How did early Automatic gear boxes on cars work?

[Simon Robinson 4Participant @simonrobinson4]

In modern Automatic cars the Engine management system detects the need for a gear change through things like the MAF sensor which detects strain on the engine. But in early automatic transmission cars there were no computers or even MAF sensors etc. So how did the planetary transmission system know when the engine was under strain such as climbing a hill or moving in slow traffic etc?

[Simon Robinson 4Participant @simonrobinson4]

[Dave HalfordParticipant @davehalford22513]

Vacuum capsules, put your foot down hard, the inlet vacuum drops so the box drops a gear

[John Doe 2Participant @johndoe2]

By "looking at" engine revs and wheel revs. But hydraulically, and mechanically, not electronically. There was a throttle pedal position input, and some gearboxes had an engine manifold vacuum input, I think, but they were mechanical, not electronic. They moved hydraulic valves in the gearbox.

Gear selection by the driver was also hydro-mechanical – just a valve block to lock-out a lower gear or whatever.

Those types of gearbox, e.g. Borg Warner 35, only had 3 forward gears and were not as good as modern 7 or 8 speed electronically controlled DSG 'boxes, and were quite inefficient. If you had a big torquey engine they were OK from a driving point of view, but MPG figures were awful – below 20 MPG, owing to the constant slipping of the torque convertor, which replaced the clutch and gave some ratio variation, as well as the actual sun and planet gear sets.

The old boxes were very fluid level and pressure dependant, if you had a small leak, they could misbehave quite badly.

My memory is of auto boxes slamming into gear when you selected 'Drive', shocking the drivetrain, which was not comfortable at all.

[Clive FosterParticipant @clivefoster55965]

Simon

There is some fairly complex fluid logic in the valve body that actually controls where the shift points occur. Oversimplifying its fundamentally it's all done by a comparison between the regulated pressure from the hydraulic pump inside the transmission and the centrifugal governor driven by the output shafts. Some early designs drove the pump via the output shaft giving a more direct relationship between speed and pump pressure so the centrifugal governor wasn't actually needed. That arrangement allowed the car to be bump started but there were design implementation issues and the control logic becomes impractically simple for most real world conditions.

Quoting from Wikicars **LINK**

https://wikicars.org/en/Automatic_gearbox

The valve body is a hydraulic control center that receives pressurised fluid from a main pump operated by the fluid coupling/torque converter. The pressure coming from this pump is regulated and used to run a network of spring-loaded valves, check balls and servo pistons. The valves use the pump pressure and the pressure from a centrifugal governor on the output side (as well as hydraulic signals from the range selector valves and the throttle valve or modulator) to control which ratio is selected on the gearset; as the car and engine change speed, the difference between the pressures changes, causing different sets of valves to open and close. The hydraulic pressure controlled by these valves drives the various clutch and brake band actuators, thereby controlling the operation of the planetary gearset to select the optimum gear ratio for the current operating conditions.

As Dave says there are usually inputs from a manifold vacuum sensor and a direct take off from accelerator cable to give the actual throttle position.

Clive

[Clive FosterParticipant @clivefoster55965]

John

Torque converters aren't actually as inherently inefficient as commonly thought. Under steady state conditions they approach that of fluid couplings which rarely have worse than 4% losses and can get close to 2% loss if properly specified and fully filled. The extra losses of a torque converter are primarily due to less efficient flow around the stator that generates the torque multiplication when needed. Under steady state conditions at whatever speed the stator is optimised for you can get down to significantly less than10 % losses. But it's generally impractical to operate a car at optimised steady state speed and load. In acceleration mode losses can be much higher due to flow inefficiencies.

If torque converters were as inherently inefficient as raw fuel consumption comparison suggests the losses would rapidly boil the transmission fluid. Assuming you need 40 hp to run at 70 mph and that raw fuel consumption comparison suggests 25% losses you'd be pushing something like 10 hp worth of heat into the transmission fluid. Basically a 7.5 kW mechanical heater. The itty bitty transmission oil cooler radiator fitted to most cars isn't going to cope with that.

For a conventional carburettor or older, more basic, fuel injection systems the tree steady state, cruise, problem is that the torque multiplication in the converter lets the engine run at a much less efficient throttle setting. Basically there is a throttle opening corresponding to peak torque at any engine speed. This engine speed is the most efficient, lowest fuel consumption, one for that throttle opening. Sympathetic drivers can generally run their cars quite close to peak torque at most speeds getting best economy.

