"Noisy" tpsDOT

This is a forum for discussing the development and testing of alpha MS2/Extra code. Documentation
(Runs on MS2 and Microsquirt)

Moderators: jsmcortina, muythaibxr

1031
MS/Extra Newbie
Posts: 17
Joined: Wed Apr 14, 2010 4:39 am

Re: "Noisy" tpsDOT

Post by 1031 »

davcol wrote: i would love to give it a try, so could you explain what polyfuse you removed and what kind of fuse you replaced it with :?:
Hello. Yes replace polyfuse F1 (seen here upper left corner of V3.0 board http://www.msextra.com/doc/ms2extra/MS2 ... htm#layout, shematics -> http://www.msextra.com/doc/general/ms2v3schems.html#p6)
Component data -> http://parts.digikey.com/1/parts/707461 ... xe050.html
Original polyfuse is rated 0.5 amp (continous) and 1 amp trip current.
I used this type of fuse -> http://www.digikey.fi/product-detail/en ... -ND/245317
it was 630mA slowblow type.

I stongly believe that polyfuses resistance is source for that noise R Min/Max0.500 ~ 0.770 Ohm So in theory that o.5A and 0.77Ohm gives about 380mV noise signal .
gslender
Super MS/Extra'er
Posts: 1072
Joined: Fri Sep 16, 2011 5:29 am
Location: Brisbane, Australia
Contact:

Re: "Noisy" tpsDOT

Post by gslender »

1031 wrote:I stongly believe that polyfuses resistance is source for that noise R Min/Max0.500 ~ 0.770 Ohm So in theory that o.5A and 0.77Ohm gives about 380mV noise signal .
Doesn't really explain the same issue existing for folks with DIYPNP boards does it?
Mazda MX5 + MS3 Pro
jsmcortina
Site Admin
Posts: 39587
Joined: Mon May 03, 2004 1:34 am
Location: Birmingham, UK
Contact:

Re: "Noisy" tpsDOT

Post by jsmcortina »

gslender wrote:
1031 wrote:I stongly believe that polyfuses resistance is source for that noise R Min/Max0.500 ~ 0.770 Ohm So in theory that o.5A and 0.77Ohm gives about 380mV noise signal .
Doesn't really explain the same issue existing for folks with DIYPNP boards does it?
I think so - the Microsquirt module has a polyfuse too.
http://www.msextra.com/doc/general/moduleschems.html

James
I can repair or upgrade Megasquirts in UK. http://www.jamesmurrayengineering.co.uk

My Success story: http://www.msextra.com/forums/viewtopic ... 04&t=34277
MSEXTRA documentation at: http://www.msextra.com/doc/index.html
New users, please read the "Forum Help Page".
1031
MS/Extra Newbie
Posts: 17
Joined: Wed Apr 14, 2010 4:39 am

Re: "Noisy" tpsDOT

Post by 1031 »

jsmcortina wrote:
gslender wrote:
1031 wrote:I stongly believe that polyfuses resistance is source for that noise R Min/Max0.500 ~ 0.770 Ohm So in theory that o.5A and 0.77Ohm gives about 380mV noise signal .
Doesn't really explain the same issue existing for folks with DIYPNP boards does it?
I think so - the Microsquirt module has a polyfuse too.
http://www.msextra.com/doc/general/moduleschems.html

James
Of course there can be different sources for noise also, and i´m not saying that "polyfuse noise" is culprit for every case. But thats not hard thing to check by changing that fuse to normal type.
racingmini_mtl
Super MS/Extra'er
Posts: 9128
Joined: Sun May 02, 2004 6:51 am
Location: Quebec, Canada
Contact:

Re: "Noisy" tpsDOT

Post by racingmini_mtl »

1031 wrote:I stongly believe that polyfuses resistance is source for that noise R Min/Max0.500 ~ 0.770 Ohm So in theory that o.5A and 0.77Ohm gives about 380mV noise signal .
I'm sorry but that doesn't explain a noise source. What you show here is a voltage drop which would just give a constant offset.

