User Tools

Site Tools



DC/DC converter

This chapter is about the “main” DC/DC converter, that powers the 12V grid from the traction grid, and its peripheral circuitry. Note that there is also a couple of small DC/DC converters, a few Watts each, too, within the circuitry.

At a voltage of 13,8 Volt, a 12V lead-acid battery can be completely charged, provided that the charging time is sufficient. Therefore, the DC/DC converter is powered not only when ignition is on, but also during charging of the traction battery. An extra battery charger is then not required any more. The 12V battery sure would appreciate an “equalisation” charge at higher voltage now and then, but in a conventional car, it won't be coddled either.

Requirements on a DC/DC converter

I learnt that it is quite hard to find a DC/DC converter that is reliable, meets the technical requirements and has enough functionality by itself so that one need not spend effort in additional gadgets. My first converter failed after TUEV had already done a first inspection of the car. Since the manufacturer did not make the impression that they would shortly deliver a replacement unit, I bought and integrated a different brand. This cost me dozens of hours of integration time. Needless to say, the replacement unit of the first converter then arrived just before the next TUEV inspection date, after only two months …. In the following, you will find pictures and peripheral circuitry of both converters.

Requirements are e.g.

* Humidity protection: A device with forced air cooling and a nervous blower wont go anywhere easily - it is too noisy for the interior, and too sensitive for the engine bay.

* Output voltage: The output voltage must be 13.8V to 14V (under load as well) so that the 12V battery can be charged and is not depleted during driving. A grid voltage near 14V is also expected by the headlights, which will otherwise not appear “bright”.

* Idle primary power consumption: Especially when the converter is also operated when the car is not running (though perhaps intermittingly), the idle power consumption plays a role. Since the 12V idle current drain of an EV vehicle is usually high and the 12V battery does not charge quickly below 14V, one will normally choose as much “on” time of the converter as possible.

* On/off control input: If such an input were available, and if the “off” primary power drain were acceptable, one would not need a primary side driver or input relay. Even if the feature had it's price, it would probably pay if you consider your own effort realistically. The implementation of a primary side driver (no control input available on the DC/DC converters that I used) is described below.

* Secondary idle reverse current: This is current that the unit will draw from the 12V grid when it is switched off. Two of the devices I have been testing had considerable idle reverse current of 60mA and even 200mA. This makes necessary an output side switch to isolate the device from the grid when idle. Unfortunately, it is not so easy to integrate a power FET here - it would be difficult to drive and would not cover all possible operative constellations. A relay, if driven without special precaution, tends to “stick” since the relay contact gets overloaded if closed too early. My solution is using a mechanical relay, but switching it on only when the output voltage of the converter has already approached the grid voltage. The implementation of an output relay with delayed switch-on is described below. Anyhow, a properly designed DC/DC converter should not require an output relay at all and thus save a lot of integration effort.

* Overload behaviour: Best behaviour will be a simple U/I characteristics - so if the output current reaches its rated maximum, the output voltage will just dip a bit to prevent further increase of the current. The most stupid behaviour to implement, in turn, would be a shutdown at maximum current, without automatic recovery. Needless to say, one of the converters I integrated really did have this behaviour. To get this device to start up reliably, an external current limiter circuit is required. Otherwise, it will shut down immediately after switch on, if the battery is just a bit discharged.

DC/DC peripheral circuits

Since the DC/DC converter will usually not meet all above requirements, some peripheral circuitry will be required to make the thing work flawlessly.

Below diagram shows an overview of the peripheral circuitry, as required by the second model that I integrated.

Link to DC/DC converter peripherals overview

On the left side of the diagram, the primary side DC/DC driver is shown. It is located in the front 100V distribution box, and has already been mentioned in the chapter “distribution boxes”.

The DC/DC driver has two parallel control inputs, one of which is served by the DC/DC control unit (a microcontroller circuit that has already been described in the chapter “BMS peripherals”). This circuit takes care that the 12V battery never runs low.

The actual switching is done by an N channel MOSFET. The transistor may appear overdimensioned with it's 500V 44A rating, but upon tragic experience with a 14A type I know that it needs some reserves to take up with the inrush current of the converter.