(Side comment. Really sympathetic drivers of a manual gearbox car with a properly set-up SU carb will know that once up to speed on motorway the throttle can be backed off to not much above idle and the car will maintain speed on much less fuel. The best speed for this effect depends on the car. For the old 2 litre Montego estate with manual box it was just over 70 mph with a couple or three mph range each side needing only tiny throttle tweaks to hold. 50 mpg being insane for those days from a car that size, well loaded too.)

Put a torque converter in the driveline and the torque multiplication lets the engine run well away from peak torque whilst still generating the power needed to drive the car. Engine efficiency plummets and fuel consumption rises the further away from peak torque you are able to run. It doesn't help that the inherent losses of a torque converter are roughly correlated with the amount of torque multiplication going on. So whilst driving the car with the engine running significantly away from its most efficient throttle position the torque converter is also running at a less efficient power transmission mode.

Loose – loose.

Modern electronically controlled, mapped, engine and transmission ECU systems seek to keep both the engine and torque converter away from the inefficient modes during normal operation. Adding more ratios to the gearbox significantly helps because the speed range demands on both is greatly reduced so much better optimisation is possible. Such controls also allow much greater use of the lock up clutch which is a great aid to economy.

Clive

Edited By Clive Foster on 05/06/2023 17:40:16

[Tim StevensParticipant @timstevens64731]

It all depends on how you interpret 'Automatic' – there were 1920s cars with infinitely variable ratios (relying on friction, and so prone to wear), and some in the 1930s where there was no actual clutch and instead you pressed a pedal and it changed gear 'automatically' (but you had to tell it you were going to need it). They were the designs (Daimler, Riley, etc) which had epicyclic boxes (or planetary if you like) and fluid flywheels, which were only a few tweaks from 'proper' automatics with torque converters. Speed sensors could rely on centrifugal 'governors', and torque (and driver need) was measured (vaguely) by sensing eg throttle opening.

The best source of information would have been in your local technical college library, but they were all re-graded (as distinct from up-graded) to universities and now concentrate on performance arts instead of performance engines. And the books went in the skip.

There, nothing political there, is there?

Cheers, Tim

 

Edited By Tim Stevens on 05/06/2023 17:30:47

[Chris Pearson 1Participant @chrispearson1]

Define "early" and "automatic".

Starting with now, automatic boxes are a thing of wonder. Mine has 9 gears which can change in 100 ms. Manual selection is an option, but can you do better?

Then you have the 1960s/70s/80s boxes with 3 gears, which were not particularly efficient so slower and used more fuel than manual ones.

So I think that automatic gear boxes have a poor reputation which is not deserved.

Back to "early": I have had the pleasure of being driven on, and occasionally having driven, a 1930s Daimler. It is difficult to compare one with a manual box. In 1st gear they can crawl along at a snail's pace without stalling. They can go pretty slowly in top gear and possibly start in top gear, but accelerate very slowly. "Automatic" is probably not the right term, but neither is "manual".

[not done it yetParticipant @notdoneityet]

An old coach, used during the early 1960s as a school bus, had an automatic gear box ‘of the earlier years’. Manual selection, but the gear box changed gear with no more driver input than easing off the throttle to change up to the higher gear. Maybe dipped the clutch pedal to change down? Can’t remember, but the gear was pre-selected and no further operation of that by the driver.

I would guess it was likely a late 1940s or early 1950s model? Likely a Bedford? A 29 seater petrol which was fairly crammed with passengers (some were transferred to another vehicle part way to school) as we normally almost filled a 41 seater. I expect the transfer was a means of reducing costs as it was not the only time we had to change school buses during the school journey – only in the mornings, when the original pick-up coach on a different school route to another town.

Our most regular coach was a 41 seater Leyland with the door about ‘mid-ships’. A large under-floor diesel which had a gear stick about 3 feet long, vertical from the floor beside the driver.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Tim Stevens on 05/06/2023 17:28:32:

…

The best source of information would have been in your local technical college library, but they were all re-graded (as distinct from up-graded) to universities and now concentrate on performance arts instead of performance engines. And the books went in the skip.

There, nothing political there, is there?

Cheers, Tim

ABBA earned more money for Sweden than Volvo. Discuss.

[Maurice TaylorParticipant @mauricetaylor82093]

Hi, What’s ABBA got to do with an interesting discussion on automatic gearboxes

Maurice

[Nick Clarke 3Participant @nickclarke3]

Manual gearboxes were extremely difficult to use until the introduction of synchromesh – mass produced cars were still being made with synchro on only some of the ratios right up until the 1960s. Until then silent gear changing without crunching was a skill and even then a moments inattention could make things noisy – and too much attention could lead to an accident.

