RaceCal 'Tech Insight' - Sequential Gearbox Control Part 1
Welcome back to Volume 2 of our Tech Insight series! We hope you enjoyed our first article which gave an overview of how basic maths channels are used in the motorsport world; if you missed it feel free to check out the article and let us know what you think here: https://racecal.co.uk/blogs/news/racecal-tech-insight-maths-channels-beginner
Our second Tech Insight is going to revolve around sequential gearbox control and the appropriate systems you need in place to do this correctly and safely.
This article will be predominantly based on how we do this with a Syvecs ECU, however the theory can be utilised on a variety of engine controllers – we'd recommend checking they have the necessary feature set you need to control the gearbox and its subsidiary systems.
I suppose we should start this article with a brief overview of what a sequential gearbox is and why we'd want to use one.
A sequential transmission is a type of manual transmission used mostly in competition motorsport vehicles. We are, however, seeing sequential transmissions being used on the road more and more, but the original sole purpose of them was always intended to be for motorsport use.
A sequential transmission can produce faster shift times than a traditional synchromesh manual transmission by using 'dog rings' whilst also restricting the driver to selecting only the next or previous gear. This minimises the chance of engine damage occurring from a missed or wrongly selected gear.
On a sequential transmission, the shift lever or electric/air powered actuator (if using paddles) operates a ratchet mechanism that converts the forward and back motion of the shifter into the rotation of a selector barrel. This barrel (or drum depending on which term you'd like to use) has tracks machined around its circumference to help guide the selector forks, which in turn initiates the gear change when the barrel rotates.
Sequential gearboxes utilise straight cut gears, which have the added benefit of less power loss to the driven axle. As mentioned above, sequential gearboxes use dog rings, which in effect force the gears together rather than smoothly engaging as they would if using a helical cut gear. Whilst this sounds brutal, it is what in fact enables the gearbox to shift so quickly and without the use of a clutch in a correctly set up system. The only major downside for some is the noise (we quite like it!), the cost and that the gearbox will require considerably more maintenance than a traditional helical cut gearbox.
In the image below, courtesy of Xtrac, you will see the straight cut gears and the barrel that drives the selectors:
Ok, so that's the brief overview done - if you'd like to learn more about the inner workings of a sequential gearbox, there is a host of great information online for your reading pleasure.
We will now move back to how we implement a control system for this.
Firstly, when deciding to move to a sequential gearbox, you need to decide how you are going to implement it into the chassis and how you want to control it. You have a range of options, from no electronic control and a manual shifter mechanism for a very basic implementation, all the way up to a closed-loop ECU controlled shifting strategy with paddles.
Starting with a simple implementation, there is nothing stopping you buying a sequential gearbox and fitting it to your car with no other external control at all; we would, however, say this defeats the object of the transmission in the first place - you'd lose the super quick shifting speeds that are associated with sequentials as you'd need to lift the throttle when changing gear. This is because you need to reduce engine torque to unload the dogs to allow the shift. You will manually need to 'blip' the throttle on downshifts also. Furthermore, you have no safety - if you downshift too early, there is nothing to stop you selecting a gear too low, thus causing engine damage, or even going from past 1st to neutral. Lastly, the driver will never be able to replicate the consistency of an electronic control system to manage the torque reduction - this can wear the dogs prematurely, and in some cases destroy the gearbox quite quickly. It works, but is not the best implementation by any stretch of the imagination.
You can of course improve on this by introducing some external control and operating a closed-loop control strategy. Again, there are multiple ways of doing this - some will utilise an ECU and a separate GCU (Gearbox control unit); our preferred method is to do this all with one ECU. Most modern aftermarket engine management systems now have a gearbox control aspect that allows on board control without the need for a second controller.
The basis of a closed-loop system is allowing the ECU to obtain additional information from a group of critical sensors on the engine/gearbox to safely control the transmission. This allows you to execute full throttle clutch-less upshifts and clutch-less downshifts in safe manner. By utilising these sensors, the ECU will always know which gear you're in, thus being able to deny a shift if you try and downshift too early, for example.
