Anyone tried PWM fuel pump control to improve reliability?

[Grant]

The idea being pumps fail from heat when very low on fuel, but PWM control should reduce heat build up. There are several RPM-based fuel pump controllers on the market. The big question in my mind is how much will we really be able to drop fuel pump current while maintaining operation of the factory venturi. As long as target fuel pressure is maintained I'd think the venturi would keep working fine.

 

The alternative is re-engineering the factory system with an external surge tank and Radium venturi jet pump or electronic lift pumps.

We've had two OEM fuel pumps fail within a year. One just dropped pressure and let us continue to win the race, the other failed catastrophically at Sebring after a session where I essentially ran it out of fuel (1% on the gauge; basically just the fuel in the basket was left).

Edited December 24, 2021 by Grant

[Ian]

We went to dual low pressure lift pumps feeding a surge.

 

Usually less is more, but in this case, high pressure factory pumps really don't like to be run out of fuel.  Where as the lp pumps seem to handle it much better, partially because pumping at lower pressure generates less heat, and moving more volume carries that heat away.

 

We run a Bosche Motorsports high pressure pump.  It's twice the size of a factory pump, and seems very resistant to failure even when occasionally run dry.

 

PWM could help pump life when run empty if you could duty cycle it low enough, which in theory would be relatively easy to set up.  The problem is, how much would it help?  If you get 2 run dry moments before pump failure instead of 1, it's not really worth it IMO.

[petawawarace]

Are you running the stock pump in the tank?  
I made a surge tank that uses the stock pump/regulator/sending unit.  I can take a bone stock OE pump,  remove the float arm and drop it in.  There’s also a float switch on the surge tank as well.  If we run the OE pump dry, it would literally be for 20 seconds.   The pickup is right at the bottom of the surge so it will run until there’s pretty much zero fuel in it.  We use a more dry run tolerant lift pump.  

Edited December 24, 2021 by petawawarace

[Grant]

1 minute ago, petawawarace said:

Are you running the stock pump in the tank? 

Yes, everything is stock except the tank vent and fill tube. Trouble is I don't really have time to re-engineer the factory fuel system while I'm building a WRL car. Meanwhile a PWM signal can be generated by my Ecumaster dash to drive the pump via a solid state relay. Very cheap and easy to set up.

 

In our case we see 76% injector duty cycle at redline, which equates to about 51 L/h or 64% of what the stock fuel pump flows. So we should be able to drop duty cycle some at WOT, more at lower RPM, and can really cut it to the minimum that maintains fuel pressure (so the venturi jet keeps working) on decel. Overall I think I'll be able to greatly reduce heat buildup.

[Grant]

If I did a surge tank I think I'd give Radium's venturi jet pump a try instead of lift pumps. I've seen too many of those fail.

[Team Infiniti]

5 minutes ago, petawawarace said:

Are you running the stock pump in the tank?  
I made a surge tank that uses the stock pump/regulator/sending unit.  I can take a bone stock OE pump,  remove the float arm and drop it in.  There’s also a float switch on the surge tank as well.  If we run the OE pump dry, it would literally be for 20 seconds.   The pickup is right at the bottom of the surge so it will run until there’s pretty much zero fuel in it.  We use a more dry run tolerant lift pump.  

We literally do the same thing except, for years had a stock Infiniti pump as the lift pump (used OEM works best for us). To date, no surge tank pumps have failed but a few lift pumps have, so we went with the run dry tolerant DW pump, only to have it fail us rather quickly. Lazy me dropped in another used OEM pump while waiting for a free moment to send in the DW pump for warranty.... 3 races ago...

[turbogrill]

On 12/24/2021 at 1:19 PM, Grant said:

If I did a surge tank I think I'd give Radium's venturi jet pump a try instead of lift pumps. I've seen too many of those fail.

 

So the INLET would be from the FPR? What would happen if the surge tank is empty. Wouldn't there be an issue with priming the system? 

 

Are you using the ECUMASTER PMU and ECU as well?

Edited December 26, 2021 by turbogrill

[wvumtnbkr]

I don't think so.  Engine pump would be primed from swirl pot.  If swirl pot is empty, lift pump has issues 

[atxe30]

@Grant on the e30 i have two Carter P4594 lift pumps pulling from both sides of the OEM saddle tank. Those pumps are dry run tolerant. they feed a surge tank, etc. The only gaps in the system at present are fuel flow sensors on the lift lines so we know if one of the lift pumps has failed, and a second HP pump in the surge tank. I was pointed at the carters for their reliability and dry run tolerance when i built the system by some hard core e30 folks (might have been jim levie on the r3vlimited board..), but I believe those pumps have been mentioned in these forums as well...

 

Edit to add: getting the lift pumps out of the tank and co-located with HP pump means emergency servicing faster easier. The Carters are cheap enough that replacing them annually as a proactive per-season refresh is pretty painless....

