Ready for More?

Auto Grade FM25V10

The automotive industry is very conservative.  New products or ideas are adopted only after their quality and integrity have been proven. Things change relatively slowly.

A short time ago Ramtron made a small change. We adopted a policy of qualifying all new products to AEC Q100 standards rather than wait for a customer to request the qualification. The first product resulting from this policy is the FM25V(N)10. The FM25V(N)10 is a 1Mbit SPI F-RAM with a few new features that might be interesting to automotive engineers.

The V family is characterised by a lower Vdd, operating from 2V to 3.6V. Almost all new automotive designs are 3V nowadays. Many FPGAs, Micros and ASICs are capable of lower operating voltages so the V family offers  a memory that supports these designs.

The serial number feature of the new FM25V(N)10 is optional. It can be customised so that only a certain range of numbers are sold to one customer. This simple feature allows the software to ensure that the product is not a clone. It also has another benefit. It is guaranteed to be unique so every system made will have a different number. This could be useful as a MAC address, IP address or a similar number that uniquely identifies a product on a network.

The FM25V(N)10 has a fast SPI at 40MHz. Using F-RAM’s ability to write as fast as data is delivered over a serial bus allows the entire 1Mbit to be written in about 27mS. This is less time than it would take to write three pages of data to an EEPROM or Flash memory.

The serial memory used in automotive applications have been slowly moving towards higher density and lower Vdd. Ramtron’s change in qualification policy may have been small but it means that automotive engineers now have a nonvolatile serial memory solution ready and waiting for them as their demands materialise rather than having to request it and wait.

Look, No Batteries!

The world of energy harvesting is getting very interesting. The idea of getting something for nothing is appealing and you get to save the environment at the same time. This is what they call a win-win situation.  There are lots of fascinating numbers on this subject but my favourite is that 253 billion batteries will not be recycled in Europe over the next 50 years. At the current run rate, that’s about 1000 tossed batteries per person.  I wonder how high a mountain 253 billion batteries would make?

It’s all very interesting, you say, but this doesn’t apply to automotive electronics. A car has only one battery and that gets recycled. On the surface this looks like a good argument but I recently attended a conference on Energy Harvesting and Wireless Sensor Networks. Now wireless sensors are something you do put in vehicles, the most common being tyre pressure sensors.

So perhaps energy harvesting is worth looking at a bit further. Everyone is aware that the wiring harness in a vehicle is expensive and heavy, which is why most manufacturers are trying to find a way to reduce its size. Perhaps wireless sensors that harvest energy from the environment are a way to make that happen. Imagine an engine bay with no wiring to any of the sensors!

Let’s start with the easy ones. A sensor that generates energy from rotational movement is relatively easy. There are magnets, coils, piezo sensors and Wiegend sensors that can all generate enough power to conduct useful electronic functions. Energy harvesters that get their energy from vibration are also possible, which can be very effective if the frequency and amplitude of the vibration is well defined. Similarly, collecting energy from an RF field is not so hard if you can tune your antenna to the right frequency. The engine bay of any vehicle is an oasis of opportunity to harvest energy from movement, heat, vibration or RF. So it would seem that harvesting the energy is not technically that difficult.

Ramtron recently released an innovative memory called WM72016. Its a wireless memory that can be read from and written to via an RF port and serial port. You can think of it as a dual-ported memory. The nice thing is that it can be powered either from a local power source or from an RF field. In a wireless sensor application the data could be retrieved even when the sensor is not generating any energy and data could be written by the harvested energy from the sensor without the need for a constant RF field.

Energy harvesting may offer solutions to automotive sensors that could reduce wiring. Ramtron’s Wireless Memory with a serial port gives a simple solution to the problem of transporting data without wires based on the energy harvested from the RF field. Now come on. Be honest. Did you start out thinking that a nonvolatile memory with low-power, high endurance and fast write could actually be a tool for the removal of sensor wiring?

The Fatal Blow

I have to confess being a fan of crime TV shows, particularly ones involving forensics that put together clues and finish by tying it all together and explaining what happened. I like the medical series ‘House’ for the same reason. I wish that I had been blessed with half his diagnostic skills when I was a design engineer. Instead I spent hours looking at logic analyser output, scope traces, and software dumps searching for clues as to what had happened to cause the failure that I was debugging.

The same problem occurs with modern electronic systems. As automotive manufacturers look for greater reliability, we are looking at how we can prevent failures in the parts-per-billion range. What chance do we realistically have of ever witnessing a failure that occurs at a 1 ppb rate?

The answer is simple. The designer has to build systems into the car that perform self test and error logging functions that leave the vital trail of clues to help identify any problem. A good example of such as system are the batteries in a hybrid or electric vehicle. To get the most out of the battery, a great deal of data needs to be collected and processed including charge cycles, charge rates, temperature while charging and discharging, etc. If the manufacturer provides a 5-year warranty on the batteries, they will need to know that during the entire life of the battery, that it has not been mistreated. If it has, they will need to implement fixes to prevent the failure from happening again. Cue the need for a data log.

