MSefi.com read only forum (Jan.15,2018)

D1 , D2, D3, D4 are not needed , because the processor has internal
protection diodes .
What about D11 .
Can it be left away also ?
It looks like it has thesame function as the other four .

Pat

Well I'm no EE, the manual states that D1,2,3,4 can be left out but step 55 says to install D11.Is there some reason why you don't want to install it?

Brad J.

Well , I am an EE .
That is why I posted this question .

If the other four adc inputs need no protection , then this one probably
won't need it either .

Pat

On the other hand,it won't hurt, will it? (for all use members that installed it)

Brad J.

D11 is the input clamp for the ADC silly, needed just in-case you short out 12v heater to input of adc, that would let all the smoke out of the CPU, and everyone knows that CPU's run on smoke. I'm not EE either, but I have years of experience with letting out smoke.

S.

I left it out because i needed space to implement the active flyback mod
on my pcb .
But it's true , there is a chance of 12V on the circuit .
Didn't think of that .

Oh well , let's be extra carefull now and nothing
bad can happen .

Thx for the replys guys

Pat

A failure of the 5v regulator would be... what? If it dies, what are the chances that rather than getting 0v on that circut, you get 12v?

Just trying to rule out all possibilities.

-Brian

Usually when you design a circuit you need to pay attention to failure modes, estimate the probability that a simple miswire in the circuit would cause failure, and protect against that. Sure the likelyhood that one would accidentally connect 12V to the ADC input are slim, but nonetheless there is small chance, and thats what you are protecting for, also don't forget that it is possible for the wires to fuse together on the hot exhaust and cause a short to the 12V feed.

All the other sensor circuits are powered off 5V, so no risk of damage there, not to mention the series resistance and the built-in protection.

S.

Just got linked to this thread...

I just built and installed a Techedge WBO2 system, and have pin 4 (linear 5 volts) feeding my Megasquirt's O2 sensor input. However, the closer the output of the WBO2 gets to 4 volts (17:1 AFR) the greater the difference between the actual voltage at pin 4 and the voltage that's displayed in Megatune's Runtime.

I haven't checked this in free air, but after cranking the engine in flood clear for 10 seconds, the Techedge display reads 45L (pinned lean) and the output voltage at pin 4 is 4.6 volts. Megasquirt displays this as 3.8 volts.

Lance suggested that this circuit was designed for 0-4 volt operation, so things are starting to make sense...

I could just eliminate D11, which would probably then allow Megasquirt to accurately read the Techedge's voltage signal, but that would put my Megasquirt at risk. So what value of diode should I install at D11 to allow me to feed Megasquirt with a 5 volt O2 sensor?

The diode is intended for protection from a shorted heater, or a miswired o2 sensor, in the case you install a WBO2 like the tech edge, you don't need external protection anymore since the tech edge output can only go to 5v and no higher. Do you really intend to run your engine at the a/f ratio depicted by 4.8V output of the WBO2?. Given you probably are running gasoline for fuel, you are going to shoot for a 14.7 to 11.0 AFR, and the output of the tech edge will be roughly 2v-2.5v. so you want to make sure that the MegaSquirt is reading most accuratly towards the middle of the scale. In other words, if your WBO2 is actually putting out more than 3v I would be suprised to hear that the engine is even running.

Look here for the output of the tech edge.

http://techedge.com.au/vehicle/wbo2/wblambda.htm


S.

No offense Scott, but rather than ignore the problem I'd like it to be a non-issue.

The discrepancy between actual input voltage and displayed voltage (from Megasquirt's point of view) is larger the closer to 3.8 volts you get. That should imply (as I intend it to) that there's ALWAYS a discrepancy between measured and displayed voltage, though down around 11:1 it's very slight (which is far richer than I'd ever go).