Additions to the standard 'crash' box were at first designed just to simplify the gearchanging process and this is where things ended for many manufacturers eg:- Hondamatic (Trimatic) two speed transmissions being a torque convertor and a manual change between ratios allowing clutchless driving.

The Wilson preselector used planetary gears and brake bands to drive but the gear was selected by the driver and the change only occurred when the change pedal was depressed. These needed a fluid coupling (fluid flywheel) to enable starting off. Other preselectors used clutches, electromagnetic in the French Cotal box and racing cars with preselectors sometimes had no primary clutch at all relying on the change from N to 1 to progressively engage the brake bad on the first gear gearset.

Small Citroens had a centrifugal clutch (Trafficlutch) to enable clutchless starting off but the gearbox was normal with a standard clutch. Two stroke Saabs had a similar device but here it was to allow the car to coast with the engine idling as lubrication when idling is marginal with a two stroke.

The whole process of changing gear was then mechanised (close throttle, depress clutch, select gear, release clutch, open throttle) with the driver having to only move a small lever – This was hydraulically driven on Citroens and by vacuum on Standard and Rover cars. Not quite so sophisticated was the VW Beetle system where depressing the gear knob closed the throttle and disengaged the clutch enabling the next gear to be selected by the driver in the normal way by moving the gearstick. Taking the hand off the gearlever engaged drive again.

With all of these development the move to fully automatic transmissions was a small step, and in fact the technology to detect driver input, speed and load had been available for many years in infinitely variable, sliding cone and belt gearboxes so just needed to be combined with the noiseless gearchange system to make the automatic gearbox as we know it – although the dual clutch gearbox where one clutch changes a gear not under load while a second drives the vehicle is such a clever idea that one wonders why it was not thought of before!

Personally I enjoy driving a range of cars with my order of preference being full automatic, a decent manual box second equal with a crash (non synchro) gearbox, semi automatic and finally, way at the bottom of the pile a synchromesh gearbox with worn out non-functioning synchro as it is unpredictable at best.

[John Doe 2Participant @johndoe2]

Modern DSG gear boxes are really good, and very clever: combining two clutches and two gear boxes concentrically. Electronic control and learning is a real game-changer as well.

As a young lad, building simple gearboxes in Meccano, I assumed that gears in a gearbox were physically moved in and out of mesh, so was surprised to learn from the Ladybird book of the car that all gears are in constant mesh, and are engaged to the shaft when required with synchromesh hubs.

I found it surprisingly easy to drive and change gears with a manual synchromesh gearbox without using the clutch, (apart from starting off, obviously). Just requires being gentle; lifting the accelerator and matching road and engine speeds to the next gear – and then gently holding the gearstick 'against' the next gear, until it slips in after the synchro does its work.

 

 

Edited By John Doe 2 on 06/06/2023 12:25:15

[noel shelleyParticipant @noelshelley55608]

Dear Nick, Thank you for the long description, brings back many memories, pre selector boxes in buses of the 50s and a mini 1963 with only syncro on top 3 gears, this taught me to let the revs drop on up change and blip the throttle on down change, a habit I never got out of ! John Doe,then there was the night my cluch release bearing failed and I had 65 miles to do with a very heavy trailer on – amazing what you can do when you try – though an understanding of the box helped ! The starter just got us moving as the engine fired, and it was late so no need to stop all the way home.  Noel.

Edited By noel shelley on 06/06/2023 12:57:42

[Bruce VoelkerdingParticipant @brucevoelkerding91659]

this was in the States, out in the Country. We had a '63 Ford Galaxie, color Light Blue, which developed a strange Problem in 1968. Sometimes, not always, it would burn "Oil" and leave a blue Cloud behind it like a war-time Smoke Bomb. Us Kids thought it was fantastic – naturally we named it the "Blue Streak". It was odd that it was an intermittent Phenomena. Even more odd, there was no detectible loss of Engine Oil Level. It dawned on me one Day to check the Automatic Transmission Fluid. It was quite low, so it went to the Ford Shop.

Back in those days Automatic Transmissions were treated as some sort of mysterious Object that most Mechanics would dare not touch. There were specialized Transmission Shops that would work on them. Common knowledge was a Trip to one of them meant quite a hefty Bill.

I remember the Repair Cost was surprisingly low. They had replaced something the Guy called a "Vacuum Diaphragm" which he said was in the Vacuum Line between the Carburetor and the Automatic Transmission. Today I doubt it was a true Diaphragm but rather some sort of Oil Limiting/Collecting Surface.