Edited December 27, 2021 by atxe30

[Robmink]

Simple answer: yes, PWMing your low side pump will increase its life.  OEMs use PWM for three reasons: 1) most all of us use true returnless fuel systems that rely on the PWM to control the low side pressure, 2) PWM reduces pump wear and extends the life of the pump, and most importantly 3) PWM reduces the heat buildup in the fuel which reduces the load on the EVAP system. 

 

How you control the PWM will make all the difference if you want to realize your goal of reducing starvation failures.  If you use fuel pressure as a feedback you will need to write some type of control system, which is never an easy task, plus you will need to put some type of timer into the control system to shut down the pump if the desired pressure cannot be reached within X number of seconds, otherwise, the control system will just set the pump to max PWM once the fuel runs out and the pump will burn out anyway.  If you use the simpler RPM based PWM, then the fact that the engine cannot rev due to no fuel will act as a fail safe and the pump will run slowly and make less heat.  One other item to consider if you are running returnless is the heat buildup at the inlet to the high pressure pump.  Too much heat from an extended idle or heat soak after engine off can allow the fuel to boil and you will get no high side pressure.  I would add in a 10s run at max PWM at ignition on and during cranking to make sure that the fuel is not boiling.  I would also add in a slow ramp up of the PWM during periods of extended idling (red flag?) for the same reason.  Using a slow ramp up should stop the pump from going to max PWM if the engine cannot rev because of starvation.

 

Just for your information, we run PWMs from roughly 30-85% depending on the engine load and other variables.

[turbogrill]

A very simple solution would be to run the pump at ~70% all the time? 

 

 

[Grant]

On 12/25/2021 at 11:45 PM, turbogrill said:

So the INLET would be from the FPR? What would happen if the surge tank is empty. Wouldn't there be an issue with priming the system? 

 

Are you using the ECUMASTER PMU and ECU as well?

Yes, yes and no. I wired up my own marine breaker board and use the stock ECU, because stock ECUs are usually the best at knock detection. In hindsight I should have used a PMU though; it would have cost a bit more but really simplified the wiring.
 

7 hours ago, Robmink said:

Simple answer: yes, PWMing your low side pump will increase its life.  OEMs use PWM for three reasons: 1) most all of us use true returnless fuel systems that rely on the PWM to control the low side pressure, 2) PWM reduces pump wear and extends the life of the pump, and most importantly 3) PWM reduces the heat buildup in the fuel which reduces the load on the EVAP system. 

 

How you control the PWM will make all the difference if you want to realize your goal of reducing starvation failures.  If you use fuel pressure as a feedback you will need to write some type of control system, which is never an easy task, plus you will need to put some type of timer into the control system to shut down the pump if the desired pressure cannot be reached within X number of seconds, otherwise, the control system will just set the pump to max PWM once the fuel runs out and the pump will burn out anyway.  If you use the simpler RPM based PWM, then the fact that the engine cannot rev due to no fuel will act as a fail safe and the pump will run slowly and make less heat.  One other item to consider if you are running returnless is the heat buildup at the inlet to the high pressure pump.  Too much heat from an extended idle or heat soak after engine off can allow the fuel to boil and you will get no high side pressure.  I would add in a 10s run at max PWM at ignition on and during cranking to make sure that the fuel is not boiling.  I would also add in a slow ramp up of the PWM during periods of extended idling (red flag?) for the same reason.  Using a slow ramp up should stop the pump from going to max PWM if the engine cannot rev because of starvation.

 

Just for your information, we run PWMs from roughly 30-85% depending on the engine load and other variables.

Thanks! This is extremely helpful. I take it you're an automotive engineer?

 

I wasn't going to use a control system, since I'd keep the OEM in-tank regulator, basket and venturi jet pump. It might be tricky to figure out how much extra FPR bypass we need to keep the jet pump working though. I was thinking of an open loop system of RPM vs. duty cycle, or maybe something more complex using throttle position and its first derivative (to lead sharp throttle increases).

We did get some fuel boiling in the tank on the Friday practice day before the first race, when we came in for a driver change. You could hear it in the tank, and we think the boiling pressure + Sebring's bumps got the discriminator valve's ball stuck to its top. Release the pressure of course increased boiling due to dropping the boiling point. We ended up removing the lighter ball from the discriminator valve (making it just a rollover valve) for the race, and used a restrictor to prevent much fuel from sloshing out of it. During refueling we removed the restrictor.

 

It's weird we've never encountered fuel boiling before Sebring. Our exhaust used to be a lot closer to the fuel tank too. From what I've googled the boiling point of gasoline can vary wildly? It didn't feel hot when it pressure sprayed it all over me. I guess the fuel could have been boiling in the lines and coming out the pump outlet?

After seeing and hearing stories of discriminator valves getting stuck, I think my ideal fuel venting solution is a rollover valve + manual valve (or maybe Jiffy-tite fitting) for refuel venting, in parallel with a very small vent for tank breathing at the highest possible point.

[Robmink]

1 hour ago, turbogrill said:

A very simple solution would be to run the pump at ~70% all the time? 

 

 

 

Good point.  Find the minimum voltage/PWM you need to supply the engine at full song and run it at that level all the time.

