Cummins QSB 6.7 HP Rail Pressure Sensor

whdiesels · Post by **whdiesels** » Sun Jun 25, 2023 6:05 pm

Hi Everyone

Been out to a Hitachi loading shovel . Customer complaint is low power , with a miss fire and greyish white smoke .

Carried out a relative compression test and all looked good ( see screen shot )

Plugged in Texa to see some live data and faults stored . Found rail pressure at 750bar on idle and with a snap wide open throttle pressure reached 1800 bar which sounds plausible . Also carried out a cylinder cut out test and found Number 5 injector not firing . we fitted a new injector to No5 cylinder , i tried to explain to the owner it would be best to fit 6 new injectors because the other injectors will be in a similar condition , but he wanted the cheap fix . So we fitted a new injector to No5 and run up the engine . The miss fire was fixed , smoke had cleared and all seemed good . Until we put the engine under full load and its black smoking .

So carried out a few more tests . Back probed into the crank sensor , rail pressure sensor and attached an amp clamp around number 6 injector wiring for a sync also added the maths channel (crank A 60 idle) on the crank sensor to see the crank speed of every cylinder contribution . (see screen shot ) looks to me like the is 4 cylinders not contributing as much power as the other 2 ?

Also when looking at the rail pressure sensor on the scope the signal voltage is making a square wave intermittantly between the signal , looks like high voltage is no pressure and low voltage is high pressure . Ive never seen it like this before , could i have a problem with the pressure sensor ? i disconnected the sensor and started the engine , then plugged the sensor back in , you can see on the screenshot that with the engine running with sensor plug disconnected there is a constant 5 volt on the signal wire , the plug sensor back in and the square wave signal is back . not sure if this is right ? there is no pressure sensor faults in ecu and the live data looks good . has anyone seen a rail pressure sensor signal look like this before ?

Im going back to this job next week and plan on smoke testing the inlet under pressure to look for leaks and hopefully going to carry out a injector leak off test with the WPS 500 if i can find the right fittings .

Im thinking this engine wants the other 5 injectors replacing . looking at the rail pressure / crank speed screen shot, it looks like the is some injectors not contribitng ?

any more tests or things to check will be greatly appreicated

Thanks Wes

Post by **ben.martins** » Mon Jun 26, 2023 11:34 am

Hi Wes,

Got to love the comment from the customer - do it as cheap as possible!

Thanks for the post and great to see Pico being used more and more in the off highway sector. Out of interest what were the fault codes?

With regards to the waveform you have, it looks like you might have the temperature instead of the rail pressure signal. I say this as looking through the waveform library I have found an X15 waveform which shows the fuel rail sensor and the channel labelled with temperature appears identical to the one you have.

I'm not sure what scope you have but to help with some of the noisy signals, if you have a 4425 or a 4425A then it's worth using the sensor ground to reference too. This will clean up your signal. Also with these scopes, it's worth switching on bandwidth limit which can be activated in the channel options.

Do remember to turn it off again though when moving to another signal!

Injector leak off will of course help especially as there is a pattern in the crank maths suggesting some poor contribution from a number of cylinders. If possible can you upload the file to the this post or the library, so we can take a better look? Might be worth updating your PS7 version as well as some improvements to the crank maths will help clean up that math channel.

EGR would be my other point to check. Wondering if the Texa can give you actual and expected values for MAF as this might help to see if EGR is stuck.

If you've got WPS500 out before do a smoke test, can always get into the intake side and see the pull from each cylinder. Again, whilst it's out, WPS500 in the dipstick for crankcase pressure is a quick and easy test. Appreciate your relative compression test looks pretty good but whilst your up there with WPS500, it's an easy one to capture and non intrusive.

That being said, my go to engine health check test is the one you've done with crankshaft, injector or cylinder ID and rail pressure. However, I would definitely take another look round that rail pressure sensor as I'm not convinced that is the output.

Kind regards

Ben

PaulW · Post by **PaulW** » Mon Jun 26, 2023 11:53 am

Hey Wes,
I have seen this pressure signal on a Cummins CM2350 ISX15 fuel rail pressure sensor before. It was a 4 wire sensor. 2 wires are supply and return. The sensor has 2 pressure transducers built in for redundancy. This started I believe in around 2014 to 2015 (possibly 2013). What year is this engine? If you use the other signal wire at the rail pressure sensor the trace will look more like a traditional transducer trace, and not a pulsed signal. I work on-highway Interactive (IS) series Cummins engine trucks and not on the Quantum(QS) series. Not sure if this engine will have the same 4 wire sensor, but I have only seen this signal used on the redundant rail pressure sensor.

whdiesels · Post by **whdiesels** » Wed Jun 28, 2023 5:13 am

Hi Ben and Paul

Thanks for the replys

The fault code i have is 3727 High pressure common rail fuel pressure limiting valve - mechanical system not responding properly or out of adjustment
Im thinking if the is high leak back from injector the pressure regulator on the rail will be out of range . so hopefully gonna do a leak off test on Friday

Bem thanks for the tips , Going back Friday to carry out some more testing . I tried up loading my psdata to the waveform library but im strugglingto upload . Can I upload saved psdata? or do i upload them straight away? I updated to the latest Pico 7 version , can also login to the library but whem i click on waveform uploads it says there are currently no waveforms being uploaded . ive got loads of saved psdata

PaulW
Thanks for Reply . This machine is 2014 . But im sure its a 3 wire sensor but i will check again on Friday . I was thinking of using my break out leads then disconnect the signal wire and then measuring the signal of the sensor without it going back to the ecu and see what the signal looks like then ? or could it be possible this has got the wrong sensor fitted ? but then im sure there would be a code for this ?