In Automatic Transmission Shops they would always have a half-sectioned Valve Body. They were about 8" or 10" square with more Holes & Cross Holes & Spools than you can imagine. A great Tool for explaining the Cost of the Repair.

[Mick B1Participant @mickb1]

Posted by SillyOldDuffer on 06/06/2023 09:18:01:

Posted by Tim Stevens on 05/06/2023 17:28:32:

…

The best source of information would have been in your local technical college library, but they were all re-graded (as distinct from up-graded) to universities and now concentrate on performance arts instead of performance engines. And the books went in the skip.

There, nothing political there, is there?

Cheers, Tim

 

 

ABBA earned more money for Sweden than Volvo. Discuss.

 

But Volvo distributed it more widely.

Edited By Mick B1 on 06/06/2023 13:56:31

[not done it yetParticipant @notdoneityet]

Manual gearboxes were extremely difficult to use until the introduction of synchromesh – mass produced cars were still being made with synchro on only some of the ratios right up until the 1960s. Until then silent gear changing without crunching was a skill and even then a moments inattention could make things noisy – and too much attention could lead to an accident.

Nick,

Only a decade out (depending on whether you included the ‘60s). Think Minis, Minors, Hillmans, Fords and maybe a few others. I expect it was the Japanese imports, that came along later in the decade, that gave the industry a sharp kick up the rear. Likely ‘68 when ford introduced synchro on all gears, but may have been released on the Mkll Cortina in ‘67?

The morris Minor never did get an all-synchro box and was last made in ‘71.

My Mum only passed her test in about ‘69/70. She was taught in a Hillman 1500. She never ever started off in first gear in her Escort Ghia, unless on a steep incline, because that was the way she was taught!🙂

[Clive SteerParticipant @clivesteer55943]

None of my motorcycles had synchros or needed then in their gearboxes. A few times I had to drive them without a clutch when the cable broke. No problem on the open road but a bit tricky in London with the traffic lights.

CS

[Andy StopfordParticipant @andystopford50521]

Posted by John Doe 2 on 06/06/2023 12:20:03:

….so was surprised to learn from the Ladybird book of the car that all gears are in constant mesh, and are engaged to the shaft when required with synchromesh hubs.

I found it surprisingly easy to drive and change gears with a manual synchromesh gearbox without using the clutch, (apart from starting off, obviously). Just requires being gentle; lifting the accelerator and matching road and engine speeds to the next gear – and then gently holding the gearstick 'against' the next gear, until it slips in after the synchro does its work.

Edited By John Doe 2 on 06/06/2023 12:25:15

Actually, sliding mesh gears were quite common at one time, for example in the Bedford 'O' series PSV and lorry chassis. Of course it wasn't possible to have helical gear teeth with sliding mesh, so the gears were straight cut giving distinctive whining and howling noises.

Some commercial vehicle gearboxes had a combination of straight cut on the bottom two gears and constant mesh with helical teeth on the higher ratios – quieter, but when worn they were prone to jumping out of gear.

Straight cut gears would carry on working despite amazing amounts of damage to the teeth – they would always be well rounded over on the edges which first came into contact, but also frequently had large chips broken off, and holes eroded into the flanks of the teeth, possibly initiated by chip fragments being crunched between them.

I agree with Nick above, an elderly synchromesh box can be a swine if the synchronising bit is worn but the baulk mechanism still works. In theory you can still change gear by double-declutching, but in practise, it's extremely difficult to judge the right engine speed because there's no crunch to give you feedback if you get it wrong, the wretched thing just refuses to go into gear.

[SamsarandaParticipant @samsaranda]

A few years ago I had a Renault with an auto gearbox, during the warranty period it developed a very obscure and very infrequent fault, you would be driving along quite smoothly and all of a sudden there would be a jolt from the transmission accompanied by a loud metallic clang. Took it to the agents that I bought it from and they gave me a blank stare when I explained what had happened, they drove the car and of course the fault wouldn’t perform, their answer was well keep driving and if it still happens bring it back but they had no idea what was causing the fault. I carried on driving the car and periodically it reoccurred so I decided to fault find it myself, so I studied the ins and outs of the type of auto gearbox and eventually after much study and head scratching I worked out that the fault was in the valve chamber, it was one of the many valves that was sticking and suddenly freeing as pressure built up. I returned to the garage and explained my findings to the service manager and he openly admitted that he knew nothing about auto gearboxes but was prepared to accept my diagnosis as it would be charged to Renault as a warranty repair. This meant replacement of the valve block assembly, which I can only assume was a very expensive piece of engineering, it formed a plate bolted to the underside of the gearbox, the assembly was a maze of oil passages and valves. After the repair procedure the fault disappeared so it looked like my diagnosis had been correct. I still drive an auto but now buy Honda cars and in the words of the service manager at the Honda dealer “they are bulletproof “. Dave W

[Ady1Participant @ady1]

I recall DAF had an "automatic gearbox" system which used a huge rubber band

[John Doe 2Participant @johndoe2]

Hardly a "rubber band". More like a heavy duty V belt.

