My (stupid) questions to Jared

puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

My (stupid) questions to Jared

Post by puff »

This thread should have been started long ago... ;-)
Here's the first question - I can't explain..

I'm continuing populating the frankenstein board (0.1).

Inp5 -VBATT - has 36K/10K divider before the opamp (still using lmv324i thing), then 10k/10k divider to PA0 on the disco board.
Playing with my power supply unit, I made just three measurements:

Input PA0 Input/PA0
7.96V 0.83V 9.590361446
9.90V 1.04V 9.519230769
14.02V 1.49V 9.409395973

Is that okay? Why is that so? Which value should I use for coefficient in tuner studio?
All the times 5V rail of the board was 4.97V
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

Here's the second one: I am using LM1815 conditioner, which is powered from 5V, and which is expected to ouput 5V pulses.
The suggested way of connecting the signal is from the crank position sensor is to feed it to Input12, through OPAMP and a jumper wire to PC6. Would it be wrong to connect my VR board (with 5V output) directly to the jumper wire (PC6)?
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

Speaking of PC6. Currently I have 1k/1k divider on the LMV324 output. Will it be safe for LMV/MCP thing output to receive 3.3V from discovery (trigger emulator)? Will it be safe for discovery to feed this 3.3 signal, which is basically grounded through 1K resistor?
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: My (stupid) questions to Jared

Post by kb1gtt »

Frankenstien R0.1, that was more than 5 minutes ago, I forget what that was about :)

Per question 1,
puff wrote:Is that okay? Why is that so?
Sure why not :) I'm assuming R250 = 36k, and R251 = 10k, as well R252 is also 10k.

I see it like the below, it appears to be fairly linear. I forget how it was skewing at the low end and top end for that chip. That was several years ago perhaps it skewed around 0.50V of either rail. From a quick eyeball, I believe you have not yet hit the skewing areas of the op-amp. I seem to recall the lower rail was the problem. We changed the resistor divider ratio to avoid the issues at the higher end, but 0V can not be scaled, so problems at the low end. To me it appears the ADC voltage just above 3% error, and consistent. Even with 1% resistor the tolerance stack up makes sense to me. I see the front end resistor divider being 1%, then the op-amp offset voltage, which is then again has 1% from the output divider.
Capture.PNG
Capture.PNG (17.26 KiB) Viewed 17252 times
puff wrote:Which value should I use for coefficient in tuner studio?
I do not know the TS constants. That would be an @ question. Either that or provide it with a solid 12V or 14V voltage and adjust the constant until you read that voltage. If you eyeball it like this, you can calibrate out some of the resistor tolerances.
puff wrote:All the times 5V rail of the board was 4.97V
As 5V dip's, so does the op-amp voltage. If your high end skewing was at 0.05V from the top rail, then you drop this by an additional 0.03V, your top end voltage would skew at 0.08 from 5V. However at 14V you are only 3V, so you have allot of room before you start to skew. The analog is based on the 3V regulator, this regulator did not drop voltage because of this variation in 5V. So it would have read the same as it always does. AKA ratiometric issues do not apply in this situation. It's basically an absolute reading, and it will function until your 5V drops significantly more than you have noted.
Welcome to the friendlier side of internet crazy :)
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: My (stupid) questions to Jared

Post by kb1gtt »

puff wrote:Here's the second one: I am using LM1815 conditioner, which is powered from 5V, and which is expected to ouput 5V pulses.
The suggested way of connecting the signal is from the crank position sensor is to feed it to Input12, through OPAMP and a jumper wire to PC6. Would it be wrong to connect my VR board (with 5V output) directly to the jumper wire (PC6)?
I believe direct connection would likely be fine, but is has not been tested or proven and you have less protection against voltage spikes. As well if the STM32 software thinks it is connected to ADC, some odd things could happen. It is safer to use the opamp as it drops your signal to a 3V signal instead of a 5V signal.
Welcome to the friendlier side of internet crazy :)
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: My (stupid) questions to Jared