Jean
jbperf.com Main site . . . . . . . . . . . . . . . . . . . . . . jbperf.com Forum
Image
1031
MS/Extra Newbie
Posts: 17
Joined: Wed Apr 14, 2010 4:39 am

Re: "Noisy" tpsDOT

Post by 1031 »

Everything that are after that polyfuse can made voltage drop to that vref voltage. And if there is ie. Hall sensor that uses vref. Then current (pulses) it takes are seen as voltage drops on vref. Also when vref/sensor supply have some source impedanse (polyfuses resistance) noise from close wirings etc. Can enter to thar vref voltage more easily. I never looked that noise more deeply, just used oscilloscope and noticed that after that polyfuse there was visible ripple voltage.. I tested some capacitors (sanyo os-cons etc..)but non of those caps didnt helped. So then i tested that normal fuse instead of that poly fuse and noise/ripple was gone.
JasonC SBB
Helpful MS/Extra'er
Posts: 120
Joined: Wed Oct 26, 2011 7:36 pm

Re: "Noisy" tpsDOT

Post by JasonC SBB »

Was the ripple voltage, mainly spikes riding on top of a constant DC?
Was the constant DC zero?

Did you say you tried Oscons across the polyfuse?
robs
Master MS/Extra'er
Posts: 564
Joined: Sun Jan 17, 2010 4:26 pm
Location: Sydney, Australia

Re: "Noisy" tpsDOT

Post by robs »

1031 wrote:Hello to all, I´m still bit lost why you are trying to corret HW errors whit software filters etc.
My own excuse is that I only have the vaguest idea about hardware, but am pretty comfortable with software. I'm the proverbial man with a hammer -- everything looks like a nail.

Jason has certainly talked about deficiencies in the MS TPS hardware filter (too high a cutoff frequency) and (IIRC) has suggested better values for the capacitor. Your polyfuse suggestion sounds promising, but your description of having a noisy TPS on the stim as well as on the car sounds strange. Mine's quiet on the stim and quiet in the car until the engine's turning.

Finally, as far as this discussion goes, for my part the TPS problem's already solved. Grant has said so too. So don't be too worried about this thread continuing. I am finding it interesting exploring Jason's ideas and learning a little about how the hardware guys think about problems. If it leads to further improvements (to TPS, MAP, RPM...) then that's a good thing too.

Have fun,

Rob.
robs
Master MS/Extra'er
Posts: 564
Joined: Sun Jan 17, 2010 4:26 pm
Location: Sydney, Australia

Re: "Noisy" tpsDOT

Post by robs »

A deficiency in my earlier graphs was that I was using gaussian noise for the jitter. Good enough for sanity checking, but the noise on my TPS has never looked normally distributed. It had a fair amount of random looking jitter, but some quite large spikes too. For some time I've been meaning to get a capture of the TPS on the scope and use it as my noise base. Unfortunately, my scope's display has gone haywire. I can read the traces, but the text is basically unreadable. I can use a screen dump and read that, but it makes it quite a chore to use.

Anyhow, I have gone through the pain and scoped 0.3 s worth of free-run data, DC coupled, throttle closed, engine running, and it does look something like I expected, but I think I've scoped the wrong thing so I won't bother trying the data against the R code until I've had a sanity check.

My MS is buried behind the dash and it's about an hour's work to get at it, so I just probed at the TPS sensor itself. Channel 1 on the TPS signal, Channel 2 on the TPS signal ground, ground lead to the battery -ve. The MS's view of the TPS should be Chan1-Chan2, or so I thought. Looking at the fluctuations of more than 0.5V makes me suspicious. I guess this is before the RC filter on the input. If I were to scope after the filter I expect I'd get a much quieter signal. Must be something like that because my TPS certainly doesn't fluctuate as much as this.

Here's a screen dump from the scope (sorry about the ghastly colours, but that skin is the least unreadable on my damaged display):
tpsscreen.jpg
And here is a csv of the numbers for the whole capture if you want to graph them for yourself:
tpsdump.zip
So, a couple of questions:
  • Does this look like other peoples' traces?
  • Seems to me that the signal ground (Chan2) is particularly noisy, appreciably above 0V and won't be filtered -- so that noise will be reflected for all sensors.
  • If this looks right, do you think I can cook these figures to generate the RC filtered input. I really don't want to have to get in to my MS. Perhaps someone else could scope their TPS input??
Have fun,

Rob.
1031
MS/Extra Newbie
Posts: 17
Joined: Wed Apr 14, 2010 4:39 am

Re: "Noisy" tpsDOT

Post by 1031 »

JasonC SBB wrote:Was the ripple voltage, mainly spikes riding on top of a constant DC?
Was the constant DC zero?