One key element of the data log is that it must be nonvolatile (so the data is preserved when the battery fails) and it helps if it can be written quickly and as often as possible. It’s no surprise then that F-RAM has been chosen in some of these applications. It has virtually unlimited endurance and its fast write allows designers the option of writing data continuously until a failure occurs. The F-RAM will then contain a short record of the operation that contains the trail of clues that would even impress Sherlock Homes!

Do I have Weak Fingers?

Maybe its me but occasionally questions pop into my head that no one else seems to ask. This happened recently when we were talking about map lights found in cars. My car is five years old now and the map light switch is the classical push-to-make/push-to-break switch. The latching action is accomplished with a combination of plastic and springs inside the switch. In more recent or maybe just more expensive cars (note to myself: check the bosses car) the push-button is a very delicate push-button. The latching action has probably been accomplished by a small micro. The implication is that we all want delicate push-buttons with electronic control to replace the firmer push-buttons. This is when that question filtered into my conscious – do I have weak fingers?

Putting that complex question aside for another time I soon realized that F-RAM could be part of a rather neat solution for the weak-fingered applications. The problem with the electronic control is that connecting the vehicle battery after servicing could power up the light in the on state and that could drain the vehicle battery. In the schematic below a Ramtron F-RAM State Saver is used to remember the status. This is a simple D-type latch where the output uses F-RAM to recall the previous state. This means that if the lamp was off before the vehicle battery was removed it will be off after the vehicle battery is reconnected. It doesn’t get much simpler than that. Now, I wonder how to measure finger strength?

Making Friends with the Production Team

Most engineers are aware of design for test and design for manufacture. The goal is to design products that are easy to test and manufacture. If you achieve this, you’ll have lots of friends in the production department.

I once heard that if you consider all the systems a typical car needs, you will need about four hours for programming and configuration. Although I’ve never been able to verify this, I was reminded of this figure when I recently came across another programming problem.

The problem was programming a data table during production. The look-up table (LUT) contains data to control an engine and maps all of the engine parameters under various operating conditions. The table was about 256Kbits. That’s no not unusually large so why the problem? The problem was that the LUT had to be measured and calculated during production.  To compound the problem further, the table is arrived at iteratively and needs to be re-written many times.

Let’s just run through the calculations if you stored the LUT in an EEPROM:

 Now I don’t know about you but I was surprised to calculate that the cost of programming a 32Kbyte 256Kbit) EEPROM is 8.5c. This of course assumes that you can only program one device at time.

Compare this to using an F-RAM. The programming time is basically the bus speed.

 Now if you go the production department and tell them that you could save them 1min 41s of production time and will save €0.82 (at €30 per hour) you will make lots of people happy!

What does every security system need? A switch.

F-RAM State Saver driving a relay

Sounds obvious but every security system needs a switch. Somewhere, at the output of the security system will be a switch that either turns the alarm on or turns some component off, disabling whatever it is that the security system is protecting. In a home alarm, the security system turns on the bell-box (the thing that makes a loud noise and flashes), and in an automotive security system, the ignition system is usually disabled in some way.

Now, as every good thief will tell you, if you cut the power to the security system and silence it soon after it starts to make a noise, then everyone will ignore it. Home security systems get round this obvious attack by having a backup battery in the bell-box and mounting it out of reach. Achieving the same in an automotive security system takes a little ingenuity. Autos usually only have one battery and everyone knows how to access it. If you are not familiar with this process just google it before you set out to steal cars.

One solution is to have a switch that remembers its previous state. That way once the alarm is triggered, cycling of the power supply will have no effect. Ramtron’s State Savers offer a simple, cheap electronic switch that remembers its previous state after a power cycle, which is ideal for security systems. You can learn more about our State Savers here.

EEPROM can be a right and royal pain

EEPROM with Micro and Capacitor

The circuit shown above is the memory sub-system that you would need to store data on power fail. The key components are a microcontroller for processing, an EEPROM for nonvolatile data storage, and a capacitor to provide enough power to give the micro enough time to finish writing as the power fails.

F-RAM and Microcontroller

Above is an equivalent circuit design using F-RAM. As you can see, the capacitor is no longer required for two simple reasons:

1. F-RAM writes 50 to 100 times faster than EEPROM, allowing the data to be written as the main power supply fails without support from capacitors
2. F-RAM uses about 1/100th the power of an EEPPROM so, in effect, the power that is dying lasts much longer

So why is this important? F-RAM allows for simpler, cost effective, and more reliable designs in automotive applications. Although F-RAM compoments are more expensive than EEPROMs, when you add up the solution cost including the EEPROM, extra capacitor, assembly issues, and extra board area, F-RAM becomes a compelling high-performance design alternative.

Also, the more data you need to write, the larger the capacitor needs to be in both size and Farads. One recent application that comes to mind needed capacitance of around 0.1F. This amount of capacitance does not come cheap. Another spec that engineers need to look at is the capacitance at max temperature. Capacitors leak and the leakage increases with temperature. Low-leakage capacitors are readily available but, again,  they are more expensive.

I recommend taking the time tocalculate the amount of capacitance you actually need with your EEPROM at the maximum operating temperature and ask yourself if F-RAM actually allows a cheaper and more effective solution.