Go Down

Topic: Sanity check required - 12V 20A motor (Read 3502 times) previous topic - next topic

autoduino

Hi,

I'm looking to use an Arduino to control a secondary fuel pump in the car. What I'm having bother with is if I've grasped what I need to do to run the motor.

The scenario is this: The pump remains dormant the majority of the time, the standard in-tank pump runs and provides enough pre-supply pressure during normal driving. Once boost pressure reaches 18PSI, the Arduino kicks the pump off at 25% duty cycle. 23PSI has 50% duty cycle and 26PSI will run the pump 100% / ON. As boost drops, the duty cycle drops to the previously mentioned set points. At full load, the pump will draw about 16A. The boost measurement and sketch logic is fine, I've tested it out with a small PC fan and a BD237 that I had lying about - it's the driving of the high current pump I'm not sure.

At first, I thought I'd get away with running an appropriately rated MOSFET and supplying the gate with the 5V from a digital pin on the board. From what I see now, doing so would cause quite a high resistance within the source and drain points and generate a shed load of heat (not ideal!) as higher powered FETs look for something closer to 10-15V at the gate.

Now, my next thought is use a smaller FET that's happy being fully saturated at the gate with the 5V from the digital pin and use that to then feed the gate of the higher powered FET - am I right?

Has anyone done anything similar and could recommend the drivers I'd need? I see many drivers seem 'rated' up to the job, but then read stories of them burning out very quickly when pushed.

Even better - are there any drivers that will do this very thing? I've seen H bridges that are capable of driving multi channels with reverse option (not required) and I've also read that H bridges don't like running 100% and are better suited at 95%?


Boardburner2


jremington

#2
Feb 18, 2015, 11:25 pm Last Edit: Feb 18, 2015, 11:26 pm by jremington
Quote
At full load, the pump will draw about 16A.
Really? A 200 watt pump for FUEL? That is over 1/4 HP!

autoduino

The standard in tank pump is PWM controlled by the ECU. Replacing it is expensive with most solutions costing about £700-1000. As the engine comes on boost, the low pressure fuel system is caught off guard and causes a momentary drop in pressure on the rail. As the engine heads towards the red line, the fuel system then starts to struggle again.

Boost pressure is the best indicator of load. The higher the load, the higher the fuel requirement. As the turbo comes on boost, the Bosch 044 inline pump will start running at 25% duty cycle to get it up to speed and keep the rail pressure steady. As boost increases further, pump duty cycle is increased to keep the rail immersed in plenty of fuel.

The engine has a high pressure pump that provides ~130bar fuel pressure on the injector rail. Providing plenty of pre-supply pressure means the high pressure pump doesn't work as hard and doesn't suffer flow problems at higher loads.

autoduino

Really? A 200 watt pump for FUEL? That is over 1/4 HP!
Yep. Independent testing of the Bosch 044 has seen it peak at as much as 16A. The pump will provide pre-supply pressure to a high pressure pump, and as such the low pressure side does not have a return line. Once the pump ramps up and pressure builds in the low pressure side, pump current increases dramatically.

I dare say in normal usage I won't see 16A draw, but for the sake of a few £ I'd rather over-engineer the electrical side. It's far cheaper than the pump controller failing at high load and causing the engine to run lean.

Boardburner2

#5
Feb 18, 2015, 11:48 pm Last Edit: Feb 18, 2015, 11:59 pm by Boardburner2
The standard in tank pump is PWM controlled by the ECU. Replacing it is expensive with most solutions costing about £700-1000. As the engine comes on boost, the low pressure fuel system is caught off guard and causes a momentary drop in pressure on the rail. As the engine heads towards the red line, the fuel system then starts to struggle again.

Am I correct in thinking you have a faulty pump and are trying to find a cheaper work around ?

Edit.
Is your vehicle a Bently or a Honda by any chance   ;D

autoduino

Am I correct in thinking you have a faulty pump and are trying to find a cheaper work around ?

Edit.
Is your vehicle a Bently or a Honda by any chance   ;D
No. The in tank pump works fine. It is just not capable of supporting the higher power that the engine is capable of producing with the new turbocharger. Adding a higher flow inline pump is common practice in high powered applications and essentially transforms the OEM pump into a lift pump.

And no, it's neither a Bentley or a Honda.

Boardburner2

#7
Feb 19, 2015, 03:58 pm Last Edit: Feb 19, 2015, 04:09 pm by Boardburner2
Hi,



I've also read that H bridges don't like running 100% and are better suited at 95%?


H bridges inevitably have a certain amount of dead time in their switching cycle.
Its to avoid a condition called shoot through , basically a dead short.

There are high current motor shields available , but most use h bridges.

You don't have to use them as such though.
http://www.elechouse.com/elechouse/index.php?main_page=product_info&products_id=2179

Used as a half bridge they can be full on.
Check with the supplier though.
(some of the driver chips require a short off time to generate internal voltages)

Google on mosfet driver ic will give lots of options they are fairly common


MarkT

#8
Feb 20, 2015, 02:21 am Last Edit: Feb 20, 2015, 02:33 am by MarkT
Really? A 200 watt pump for FUEL? That is over 1/4 HP!
Fuel pumps for the Saturn V rocket were 7,000,000 horse power, if I remember rightly :)

[ edit: I remembered wrong, 55,000 hp only! 7 million pounds of thrust... ]
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

noisymime

#9
Feb 20, 2015, 06:33 am Last Edit: Feb 20, 2015, 06:33 am by noisymime
It sounds like what you're after is a dedicated MOSFET driver, they're designed to do exactly what you need (ie driving a high power MOSFET from a low voltage/current IC). I can highly recommend something like a tc4424 in automotive applications.

MorganS

The chip used in the Sparkfun Monter Moto Shield is originally intended for automotive applications. It can handle the harsh environment. It doesn't even need any heatsinking until the motor currents get over 20A.

A dual H-bridge is a bit of overkill for a fuel pump but it's simple to program and reliable.

I thought the original explanation was really good. It didn't say that this is a turbocharged car with the turbo boost much higher than standard, which seems to be why there are so many "Why?" questions.
"The problem is in the code you didn't post."

Boardburner2

That sparkfun shield is quite expensive.
My unasked question though is whats the fuel consumption and performance of this beast.

MorganS

I'm sure he gets at least 20 miles per Sparkfun.
"The problem is in the code you didn't post."

autoduino

Many thanks for the replies. I actually picked this up for testing as it's quite compact and was available for quick delivery in the UK. It can do 15A without cooling, or 21A cooled. I'll be swapping the cap over to the other side and mounting a heat sink for the drivers.

Performance wise, it's an A3 2.0 TFSI Quattro that is being built for the track. Stronger S3 engine plus larger turbocharger. Expecting mid 400 horse power with similar torque. The low pressure fuel system is just about maxxed out at 400, so providing that little extra grunt on the LP side will give the high pressure side an easy job and keep injection times down while also future proofing for if things get serious with bigger turbos etc.

autoduino

That sparkfun shield is quite expensive.
My unasked question though is whats the fuel consumption and performance of this beast.
Fuel consumption isn't a huge concern, but the car will be road legal and should still be good for an easy 30mpg at steady motorway speeds.

Go Up