[Robmink]

7 minutes ago, Grant said:

It's weird we've never encountered fuel boiling before Sebring. Our exhaust used to be a lot closer to the fuel tank too. From what I've googled the boiling point of gasoline can vary wildly? It didn't feel hot when it pressure sprayed it all over me. I guess the fuel could have been boiling in the lines and coming out the pump outlet?

 

 

Welcome to hot weather areas with winter fuel!  Winter fuel has a much higher volatility and boils easier, and it was nice and warm at Sebring this year. 

 

It doesn't feel so hot because it starts evaporating almost immediately.  As a good rule of thumb, assume the boiling point of fuel in a vented tank to be about 150F. 

 

Another note, if you have pressure in your tank something is wrong with your vent!  Fuel tanks in race cars should be at ambient pressure at all times.

[Grant]

2 minutes ago, Robmink said:

 

Welcome to hot weather areas with winter fuel!  Winter fuel has a much higher volatility and boils easier, and it was nice and warm at Sebring this year. 

 

It doesn't feel so hot because it starts evaporating almost immediately.  As a good rule of thumb, assume the boiling point of fuel in a vented tank to be about 150F. 

 

Another note, if you have pressure in your tank something is wrong with your vent!  Fuel tanks in race cars should be at ambient pressure at all times.

Ah thanks that makes sense, the fuel was probably from AMP in November.

The vent was stuck closed. I think Sebring's bumps knocked the lighter ball in the discriminator valve up, and the boiling pressure made it stick to the top.

[atxe30]

36 minutes ago, Grant said:

Ah thanks that makes sense, the fuel was probably from AMP in November.

The vent was stuck closed. I think Sebring's bumps knocked the lighter ball in the discriminator valve up, and the boiling pressure made it stick to the top.

well this is interesting. i have a radium discriminator valve in hand i have not yet installed on the e30. what valve were you using?

Edited December 27, 2021 by atxe30

[Grant]

3 minutes ago, atxe30 said:

well this is interesting. i have a radium discriminator valve in hand i have not yet installed on the e30. what valve were you using?

Fuel Safe's 12 AN. I also looked at Harmon's -10 ORB valve, but the lighter ball would stick to the top after turning it upside down in your hands... and stay there as you shook it. Only way to loosen it up again was to poke the ball with your finger.

[atxe30]

17 minutes ago, Grant said:

Fuel Safe's 12 AN. I also looked at Harmon's -10 ORB valve, but the lighter ball would stick to the top after turning it upside down in your hands... and stay there as you shook it. Only way to loosen it up again was to poke the ball with your finger.

ok, so the sticky thing is good to know. i might wait on the e30 install.....my more pressing effort is surge tanking the boxster. i will run an almost identical playbook on the boxster as i did with e30. i'd prefer a fuel cell with integrated surge tank, but OMG, that is a 10k line item for a 996/986 fuel tank pattern...ugghhhh

 

that said, i need to sort the vent system after emissions delete, etc....

 

@Grant if you don't mind me asking do you retrofit aeromotive or other aftermarket HP pump on your miata?

[frankrehnelt]

We’ve had the Radium float ball in the valve get stuck. We think the ball was swelling . We opened up the id .5 mm. 

[atxe30]

Also....for fuel flow, I would be curious to see what folks are using. I am looking at two potential paths: 1) aviation sensors and guages, and 2) marine. in digging around it appears the baja truck crowd has had success with Floscan: http://www.floscan.com/html/blue/seriesdetail.php?sid=14&catid=2. For aviation, I have an ask out to EI on these: https://www.iflyei.com/product-category/accessories/fuel-flow-transducers/

 

 

6 minutes ago, frankrehnelt said:

We’ve had the Radium float ball in the valve get stuck. We think the ball was swelling . We opened up the id .5 mm. 

thx

[Grant]

27 minutes ago, atxe30 said:

@Grant if you don't mind me asking do you retrofit aeromotive or other aftermarket HP pump on your miata?

My fuel system is all stock, with the exception of a Radium rail, originally so I could run a fuel pressure sensor. After adding the sensor I put a pressure damper on the front of the rail.

If you're asking why don't I retrofit an aftermarket system, well I might. In my experience OE stuff is usually a lot more reliable though. If the problem is simply the pump overheating due to being run low on fuel, the PWM option should be the simplest and most reliable solution.

There are also pumps that can be dropped in to the OE basket which might withstand heat better. I'm going to ask Mazda Motorsports about it.

Edited December 27, 2021 by Grant

[shanehutton]

Could vacuum in the sealed tank be killing your pump?  It will pressurize with heat but if you empty it and it's sealed I assume you have a fair amount of vacuum once it gets low.  Just a random thought.

[Grant]

6 hours ago, shanehutton said:

Could vacuum in the sealed tank be killing your pump?  It will pressurize with heat but if you empty it and it's sealed I assume you have a fair amount of vacuum once it gets low.  Just a random thought.

It only sealed because the discriminator valve ball got stuck. Vacuum would just un-seat the ball and make it no longer sealed. It takes very little pressure to hold that ball in place.

Edited December 28, 2021 by Grant