Cheers Wes

Post by **ben.martins** » Wed Jun 28, 2023 6:06 pm

Hi Wes,

Keep us posted on how Friday goes. It'll be interesting to see if that is the signal you are supposed to be getting or if it's a different type. As you say might be worth breaking out the plug and see what each pin is doing even if that's the first test you do with a simple key on, engine off.

With regards to the waveform library might be worth taking a look at the video I did. I've skipped to the part where I walk through the upload. In the meantime I'll check in on the Admin side to see if it's there. https://youtu.be/M4YPL-zmZw4?t=464

There is a bulk upload which forms part of the waveform manager. We talk a bit about it in this video but we should look to do a quick update on the library. Steve did talk through it on the following forum post as well - viewtopic.php?p=103911#p103911, but maybe a video might be useful as well.

Kind regards

Ben

whdiesels · Post by **whdiesels** » Sun Jul 09, 2023 9:45 pm

Hi Everyone
Got this one sorted with a new set of injectors

Ben , I was on the wrong signal wire . PaulW, Yes this sensor has 4 wires on HP rail sensor , Power, Ground, and 2 sinal wires. I back probed into the other signal wire and found the rail pressure signal and everything looked like i expected , So thanks for the help with that.

I also managed to plumb into the injector leak off gallery at the flywheel end of the cylinder head with the WPS500 . But i find it difficult to call a bad injector looking at the leak off waveform . I tested the leak off 6 times with a snap wide open throttle and the amp clamp around a different injector wire each time , and the leak off waveforms looks different every time . I ended up removing all injectors and taking back to the workshop and bench testing them and sure enough the other 5 injectors was bad with high leak back . I wanted to do another leak off test with the new injectors but the owner wanted his machine back so that didnt happen.

But i did manage to back probe into the crank sensor , rail sensor and amp clamp around number 6 injector , and you can see clearly that each cylinder is making a nice even contribution with the new set of injectors compared to when i did the same test with the bad injectors .
This is defo gonna be my go to test when im looking for bad injectors in the future .

Thanks for the help

Cheers Wes

Post by **ben.martins** » Mon Jul 10, 2023 11:43 am

Thanks for the update Wes, and great work on getting it sorted.

Awesome use of the crank math channel as well. I use it a lot especially for performance based issues where there is either multiple misfires or worse - no fault codes. You can see more on it here https://www.picoauto.com/library/case-s ... ic-process.

Keep your eyes peeled as this will hopefully be a guided test making it's way to PS7 in the future which will hopefully help.

Thanks for the images as well Wes, helps to bring it to life when you can see the difference between before and after. Not necessarily when in the middle of diagnosis, but if you're having to do reports for the customer using the option to include an other view will let you have the option to have the before and after capture on the same screen. That being said, if you need your known good on the screen whilst capturing against the machine being tested, multiple views can be useful. To add a new scope view click More > Views you will then have the option to add a new scope spectrum, XY or persistence view. For the most part scope will be the most common used but the others do have a use. XY for example is one I use when perform PQ tests on hydraulic systems.

The other thing I noticed was the current clamp appears to be inverted. In PS7 there is a quick invert button under the channel options which can be used during both capturing and analysis. Click the channel and then hit invert.

If you get a chance feel free to upload these waveforms to the library and we can all take a look. It'll be interesting to see the relationship between the injectors and rail pressure. Looking at the image it's not completely clear as I think you have a lowpass filter active. If you're using a newer 4425A or even a 4425, you can make use of the floating input. I find using the rail pressure ground to reference the signal to provides a much cleaner waveform with less interference and needs less filtering post capture. If you're using a common ground scope though, like a 4423, then stick with chassis or battery negative as your ground reference.

Keep up the good work!

Kind regards

Ben

PaulW · Post by **PaulW** » Wed Jul 12, 2023 11:53 am

Hey Wes,
Thanks for the update. Glad to hear you got this figured out. I'm sure the owner is happy to have his machine back. Thanks for the link Ben.

PicoScope 7 Automotive

PicoScopes
& kits

Noise, Vibration and Balancing

Breakout
leads/boxes

For vehicle manufacturers

Milliohm

Current clamps

EV kits and
accessories

Pressure sensors

Accessories

PicoScope forum

Contact

Training

Automotive guided tests

Case studies

Newsletters

Training

A to Z of PicoScope

Documentation

Waveform library

Videos

Reviews and awards

Automotive Support Forum