triumph Stag was aimed at more or less the same market segment as the W113 and R107. Unfortunately for the Stag, it was developed and introduced at a time of extreme dysfunction at British Leyland. It quickly acquired (and never really lost) a reputation for poor build quality and engine problems.
Thanks for that fine article. I recently purchased a 1963 MB 230SL
standard transmission. The car seems to be treated poorly at sometime during it’s past. The steering wheel cracked, the top of the ashtray is missing, but it is a fun car to drive, I am slowly bringing it back to good form. BTW, any idea where I can get a top to the ashtray? I fear that I might have to get it fabricated.
I’m afraid I can’t help with parts or repairs, sorry!
The Stag was a later car than the W113, and was an occasional 4 seater. Its main competitor , on the UK market a least , was the Reliant Scimitar GTC
All true. On the other hand, the W113 was one of the key inspirations for the Stag (whose development was quite protracted) and Triumph explicitly hoped to make the Stag that kind of car. Obviously, the Mercedes was a lot more expensive (and the R107 350SL was even more so), and Triumph didn’t have that kind of cachet, but that’s the sort of buyer they wanted to court.
Of course, it didn’t turn out that way, but massive disparities between intention and outcome are a central theme of the Stag story…!
Rather intrigued by the 115-134 hp 2.2-litre output of the M127 engine in the Mercedes-Benz W127 prototype, both it not being produced with neither it or the Mercedes-Benz W121 featuring a coupe bodystyle were missed opportunities.
Quite perplexed by the outputs of the M180 6-cylinder engine as find it difficult seeing capable of putting out 110 hp whereas the M127 was comfortably able to exceed those figures, also interested to know the M180 family’s relation to the larger M186 family.
The issue was pretty plainly torque, and in particular the shape of the torque curve. Torque output with normally aspirated engines is directly related to displacement. Many of the changes that will provide more horsepower from a given displacement either won’t produce a commensurate increase in torque. Some will actually hurt torque output or at least alter the torque curve in undesirable ways. The M127’s torque peak was 4,500 rpm, which was quite high for the time and forced some awkward compromises in gearing to compensate. In the heavier sedans and coupes, that would have been even worse.
It’s easy to lose sight of that issue when one becomes accustomed to modern engines, which are vastly more flexible. With modern engines, especially with forced induction, the torque peak doesn’t tell you much about the shape of the curve, whereas with early sixties technology, a torque peak above 4,000 rpm was likely to mean a very peaky, high-strung engine.
Can understand as way variations on the M180 were still used to some extent instead of the M127, just brings up the question of whether even further could have been extracted from the M180.
Like the idea of the Mercedes 190SL W121 receiving an 110 hp M180 or 115-134 hp M127 in W127 prototype form as a German equivalent to the Big Healey, which also switched from 4-cylinder to 6-cylinder engines.
Despite the potential compromises in refinement and tractability, is it known how much more power was capable of being extracted from the 105 hp 1.9-2.0 M121 4-cylinder engine or what it would have likely entailed?
I don’t have any details regarding what Nallinger tried in getting more power from the M121, but in that era, the steps involved in getting more power out of a given engine of given duration were generally straightforward: bigger ports, bigger valves, greater valve lift, longer duration, more carburetion. Not all of those are necessarily practical; for example, existing cylinder head design may impose limits on port size and shape, or oil system design may be inadequate for sustained high-rpm use. This makes your question more complicated than it might appear because some modifications may have required extensive and expensive design changes, such as new cylinder heads, for very modest improvements in power and significant cost to low-speed response.
The Big Healey is potentially a misleading comparison because the switch from the big four to the six was initially a matter of production economy rather than a product planning decision; the initial 100-6 lost some ground in performance. (It’s also significant there that the 2.6-liter C-series six was actually about 21 cc smaller than the big four!)
Read the M121 later M115 4-cylinder and the M180 family up to the M110 6-cylinder were related to each other (along with the dieselized 4/5-cylinder units), given the M121 and M180 engines shared history (and other related derivatives) it is strange the 6-cylinder was not capable of reaching 3-litres+.
In theory an 1897cc/1988cc 4-cylinder M121/M115 of similar spec to the 134 hp 2196cc M123 or 168 hp 2778cc M130 have a potential output of 115-116 hp (1897cc) and 120-121 hp (1988cc) for an alternate 190SL W121. Meanwhile using the much later 182 hp 2749cc M110 Twin-Cam as a rough guide equates to a potential 1988cc 4-cylinder output of 132 hp.
Brought up the Big Healey comparison because it immediately comes to mind even if the background for it differed from the W121 and W127 prototype.
I can’t claim to be sufficiently familiar with the fine details of Mercedes engines to tell you how closely their four- and six-cylinder engines of this vintage relate to one another. However, comparing specific outputs the way you’re doing is potentially misleading. A six-cylinder engine has a number of intrinsic advantages, including in breathing, and if the engines are part of the same family, the six will also have greater displacement and thus more torque. Even for engines of identical displacement, the four will need significantly bigger valves to match the total valve area of the six, which the standard port design might not easily accommodate, and would need greater valve lift and longer duration to compensate, hurting low-end response.
My suspicion — which again I’m not familiar enough with the engine development saga to confirm — is that porting ended up being a limiting factor for the M121. Small ports have advantages; they tend to give better low-end torque because you have greater intake velocity at low rpm, and they allow the head to be physically smaller. They obviously become a bottleneck when it comes to extracting more power, but if you have a closely related small-displacement six, that may not seem like a significant problem. However, if you want to get more power out of the four, the port design may leave you boxed in, wrestling with questions like whether it’s worth the cost to completely redesign the head for a gain of a few horsepower.
As for the displacement, I don’t think that one should assume an engine not being expanded beyond a particular threshold meant that it wasn’t capable of being expanded that far. The 2.8-liter threshold was a pretty hard limit in terms of taxation in some markets, like France, and Mercedes-Benz wasn’t yet centering the U.S. market in their product planning decisions. They may well have said, “We’ve already got a 3-liter engine for the few applications that need it,” and left it at that.