So what do you need for an entry level closed-loop sequential gearbox control system?
- ECU with on board gearbox control
- Gear position sensor
- Load Cell
- DBW (also sometimes referred to as FBW/ETB) throttle body or at least a throttle blipper if using a cable throttle
- Manual shifter mechanism (this will usually be supplied with your gearbox, either cable or solid bar operated the latter being preferable)
Some of the above will look familiar and self explanatory (shifter and ECU!), however some may be new to you, so lets overview how these are used.
Gear position sensor
Does what it says on the tin - indicates to the ECU the position of the barrel which is then translated by the ECU into a gear position. The sensor will traditionally output a 0-5v range with the calibration engineer inputting the gear thresholds into the ECU.
The below shows how this looks in a Syvecs ECU. These are default values for this example; you will need to rotate the barrel and input suitable voltages.
The load cell is used to indicate to the ECU when the driver is requesting a gear shift. The load cell can either be 'inline' i.e. it sits inline between the shifter and the gearbox shift mechanism, or can be contained within the gear knob.
A Load Cell is essentially a strain gauge that will output either a rising or falling voltage which indicates to the ECU that the driver is requesting a shift. In most instances these will rest at 2.5v. When going up a gear it will raise towards 5v; when going down a gear it will fall towards 0v or vice versa. Again, it's the calibration engineer's job to input these thresholds into the ECU - if the threshold is too low, you may find the ECU thinks the driver is requesting a shift, when in fact it's just them resting their hand on the gear lever – a Load Cell can be very sensitive.
Below is an example of the threshold the ECU needs to see to initiate the cut. Again, choose suitable voltages depending on the sensitivity you require. These are default values for this article:
Whilst a DBW throttle isn't 100% mandatory, it makes the implementation a lot easier than a cable throttle. The main reason for using a DBW is for the 'blip' request (rev match) when downshifting. If you have a cable throttle this can be done, however you will need to fit a throttle blipper.
As an example, this is a blipper Geartronics offer for a cable throttle:
So we now have everything we need for our closed-loop control. We have our gearbox fitted, we have the Load Cell set up, the gear position set up and a DBW throttle – now it's time to do the fun part and set up the gear shift cut and blip.
In a professional motorsport ECU there are quite literally hundreds of parameters you can change - we'll cover the main overview, but if you have any more in-depth questions please drop us an email.
What we need to understand is exactly what happens when you pull the gear lever to request a shift; all ECUs will manage this process slightly differently, but the below is a basic flow diagram of the process. There are a lot more steps for retries / retry counters etc. not shown, but these don't need to be covered in this article for the moment.
Exactly that – you want to change gear! The ECU will see the Load Cell voltage change and realise you are requesting a shift. From here forward the ECU will now run through its shift strategy.
Most strategies will then check a 'Mask Time' before moving forward - when a gearshift cut request completes normally, the gear cut mask time is a time period in which the gearshift cut request cannot be invoked again. Once this period has elapsed, the gearshift cut request can be invoked as normal.
Ramp Out Stage 1:
When the gear cut becomes active, the ECU will look at the Ramp Out strategy. The Ramp Out strategy is there to start reducing engine torque, allowing the dogs to unload and a shift to complete.
This stage will traditionally involve a time setting (how long the stage should last for), a 'mode' (i.e. using ignition or fuel cuts), a severity function (i.e. how aggressive the cut is), a fuel multiplier stage and an ignition retard stage.
All of the above allows the engine calibrator to start reducing engine torque in a smooth and quick manner to get the shift underway. There are no hard and fast rules for what to use here - this is down to the gearbox, the engine and how you want the cut. You can be quite aggressive and target the quickest shift possible, or you can be a little more relaxed to improve smoothness and prepare the engine for the main cut.