Post by kb1gtt »

puff wrote:Speaking of PC6. Currently I have 1k/1k divider on the LMV324 output. Will it be safe for LMV/MCP thing output to receive 3.3V from discovery (trigger emulator)? Will it be safe for discovery to feed this 3.3 signal, which is basically grounded through 1K resistor?
I'm making many assumptions, but probably this would be OK. A shared GND via 1k might cause the low side to not trigger as a 0 instead of a 1. But the inputs of the discovery are 3V friendly so your not going to damage anything. If you have a problem I believe it would be failure to change the input condition.
Welcome to the friendlier side of internet crazy :)
stefanst
contributor
contributor
Posts: 703
Joined: Wed Feb 17, 2016 12:24 am
Location: USA 08530

Re: My (stupid) questions to Jared

Post by stefanst »

IMHO the absolute voltage value displayed isn't of grave importance. Just pick a multiplier value and live with it. I'd personally pick the value that displays the correct voltage under normal operating conditions- i.E. alternator charging.

There's only two places where the ECU actually uses the measured voltage value afaik: Injector dead-time compensation and alternator regulation. I believe your alternator is self-regulating, so that leaves us with injector dead-time. And here the voltage that the injectors see is likely slightly different from the voltage your ECU sees. Additionally, since you likely haven't determined your dead-times with a proper injector-testing rig, you are likely running with guessed-at values anyway. High precision in measuring Vbatt is thus not needed.
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

probably that's true. so far no more stupid questions.
User avatar
andreika
donator
donator
Posts: 461
Joined: Mon Feb 13, 2017 2:35 pm
Location: Kiev

Re: My (stupid) questions to Jared

Post by andreika »

stefanst wrote:There's only two places where the ECU actually uses the measured voltage value afaik: Injector dead-time compensation and alternator regulation.
Don't we have the ignition dwell table ("coil accumulation time based on voltage" curves)? :(
stefanst
contributor
contributor
Posts: 703
Joined: Wed Feb 17, 2016 12:24 am
Location: USA 08530

Re: My (stupid) questions to Jared

Post by stefanst »

andreika wrote:Don't we have the ignition dwell table ("coil accumulation time based on voltage" curves)? :(
You are absolutely right. I had forgotten about that. I guess my AFAIK should have been IAMAICR (Inasmuch as I can remember). :D
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: My (stupid) questions to Jared

Post by AndreyB »

andreika wrote: Don't we have the ignition dwell table ("coil accumulation time based on voltage" curves)? :(
http://rusefi.com/wiki/index.php?title=Manual:Software:Ignition_Control
Dwell is controlled by a curve by RPM. At the moment there is no voltage correction for dwell time based on VBATT.
Do we need to correct running dwell by VBATT?
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

once again, here's the graph for my ignition module. without this voltage correction, cranking becomes sort of a problem.
If adding this curve is that difficult, i'd suggest at least a separate correction for cranking...
Снимок экрана 2017-05-12 в 12.06.48.png
Снимок экрана 2017-05-12 в 12.06.48.png (95.57 KiB) Viewed 17134 times
Have no idea of its correlation with RPM.
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

I'd also suggest to implement double-triple ignition during cranking
these new COPs and ignition modules won't give you any benefits of increased dwell (spark energy will always be the same), but they can easily allow to make several sparks during the single engine cycle.
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: My (stupid) questions to Jared

Post by kb1gtt »

I agree that multi-spark during cranking would be helpful, especially with cold weather cranking.
Welcome to the friendlier side of internet crazy :)
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: My (stupid) questions to Jared

Post by AndreyB »

puff wrote:without this voltage correction, cranking becomes sort of a problem
I prefer to keep it simple, do how do we prove that VBATT correction is needed during cranking?

http://rusefi.com/wiki/index.php?title=Manual:Software:Ignition_Control

Durining running dwell is controlled by a curve by RPM. At the moment there is no voltage correction for dwell time based on VBATT.
During cranking dwell time is set in either ms or in engine revolution degrees.
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

By RPM curve you mean just two states: below cranging RPM or above cranging RPM?

