Toyota GR86, 86, FR-S and Subaru BRZ Forum & Owners Community - FT86CLUB - View Single Post - Stock pwm fuel pump controller capacity, max current throughput etc

matt88 · 11-25-2021, 10:56 AM

Well I had some fun today doing fuel pump controller tests.

Disclaimer: I'm using the canbus translator from AGT engineering. I'm not sure of what effect this translator has on the waveform that drives the fuel control module outside of what a stock Subaru ecu would do, however I would like to assume that they've done their best at replicating stock Subaru behaviour.

So first, I scoped the output of the e38 x1 pin 50 "Fuel Pump Relay Control Primary". Just as per the work I've been doing on solving my radiator fans, the output is a nice 10 volts usually high, pulls to low style pwm square wave running at 128hz. Various fuel pump percentage settings, all the way from 1% to 100% produce the 128hz waveform with lowest duty cycle (1%) spending most amount of its time high at 10 volts and the least amount of its time at 0v. Enter 73% and 73% of the time the waveform is 0v and 27% of the waveform is 10v. All as expected.

But then scope the input to the fuel pump controller and things change. The biggest difference is that the controller is basically being driven with pulse frequency modulation at various frequencies to represent the percentage of fuel pump as controlled by the e38 (not by a varying duty cycle 128hz pwm as what's coming out of the e38). So for example the e38 punches out 128hz and 46% duty cycle, well the fuel pump controller, from the canbus translator, is now receiving a sync pulse at 48hz. Dial the e38 up to 85% and now the fuel pump controller is receiving a sync pulse of 85hz.

Another interesting find is that anywhere from 1% to 35% duty cycle as pumped out of the e38 does not produce any sync pulse from the canbus translator to the fuel pump controller. The e38 is happily pumping out its 128hz waveform at various duty cycles between 1 and 35% but lo and behold any of these numbers don't produce a result at the fuel pump controller input. It is only when you pump in 36% does the translator suddenly kick into life and what do you know, it's a 36hz pulse.

So does anyone know if A: the stock Subaru ecu also drives the fuel pump controller via a variable frequency sync pulse? And B: does the stock ecu have this "only starts driving the fuel pump" at 36hz or more issue?

Then I moved on to the output of the fuel pump controller. Basically all that is happening is that the controller is dumping out a DC level. At the lowest working frequency of input of 36% the voltage output was something like 8 volts. Push the controller to 100% and it is dumping out a good 12 volts. No matter where I set my scope for time scale, I cannot see any pulse width. It is honestly just pure DC driving the pump at a varying voltage.

So again, some of these behaviours may be from the canbus translator, however they could well be in line with the standard Subaru behaviour.

Does anyone else have the same experience as this?

For me, I want to drive my camaro pump with true pwm, something like 20khz. Also I want to be able to have the range from 1% to 100% as needed. Why? Because I want to. Maybe it's pointless or overkill but having this level of control will make me think I've done it "properly". So I am going to splice into the output of the e38 pin 50 directly and have that running the fuel pump controller instead of passing through the canbus translator. I will also find a controller that can take a 128hz in, variable duty cycle pwm and then dump out a 19/20khz pwm to the pump.

Much of the work I'm doing on solving the pwm radiator fans will end up applying here, I'm sure.

11-25-2021, 10:56 AM	#21
matt88 Member Join Date: Oct 2019 Drives: 2017 Toyota 86 GTS Manual Location: Australia Posts: 33 Thanks: 4 Thanked 24 Times in 9 Posts Mentioned: 0 Post(s) Tagged: 0 Thread(s)	Well I had some fun today doing fuel pump controller tests. Disclaimer: I'm using the canbus translator from AGT engineering. I'm not sure of what effect this translator has on the waveform that drives the fuel control module outside of what a stock Subaru ecu would do, however I would like to assume that they've done their best at replicating stock Subaru behaviour. So first, I scoped the output of the e38 x1 pin 50 "Fuel Pump Relay Control Primary". Just as per the work I've been doing on solving my radiator fans, the output is a nice 10 volts usually high, pulls to low style pwm square wave running at 128hz. Various fuel pump percentage settings, all the way from 1% to 100% produce the 128hz waveform with lowest duty cycle (1%) spending most amount of its time high at 10 volts and the least amount of its time at 0v. Enter 73% and 73% of the time the waveform is 0v and 27% of the waveform is 10v. All as expected. But then scope the input to the fuel pump controller and things change. The biggest difference is that the controller is basically being driven with pulse frequency modulation at various frequencies to represent the percentage of fuel pump as controlled by the e38 (not by a varying duty cycle 128hz pwm as what's coming out of the e38). So for example the e38 punches out 128hz and 46% duty cycle, well the fuel pump controller, from the canbus translator, is now receiving a sync pulse at 48hz. Dial the e38 up to 85% and now the fuel pump controller is receiving a sync pulse of 85hz. Another interesting find is that anywhere from 1% to 35% duty cycle as pumped out of the e38 does not produce any sync pulse from the canbus translator to the fuel pump controller. The e38 is happily pumping out its 128hz waveform at various duty cycles between 1 and 35% but lo and behold any of these numbers don't produce a result at the fuel pump controller input. It is only when you pump in 36% does the translator suddenly kick into life and what do you know, it's a 36hz pulse. So does anyone know if A: the stock Subaru ecu also drives the fuel pump controller via a variable frequency sync pulse? And B: does the stock ecu have this "only starts driving the fuel pump" at 36hz or more issue? Then I moved on to the output of the fuel pump controller. Basically all that is happening is that the controller is dumping out a DC level. At the lowest working frequency of input of 36% the voltage output was something like 8 volts. Push the controller to 100% and it is dumping out a good 12 volts. No matter where I set my scope for time scale, I cannot see any pulse width. It is honestly just pure DC driving the pump at a varying voltage. So again, some of these behaviours may be from the canbus translator, however they could well be in line with the standard Subaru behaviour. Does anyone else have the same experience as this? For me, I want to drive my camaro pump with true pwm, something like 20khz. Also I want to be able to have the range from 1% to 100% as needed. Why? Because I want to. Maybe it's pointless or overkill but having this level of control will make me think I've done it "properly". So I am going to splice into the output of the e38 pin 50 directly and have that running the fuel pump controller instead of passing through the canbus translator. I will also find a controller that can take a 128hz in, variable duty cycle pwm and then dump out a 19/20khz pwm to the pump. Much of the work I'm doing on solving the pwm radiator fans will end up applying here, I'm sure.