Here you will see these settings plus others listed in the Syvecs ECU:
If you ever see f(***) this means 'function of'. Essentially if we use the highlighted as an example, the 'Ramp Out Time' is a function of 'Gear' and 'Gear Shift Direction'. You can adjust the parameters based on these functions. As per the below, you can run a different Ramp Out in second to third, and whether you are going up or down the gearbox:
When the above 'Ramp Out' timer has finished, you will then transition into the 'Main Cut' stage. You will notice similar parameters as before: timers, Fuel and Ignition Cut Severity, Fuel Multiplier, Ignition Retard (some ECUs will even allow a different ignition map to be looked up in the Main Cut stage), and many more.
The strategy will then stay in these Main Cut maps until either the next gear voltage is achieved or the maximum allowed cut time is reached (you set this yourself in the calibration). If the next in-gear voltage is achieved before the maximum cut time is reached, the strategy will then default to the Ramp In section shown below.
If the gear shift is not completed in this time, traditionally full power (engine torque) will be reapplied for a set chosen time. Shortly after this, the gear cut will enter the retry cut sequence of the Main Cut parameters in an effort to complete the shift:
As before, the whole point of this stage is also to reduce engine torque to allow the dogs to unload and a shift to complete. Inputting too short a time into the ECU to complete this stage will result in the dogs not unloading and retries being triggered.
Ramp In Stage 1:
We have now reached the final stage of the shift - you will notice this is pretty much the opposite of the Ramp Out stage that the strategy starts with. You will start re-introducing engine torque to finalise the shift and return back to full power. Your chosen inputs here will affect the speed at which you return to full engine torque, as well as the smoothness.
Much like before, you will see the parameters are very similar:
So there we have it - you can complete the gear shift in a very short period of time (depending on the gearbox this can be anywhere from 30-100ms), without the clutch and without closing the throttle. Keeping the throttle open is useful in a turbocharged car to remove any post shift lag that would be associated with a shift that requires the throttle to close.
The above is mainly focused on an upshift, but the theory for your downshift follows a similar process. The main difference this time, however, is your 'Blip Request' - this is the ECU blipping the DBW or controlling the cable mounted throttle blipper to rev match.
Depending on which ECU you are using, some ECUs also allow a Ramp Out, Main Cut and Ramp In for the downshift. As an example, below we can see how a Cosworth SQ6 ECU handles this.
Cosworth duplicate Ramp In and Out stages, so you have a Ramp Out 1 and a Ramp Out 2 etc. This is traditionally used so you can preload the actuator in one of the Ramp Out stages if using a paddle shift, thus removing any delay from the shifting mechanism. To add further flexibility you also have the whole strategy duplicated across two gear groups - Gear Group 1 and Gear Group 2. This allows a multitude of different settings to be used depending on the Cal Pot selection, which can also be changed virtually subject to gear position.
Below is a screenshot example of Pi CalTool with a small selection of parameters shown where the calibration engineer can input their chosen values. You will notice the similarities with the Syvecs system:
The Cosworth/Pectel ECU also allows an Ignition Map to be looked up during the Main Cut as shown below:
Moving back to the Syvecs, the below shows this process in a log extracted from a Syvecs equipped car running a sequential gearbox and going through an upshift.
If you look at 'GearCutState' you will see it moves through Idle, Ramp Out, Main Cut, Ramp In and Masked. In this instance the whole shift strategy was completed in 94ms. You have to be careful with how some people quote shift time - if you quote the time that the barrel moves it'll always look quicker than accounting for the full process, so this is from initial Ramp Out back to full engine torque reintroduction on a 4th to 5th upshift.
For a downshift, you can see the blip status change and the gear move from 4th to 3rd
This now concludes Part 1 of this Tech Insight - we hope this was informative and welcome any questions via Facebook, Instagram, email or even here directly on the blog! We'd like to hear your feedback (good or bad!) so please get in touch.
Part 2 will follow shortly and will further expand on a closed-loop paddle shift system.
Thanks for reading!