Once again, what is that dependency on RPM? Why is it so?

It is understandable that dwell time should be short enough, so that the coil to be able to serve all the designated plugs. At high RPMs the coil might not have enough time to get fully charged, but then again, it's the limitation of design, why introducing these artificial limits?

From my understanding, dwell is a funciton of (from) voltage, not RPM.
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: My (stupid) questions to Jared

Post by AndreyB »

puff wrote:By RPM curve you mean just two states: below cranging RPM or above cranging RPM?
@ can I volunteer you to make a video of all TS dialogs? Just begin from top left to bottom right?

https://github.com/rusefi/rusefi/issues/373

As maybe a more useful option we can start by .png screen shots of all menus, we will post on the wiki and we would be linking to them?

I know it's boring labour but it would help, also it would answer you question. I only have that much bandwidth to point people to TS dialogs.
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

you are talking about this one?
Снимок экрана 2017-05-12 в 15.25.27.png
Снимок экрана 2017-05-12 в 15.25.27.png (131.73 KiB) Viewed 17113 times
the question is why you make it dwell/RPM, not dwell/voltage?
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: My (stupid) questions to Jared

Post by kb1gtt »

My un-educated guesses, so take this with a grain of salt, preferably from the rim of a glass filled with a margarita.

My general theory in the below is based on the assumption that the coil is not normally fully saturated, such that an increased dwell results in more energy absorbed by the spark event. The block and wires do not change, so the assumption is more absorbed energy means hotter spark.

-- Most coils these days cannot with stand a 100% duty cycles. If you turn your coil on for to long, it will over heat and then fail. It's possible that at lower RPM they identify that a longer dwell is not going to over heat the coil, so perhaps they allow for a longer dwell as they know you will not damage the coil, and why not allow for a hotter spark.
-- If your system is capable of driving too hot of a spark, you could fatigue your plug via over heating the electrodes. If you are full RPM you you're going to spark much more frequently with less time between spark event to allow cooling down. Resulting in your electrode having to absorb more heat. If you decrease the dwell, you can decrease the heat absorbed by the electrode.
-- At low RPM you may get less mixing as the air is less violently trashed around before it enters the cylinder. So perhaps it needs a hotter spark to ensure ignition.
-- I would also guess that a hotter spark at low RPM would help increase the flame front speed as the flame front is an exponential energy growth, if you can start it off a bit hotter it probably makes for a faster starting point, which could potentially make a bit more power.

What is that program? Is that a professionally develop and commonly used tool, or is that just someone who made what ever they though would work?
Welcome to the friendlier side of internet crazy :)
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

Which program are you referring to?

In my case of a low-rev engine, let's suppose the max RPM is 6000, which is 100 spark events per second, which is 10ms per event, while the suggested dwell time for 14.5V is just 3ms, thus the duty cycle is below 30%. But then again, there are engines reving to 12000 and even higher.

As far as I understand, those smart logic-driven COPs and ignition modules have some sort of protection against coil overcharge, which is seemingly why making higher dwell times at the given voltage is unnecessary and excessive.
It's possible that at lower RPM they identify that a longer dwell is not going to over heat the coil
- didn't catch it.

Btw, does it all have anything to do with hotter/colder plugs? Do the hotter plugs withstand higher energy and require less cooling? Then why don't they use hot plugs all the time?
Flying
Posts: 13
Joined: Fri Apr 21, 2017 8:35 am

Re: My (stupid) questions to Jared

Post by Flying »

puff wrote:Btw, does it all have anything to do with hotter/colder plugs? Do the hotter plugs withstand higher energy and require less cooling? Then why don't they use hot plugs all the time?
The heat rating of spark plugs have more to do with their ability to transfer heat from the electrode into the rest of the spark plug body.

Hot plugs are more insulative and cold plugs less so. This doesn't necessarily reflect the properties of the spark produced.
Spark plugs with a higher heat rating are preferred for daily driving OEM applications because they reduce carbon buildup due to the heat retained.