Did you say you tried Oscons across the polyfuse?
I´m going from memory here.. Ripple was random spikes here and there and low amplitude square wave like ripple riding top of dc.
I tested that capacitor across, before/after between vref-gnd, improvement was so small whit cap that it was near impossibe to see at o-scopes screen.
Last edited by 1031 on Tue Apr 03, 2012 10:53 pm, edited 1 time in total.
1031
MS/Extra Newbie
Posts: 17
Joined: Wed Apr 14, 2010 4:39 am

Re: "Noisy" tpsDOT

Post by 1031 »

robs wrote: My own excuse is that I only have the vaguest idea about hardware, but am pretty comfortable with software. .
My case is another way around.. Software is mumbo jumbo.. to me :)
Peter Florance
Super MS/Extra'er
Posts: 3653
Joined: Fri Apr 02, 2004 8:40 pm
Location: Virginia Beach, VA
Contact:

Re: "Noisy" tpsDOT

Post by Peter Florance »

Rob, try attaching 330uf-470uf/16v on TPS between Vref and Gnd (DO NOT attach to engine or chassis ground). and see if it is quieter. If it is, then Vref needs some help.

Just to confirm, your sensor ground only terminates at megasquirt connector and is not grounded to car anywhere
Peter Florance
PF Tuning
81 BMW Euro 528i ESP Car
60-2 Wheel LS2 Coils, Low Z Inj
Co-Driver 1999 BMW E46 DSP car.
JasonC SBB
Helpful MS/Extra'er
Posts: 120
Joined: Wed Oct 26, 2011 7:36 pm

Re: "Noisy" tpsDOT

Post by JasonC SBB »

robs,

The noise you see is typical in a noisy environment like a car.

Where did you ground the scope probe, physically? How far from the TPS?
Are you powering the scope from the mains, from an inverter powered by the car, or from a battery?

Try this experiment. Take the scope probe and place it on the same point as where the probe's ground clip is connected. i.e. the tip is shorted. You would think it should read zero, right? You may need to turn up the gain on the scope (make it read mV/div - think about the A/D resolution being 5 mV for 1 bit)
If it still reads some spikiness, the spikey noise is "common mode noise" getting through the scope's front end amplifier. Repeat with the scope grounded to the TPS ground pin.

If you get no noise, for fun, disconnect the probe, clip the ground clip onto the tip, then touch various ground points on the engine. (Do this with no other probes connected). On the valve cover near the ignition coils, near the alternator, etc.
Try it again with a 2nd probe whose ground clip is connected elsewhere.

Then again connect only one probe to the scope. Take the ground clip and short it again to the tip, and make sure the ground clip forms a large loop with the tip. This time, don't touch the tip/ground anywhere, but imagine it's a loop antenna. Try to see if you can pick up noise near the ignition coils and its wiring, alternator, injectors, etc.

The above demonstration shows you how easily noise can get picked up by circuits.... I do this demonstration for newbie engineers to teach them proper oscilloscope measurement techniques and noise coupling principles.
JasonC SBB
Helpful MS/Extra'er
Posts: 120
Joined: Wed Oct 26, 2011 7:36 pm

Re: "Noisy" tpsDOT

Post by JasonC SBB »

1031 wrote:
JasonC SBB wrote:Was the ripple voltage, mainly spikes riding on top of a constant DC?
Was the constant DC zero?

Did you say you tried Oscons across the polyfuse?
I´m going from memory here.. Ripple was random spikes here and there and low amplitude square wave like ripple riding top of dc.
I tested that capacitor across, before/after between vref-gnd, improvement was so small whit cap that it was near impossibe to see at o-scopes screen.

Sounds like the scope may have been lying to you about the presence of spikes (see my post about common mode noise getting past the scope's front end "common mode noise rejection").

In any case, several pages ago, I explained why shorting out the polyfuse works.
I did something similar, when I hacked my AEM ECU. It does NOT have a polyfuse, the analog and digital sections were connected on the PCB at one point. Its layout had ~50 milliohms between the analog and digital section of the PCBs *at DC*, but at high frequencies it was a much higher impedance. And so spikey noise got into the analog readings. I fixed the layout in a manner described in a link.