I stated above that I was using a Techedge sensor, and I stated I was using pin 4, but I apologise for not mentioning that I was using a version 2 unit, specifically a 2A0. See the chart for pin 4 of that unit here: http://wbo2.com/2a0/2a0info.htm As you can see, the output of this pin is linear, with the voltage corresponding to (AFR-9)/2. Therefore, 14.7:1 AFR is 2.85 volts with this sensor, and 16.0:1 (peak cruising efficiency) is 3.5 volts. Once you get to that range, the difference between actual voltage and Megasquirt displayed voltage is about 1/2 volt, or 1 full point of AFR. This makes datalogging ugly, as you might imagine. I suppose I could use the SVout, but then there is no easy formula for calculating between voltage and AFR, and the inaccuracy at 2.9 volts (which is still present) would be more costly, though I don't expect to be running near 19:1 for too long.

I'll just remove the diode.

Done and done. Now my AFR displays just fine, within acceptable (to me) accuracy, and reads all the way up to 19:1.

Thanks for the help, folks.

Matt, none taken, you seem to have the problem under control, I was going to echo the leakage that Lance mentioned, but he beat me to it.

Personally, I would have removed D11 also, due to the reasons stated.

Sounds like overall good information for others to learn from.


S.

Just to confirm, all I need to do is cut one of D11's legs?

SVout is not smooth enough for my tastes, it jumps from 15.44 to 15.71 to 16.17 or some such, hard to get the target AFR's spot-on.

Clip one leg , or just desolder it and remove. If your using a WBO2 Its not required, Its just in there for protection in case of a miswire of the Heater on a NB.

Lance, what is your EGO correction factor set at? as a NB user, and possible PWC user down the road, can you explain what the difference between EGO correction, and EGO scaling is? TIA

Thanks for the fast reply, Lance! ...now to make the target AFR's work with the WBLin signal...

I mean't EGO switch point, not scaleing, Sorry for the confusion.

how 'bout that, "now that's what i'm talk'in about" !! Clap, Clap, Clap (the sound of Scotts hands coming together)

Thanks again Lance!!!

There are a couple of things that can cause an offset or noise on a wide band out signal and some things you can do help the problem.

Because you are dealing with multiple grounds, the sensor heater ground, the controller ground and the MS ground it is not unusual to get some amount of voltage offset between the wideband and MS. make sure you have a good ground for each to a common point if possible. Then you can check for voltage offset a couple of ways. If you can set the voltage out for the WB ( I read that you can do this for the Tech Edge 2.0 using the Configuration utility ) set it for a constant voltage something like 1v for both AFR=10 and AFR=20. datalog for a couple of minutes then look at the log and see what MS is seeing for voltage. Set your voltage out for a constant 4v and get another datalog. If MS is seeing 0.9v when the wideband is set at 1v and 3.8v when it is set for 4v then you would configure the wideband for 1.1v at the AFR that corresponds to 1v in MS and 4.2v at the AFR that corresponds to 4v in MS. If your WB out can't be changed just get a datalog with the sensor out of the exhaust system in free air to get a constant voltage and compensate for the voltage offset in MS. That will take care of any voltage offset.

Now for noise. Widebands normally use PWM to control the sensor heater and the Tech Edge uses PWM for the output voltage as well, this often sets up noise on the signal and ground wires causing the output to fluctuate. You can put a simple RC filter on the output to reduce this. put a 10k resistor on the WBlin out and a capacitor after the resistor between the WBlin out and the WB ground. You can start with about a 0.2uf capacitor up to 1.6uf. Don't go any larger than that or you will slow down the response time of the output too much. Once you have the noise filtered you can recheck the voltage offset.

I want to toss this out on this thread. Apparently a lot of us have had issues lately with this especially dealing with the LM-1 and LC-1. It seems that these are shipped default with the outputs programed for high impedance further throwing off the readings. To many of us, this is an obvious issue and fixed without a second thought. Its a simple matter of reprograming the outputs for low impedance.

I can build these, but sometimes electronics theory slips by me. I never realized this would cause a problem! I want to thank cmacwilly for linking over to this thread also since I never knew I had a reason to look for info on D11.

Lance, perhaps a section should be added to the MegaManual about this and setting up widebands? Or maybe a sticky containing info on D11 and high impedance outputs here.... I have seen a lot of posts recently on "whay doesn't my LC-1 read correctly?"...