I owned a modern Audi with a CVT (constantly variable transmission) auto gearbox, which used two hydraulically variable cone pulleys and a belt, and it was really good, although it required a different driving technique: Instead of setting an engine power with the throttle, you used the throttle to set a speed and the gearbox adjusted its ratio according to the road gradient to produce that speed. This was counter-intuitive to many drivers, but as an Airbus pilot, I was open to this type of automation, and got on with it very well.

[Tim StevensParticipant @timstevens64731]

Actually, the Daf had two V belts, one to each half of the rear axle, and this meant they relied on the two belts settling into the pulleys* more, or less, on corners, and so there was no need for a differential. Very nearly a good idea.

* pulleys with flanges that could move in or out, so altering the ratios and making proper gears unnecessary.

And while I am writing, remember the effect on driving, and engine design, resulting from crash geaboxes. It meant that anyone even slightly timid or nervous was discouraged from driving or owning a car, and designers relied on very flexible (but less efficient) engines to avoid gear changes wherever possible. It even affected motorcycles, with Brough Superior advertising '8 to 80 in top gear'.

Cheers, Tim

Edited By Tim Stevens on 07/06/2023 10:31:09

[Howard LewisParticipant @howardlewis46836]

The post WW2 automatics were basically epicyclics where which brake band was operative was controlled by the manifold vacuum. hence the vacuum capsule. Road speed was sensed from the pressure produced by an oil pump within the transmission, and driven by the output shaft.

The Borg Warner 35 was relatively unsohisticated and in some applications could produce quite a rough ride!

Renauklt sophisticated this in the 70s, using electrical sensing of road speed and load by using a small generator, where the output was moderated by shading between stator and rotor, to snse throttle poisition (implying load )

A torque convertor was a development of the fluid dlywheel, wher a set of intermediate vanes produced torque multiplication.

Matching a torque convertor to a particular engine and installation could be quite a task. At Rolls – Royce, both the car and oil engine divisions employed specialists to calculate the best match. O E D used Twin Disc transmissions, made under licence at Crewe.

A torque convertor required fewer gear ratios. than a fluid flywheel, because it already provided torque multiplication

Unsynchronised gear boxes were used in commercial vehicles until the late 60s, although semi automatic transmissions were available. The Leyland "Moncontrol" was direct air control of the brake bands rather than the electromechanical system where the the tiny gear lever controlled solenoid valves which fed air to the approriate air cylinder to apply or release the required brake band.

The London Transport Routemaster used a more sophisticated electcal contol of the gear box, supercediung the fluid flywheel and epicyclic box prviously used on the RT and RTLs.

An unsynchromised box provided a lighter gear change than a synchromesh box, even if you double declutched any way!

Used cortrectly, a cluch brake (Clutch stop ) upward chnages could be made as fast as or faster than a syncromesh box!

Relaints used unsynchronished boxes into the 50s and 60s.

As Peugeot said of the unsynchronised gear box "C'est brusque; mais sa marche"!

Howard

[Clive SteerParticipant @clivesteer55943]

In the late sixties my flat mate had a French car, I believe a Renault, which had an electrically operated automatic transmission. I believe it had a magnet particle clutch and a normal type gearbox but operated sequentially, up or down like a motorcycle gearbox, by electrical actuators.

I didn't take that much notice of it as it was a car and I was only interested in motorcycles.

I only had one ride in it as a passenger and it seemed to drive like a normal car. When stationary one would put it in drive and press the accelerator pedal and off you would go. At an appropriate speed the engine would throttle back slightly, irrespective of the accelerator pedal position, and the next gear would be engaged with the same sort of timings if done manually. If you took you foot off the accelerator pedal the car would slow and change down and I believe there was a link from the brake pedal to disengage the clutch to bring the car to a complete stop.

He did demonstrate that it could get confused sometimes if the accelerator was too on/off but otherwise it was reasonably reliable but I don't think he kept it long.

CS

[Author]

Posts