I showed before and after datalogs showing a huge improvement in spikey noise. I also linked to a page explaining why a certain "common wisdom" method of laying out mixed signal PCBs is bad advice (and was followed by AEM).
robs
Master MS/Extra'er
Posts: 564
Joined: Sun Jan 17, 2010 4:26 pm
Location: Sydney, Australia

Re: "Noisy" tpsDOT

Post by robs »

Peter Florance wrote:Rob, try attaching 330uf-470uf/16v on TPS between Vref and Gnd (DO NOT attach to engine or chassis ground). and see if it is quieter. If it is, then Vref needs some help.

Just to confirm, your sensor ground only terminates at megasquirt connector and is not grounded to car anywhere
Thanks Peter. I'll have a rummage in my (pretty small) collection of caps and see if I have one -- or else add it to my shopping list.

And yes, all the sensor grounds go to one pin on the MS connector and then, as I recall, a bunch of ground wires went from the MS to a common ground terminal (as originally used by the stock Bosch setup). Some of those grounds go through spade connectors from the original loom. The TPS spade connector being loose certainly made a lot of noise. Is it possible that a spade connector on (say) the CLT ground would be causing noise on the TPS ground? Guess I don't really need to ask since I've just said they all come back to the same place. Perhaps time I went through the whole harness looking for potential problems.

Have fun,

Rob.
robs
Master MS/Extra'er
Posts: 564
Joined: Sun Jan 17, 2010 4:26 pm
Location: Sydney, Australia

Re: "Noisy" tpsDOT

Post by robs »

JasonC SBB wrote: The noise you see is typical in a noisy environment like a car.

Where did you ground the scope probe, physically? How far from the TPS?
Are you powering the scope from the mains, from an inverter powered by the car, or from a battery?
Thanks Jason.

Probe was grounded to the battery -ve which is about 2' as the crow flies from the TPS, and I'd guess about 6' of copper by the time the electron makes its way into the cabin and back to the battery. Scope was powered from mains.
JasonC SBB wrote:
Try this experiment. Take the scope probe and place it on the same point as where the probe's ground clip is connected. i.e. the tip is shorted. You would think it should read zero, right? You may need to turn up the gain on the scope (make it read mV/div - think about the A/D resolution being 5 mV for 1 bit)
If it still reads some spikiness, the spikey noise is "common mode noise" getting through the scope's front end amplifier. Repeat with the scope grounded to the TPS ground pin.

If you get no noise, for fun, disconnect the probe, clip the ground clip onto the tip, then touch various ground points on the engine. (Do this with no other probes connected). On the valve cover near the ignition coils, near the alternator, etc.
Try it again with a 2nd probe whose ground clip is connected elsewhere.
I would like to better understand oscilloscopes, but I'm afraid mine is not much fun to use at the moment. I did do the first test and the trace was pretty much zero, just a slight flutter at highest frequency, lowest voltage range, so not much sign of common mode noise there.

I have previously seen how readily the scope can pick up apparently random voltages (often when the probe slipped off) and have learnt to be wary of signals that don't make sense (as far as any of these deep mysteries make any sense to me). It usually comes down to something being disconnected, or some scope control set in the wrong range. I do understand, in a theoretical sense, how magnetic fields abound and can induce voltages all over the place, and higher frequencies can be much weirder to deal with than lower ones, and almost any bit of wire can act like an antenna; but this understanding doesn't get me terribly far, except to feel pretty happy that I ended up with a career in software. Anyhow, when I got the above trace I had verified that the TPS trace did go up and down with the throttle and everything did make reasonable sense, noisy as it was.

So if this noise is fairly typical, other people who have probed their TPS will see something like the above trace?

Assuming yes, was I right in what I asked before? These values surely aren't what the MS CPU's A2D unit is seeing? In order to get suitable real-world input for testing the software filters, I need to get the output of the hardware filter don't I? Or perhaps feed the figures from my trace into a filter function to simulate what the hardware is doing? Or maybe this is a bit of a fool's errand?

Have fun,

Rob.
Peter Florance
Super MS/Extra'er
Posts: 3653
Joined: Fri Apr 02, 2004 8:40 pm
Location: Virginia Beach, VA
Contact:

Re: "Noisy" tpsDOT

Post by Peter Florance »

robs wrote:
And yes, all the sensor grounds go to one pin on the MS connector and then, as I recall, a bunch of ground wires went from the MS to a common ground terminal (as originally used by the stock Bosch setup). Some of those grounds go through spade connectors from the original loom. The TPS spade connector being loose certainly made a lot of noise. Is it possible that a spade connector on (say) the CLT ground would be causing noise on the TPS ground? Guess I don't really need to ask since I've just said they all come back to the same place. Perhaps time I went through the whole harness looking for potential problems.