You wouldn't want to use hot plugs all the time because of the risk of inducing knock in applications that aren't as tolerant of hot spots in the combustion chamber.
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: My (stupid) questions to Jared

Post by AndreyB »

Code: Select all

	/**
+	 * On single-coil or wasted spark setups you have to lower dwell at high RPM
 	 * offset 336
 	 */
 	float sparkDwellRpmBins[DWELL_CURVE_SIZE];
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: My (stupid) questions to Jared

Post by kb1gtt »

puff wrote:Which program are you referring to?
In the picture you provided is shows a PC program with RPM and dwell. What is that program and what produce does that program work with?
Welcome to the friendlier side of internet crazy :)
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

If it's not tunedstudio with rusefi, then the second one is just a graph, made in Excel, I guess, based on data, collected with oscope...

Sent from my XT1058 using Tapatalk
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

Since the current plan is to replace my old instrument cluster with the one from rx-8, I need to do something about the fuel gauge. On the rx-8 cluster the needle is driven by stepper motor, and to control the motor, the cluster reads signals simultaneously from two sending units, which are basically potentiometers.
Supposing, I have an Arduino that knows the level of fuel in my tank and stores it as int8, with 0 meaning empty and 255 meaning full, linear characteristic.
What's the proper way of feeding the signal to the cluster?
My plan is to check the voltage from the cluster (expect it to be 5v?) Then, use some small value resistance and measure the current, but then I have no more ideas...

Sent from my XT1058 using Tapatalk
stefanst
contributor
contributor
Posts: 703
Joined: Wed Feb 17, 2016 12:24 am
Location: USA 08530

Re: My (stupid) questions to Jared

Post by stefanst »

I'm doing something similar on my Miata and I'm just using a PWM to GND output (The input into my cluster uses a 5V pull-up). It's not perfectly linear, so I have to correct for that in code, but that's fairly trivial.
You could also use a digital controlled potentiometer. They're fairly expensive though.
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

Could you please share the schematics? It could be not linear because the original tank is not linear? And I'd need two channels..
1k two channel digipot is somewhat 3$ here. But then again, I need schematics for analogue part - I doubt very much I can use digipot to directly substitute the fuel sending unit - if the pull-up to five volts in the cluster is 200ohm (how do I measure it without taking it apart?), and I set up digipot to 30ohm, the current would be 5/230=~20ma - will the digipot like it?
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

Here's the list of available digipots:
https://www.chipdip.ru/catalog/ic-digital-potentiometers
I guess, I'd need an opamp for that circuit? like our lmv324i?
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

just got an idea: if I measure the voltage on the input pin and it's 5V, if I add a 300Ohm resistance to ground - I can measure the voltage again and then calculate the upper leg of the divider?
probably, this might help when designing my own converter?

UPD: just measured those pins: 9.94V on both of them. So, it's not 5V ;-(((
UPD2: just measured resistance between the two 'upper' pins - it reads 647 ohms, which suggests that if both of these circuits are similar to each other, they are connected to 10V regulator through 320Ohm resistors?



btw, here's a couple of screenshots:
Clipboard01.jpg
Clipboard01.jpg (78.62 KiB) Viewed 16987 times
Снимок экрана 2017-05-13 в 16.15.14.png
Снимок экрана 2017-05-13 в 16.15.14.png (115.51 KiB) Viewed 16987 times
Test mode is turned on by holding down the trip reset button while turning on ignition.
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: My (stupid) questions to Jared

Post by puff »

Okay, here we are:
Снимок экрана 2017-05-13 в 23.36.29.png
Снимок экрана 2017-05-13 в 23.36.29.png (106 KiB) Viewed 16957 times
Will this work? We need 0V to 5V at output point, but it's pulled up to 10V through 320Ohms. How will this pullup affect the opamp?
Probably, the most appropriate approach would be just remove that pullup, or,even better - pull it up to 5V. But I am afraid of taking the whole thing apart.
Post Reply