The MOSFET is controlled via a small solid state relay, that provides the required isolation between 12V grid and traction grid.

A thermo switch next to the source of the MOSFET interrupts the current when the heat sink temperature exceeds 75°C. I have experienced that such situations may occur if water gets into the distribution box and leakage currents elevate the gate voltage into an “intermediate” range. Then the MOSFET will switch incompletely and produce a lot of heat.

In the case of the DC/DC primary side driver, it would have been difficult to make a “water-proof” control circuit that drives the MOSFET's gate with with low impedance, has a hysteresis that avoids intermediate states and still draws no quiescent current. For some of the other drivers, this might be a possible improvement.

Below a view of the DC/DC converter primary side driver.

The bottom center building block is the current limiter, that avoids overcurrent shutdown. It is required only for the second device that I integrated, and the device really should feel ashamed for it's deficiencies. You may notice that I placed the shunt resistor and the power MOSFET into the return line, so that the negative output of the DC/DC converter is below ground potential. Reason is simply because this will work with an n channel FET, which particularly has lower “on” resistance than it's p channel counterparts. The control circuit based on an old-fashioned operational amplifier allows to use a very small shunt resistor (10 mOhms), that gives a voltage drop of only 300mV at 30Amps. Despite my concerns, the circuit is reacting quick enough to prevent shutdown of the converter, and so far proved to run stable without oscillations.

On the right side of the diagram, you find the output relay already mentioned above (and also in the chapter “distribution boxes” - since it is located in the front 12V distribution box). Delayed switch-on is achieved by sensing the output voltage of the DC/DC converter (after the current limiter). The relay will only be triggered after this voltage has exceeded around 10V. Without this delay, there would be a high reverse inrush current from the 12V battery to the DC/DC converter, which would cause the relay contact to “stick”.

The output relay is controlled by a separate diode combiner located on the DC/DC primary side driver board.

Both primary and secondary side are protected by fuses, see 100V distribution box - fuses and 12V distribution box - fuses.

practical implementation

The sealed converter (the light grey box at the left edge of the first picture) mounted onto the "controller" auxiliary frame. The mounting position is close to the radiator grill to ensure good ventilation. This device has an output voltage of 13,8V and delivers up to 36Amps.

The second device has forced air cooling, so it required a protective shroud.

A short “making - of”:

The current limiter utilizes the shroud as a heat sink (with additional cooling by the converter's air flow).

On the right, the opened 100V distribution box can be seen, with primary side fuse and the terminal area of the DC/DC converter primary driver. The output relay is hidden below in the front 12V distribution box.

One device had failed at low temperature. Now, my DC/DC converters have to undergo a harsh and highly professional “deep freeze” test, to prove that they do start up at -18 degrees Celsius.

how to spare the weight of the 12V lead-acid battery

In principle, it would be possible to keep the DC/DC converter on permanently, or to add a second, smaller device that feeds the 12V grid while the car is idle.

An additional, small DC/DC converter could be optimized for prolonging the battery life by doing all that ctek rituals of switching between different charging phases at different voltages. I am imagining to use a small 15V output DC/DC converter module as a power supply, followed by a microcontroller circuit that fine tunes the output voltage.

With a permanent power supply, the battery is not drained any more while the car is idle, and it's size can be substantially reduced.

The extreme case would be to omit the 12V battery at all. Of course, without a battery, it must be ensured that at any time that the DC/DC converter can deliver sufficient current to feed all 12V consumers. Especially for the anti-lock brakes, I am however not so sure about their peak current drain. A 12V battery will also come in handy as a back-up supply when the traction circuit is interrupted (maintenance) or when the DC/DC converter is broken.

So, it is probably the best option to keep the 12V battery, and only make it smaller. Unfortunately, batteries below some 40Ah dont have proper sockets, so a mechanical adapter bracket will be required.

*** back to top of this page ***

astra_conv/conversion/dcdc_converter/dcdc_converter.txt · Last modified: 2014/06/20 13:02 by richard