Have fun,

Rob.
The test would be to unplug the MS; at that point the gnd terminals of sensors (TPS CLT MAT) should not have continuity to the vehicle ground.
Peter Florance
PF Tuning
81 BMW Euro 528i ESP Car
60-2 Wheel LS2 Coils, Low Z Inj
Co-Driver 1999 BMW E46 DSP car.
JasonC SBB
Helpful MS/Extra'er
Posts: 120
Joined: Wed Oct 26, 2011 7:36 pm

Re: "Noisy" tpsDOT

Post by JasonC SBB »

robs wrote: Assuming yes, was I right in what I asked before? These values surely aren't what the MS CPU's A2D unit is seeing? In order to get suitable real-world input for testing the software filters, I need to get the output of the hardware filter don't I? Or perhaps feed the figures from my trace into a filter function to simulate what the hardware is doing?
You need to connect the scope probe ground clip to the TPS signal ground, and nowhere else. Other channel probes should not be grounded elsewhere.

The signal at the A/D input will be the same, plus noise picked up by the cabling, plus noise that gets through bad PCB layout and grounding. This assumes that there is only one path for the current to flow from the TPS signal ground, to the A/D analog ground pin. (If not there will be ground offsets which can manifest as a DC offset, high frequency noise, or a combo).
1031
MS/Extra Newbie
Posts: 17
Joined: Wed Apr 14, 2010 4:39 am

Re: "Noisy" tpsDOT

Post by 1031 »

JasonC SBB wrote:
1031 wrote:
JasonC SBB wrote:Was the ripple voltage, mainly spikes riding on top of a constant DC?
Was the constant DC zero?

Did you say you tried Oscons across the polyfuse?
I´m going from memory here.. Ripple was random spikes here and there and low amplitude square wave like ripple riding top of dc.
I tested that capacitor across, before/after between vref-gnd, improvement was so small whit cap that it was near impossibe to see at o-scopes screen.

Sounds like the scope may have been lying to you about the presence of spikes (see my post about common mode noise getting past the scope's front end "common mode noise rejection").

In any case, several pages ago, I explained why shorting out the polyfuse works.
I did something similar, when I hacked my AEM ECU. It does NOT have a polyfuse, the analog and digital sections were connected on the PCB at one point. Its layout had ~50 milliohms between the analog and digital section of the PCBs *at DC*, but at high frequencies it was a much higher impedance. And so spikey noise got into the analog readings. I fixed the layout in a manner described in a link.


I showed before and after datalogs showing a huge improvement in spikey noise. I also linked to a page explaining why a certain "common wisdom" method of laying out mixed signal PCBs is bad advice (and was followed by AEM).
I had scope connected to isolationtransformar, so no groundloops. Also i measured that noise in my workplace, whit megastim wich was feeded from lab.supply. scopes gnd was at mega´s pcb senror gnd. line. Also if you read my previos posts, that noise came out of sudden. If that noise was related to poor PCB layout, how that can work ok for one year and start doing that noise out of sudden.. I must look if i have that polyfuse still somewhere and test if it´s resistance has changed.
Peter Florance
Super MS/Extra'er
Posts: 3653
Joined: Fri Apr 02, 2004 8:40 pm
Location: Virginia Beach, VA
Contact:

Re: "Noisy" tpsDOT

Post by Peter Florance »

1031 wrote:I had scope connected to isolationtransformar, so no groundloops. Also i measured that noise in my workplace, whit megastim wich was feeded from lab.supply. scopes gnd was at mega´s pcb senror gnd. line. Also if you read my previos posts, that noise came out of sudden. If that noise was related to poor PCB layout, how that can work ok for one year and start doing that noise out of sudden.. I must look if i have that polyfuse still somewhere and test if it´s resistance has changed.
Most importantly, the probe's ground wire should be short for best results. Long ground wire will have more inductance and show noise that isn't really there. When testing SMPS, I used to make a ground wire out of pape-rclip and wrapped it around end of probe's ground ring. It was as long as the probe tip itself.
Peter Florance
PF Tuning
81 BMW Euro 528i ESP Car
60-2 Wheel LS2 Coils, Low Z Inj
Co-Driver 1999 BMW E46 DSP car.
Post Reply