another one for the MSQ gurus

r2d2hp · Post by **r2d2hp** » Mon Oct 08, 2012 11:53 am

Sensors use a regulated 5v supply

This shows the circuit
http://msextra.com/doc/general/ms3v3schems.html#ms3x

The sensor is in series with R7 and the voltage differential is fed to AD2-1 is via R8 (2.2K).

Being your sensor has a very high resistance the voltage seen by the input to the A>D will be at top end of the scale which could have an effect on accuracy

Volts Temp
4.975949577 -9
4.898530441 13
3.505154639 90

Not sure why the instructions advise to change the bias resistor, one for Phil to answer

kiwicar · Post by **kiwicar** » Mon Oct 08, 2012 12:22 pm

Hi Reg
So it does. . . Oh! not a voltage I would have chosen, especially in a noisy environment and especially using high impedence inputs

At least I don't have to worry about a stable battery supply.
Best regards
Mike

DaveEFI · Post by **DaveEFI** » Mon Oct 08, 2012 12:52 pm

The 5 volt supply is used for lots of things on the MS - why add another?
Also, if you want a regulated supply off a car battery in near any state, it's the obvious 'standard' to use.

kiwicar · Post by **kiwicar** » Mon Oct 08, 2012 1:14 pm

Hi Dave
Just for noise reduction, you want as many volts as you can feeding a decent sized current into a low inpedence sink, like the classic 600 ohms sink of a classic BT transmission line. . . my preferance on an automotive environment would have been stabalised 10v referance with a charictoristic impedence around 2k looking back into the source and driving into a 1k load give you 5v at the mid point of the range and 5 mA into the load you would need some serious noise to overwhelm that!
Best regards
Mike

stevieturbo · Post by **stevieturbo** » Mon Oct 08, 2012 3:58 pm

Why on earth does MS make such a simple task as calibrating a sensor so difficult ?

And only 3 points for calibration ? It will never be accurate over a wide range as most sensors are not linear.

It may be a budget ecu, but that really wouldn't inspire confidence.

kiwicar · Post by **kiwicar** » Mon Oct 08, 2012 4:11 pm

stevieturbo wrote:Why on earth does MS make such a simple task as calibrating a sensor so difficult ?

Hi Stevie

I'm not sure I can answer that but I do find the shear quantity of information needed to do the simplest of things with it more than a little challenging. . . I have found a refernce to "output test mode" in the manuals which looks really usefull for checking out the ignition.. Trouble is I have spent 4 hours since yesterday afternoon looking through the menus and the online manuals and I can't for hte life of me find out how to access it, you would think it would come under "tools" as it is a Testing tool er not there

. but you would also think "battery calibration" would ask you to measure the battery voltage with a meter and enter that, no chance!
Best regards
Mike

DaveEFI · Post by **DaveEFI** » Mon Oct 08, 2012 4:16 pm

kiwicar wrote:Hi Dave
Just for noise reduction, you want as many volts as you can feeding a decent sized current into a low inpedence sink, like the classic 600 ohms sink of a classic BT transmission line. . . my preferance on an automotive environment would have been stabalised 10v referance with a charictoristic impedence around 2k looking back into the source and driving into a 1k load give you 5v at the mid point of the range and 5 mA into the load you would need some serious noise to overwhelm that!
Best regards
Mike

One obvious reason is the battery volts can drop below 10 when cranking on a cold day. And a regulator needs more than its nominal volts to regulate. Of course you could go to a SMPS and produce a regulated voltage at anything you want - but really why go to the bother. 5 volts is plenty with proper cabling and connectors. Even Lucas used it on the original injection.

BTW, as I said, you do get some serious noise if you use the block as a return. The starter motor current flows through it. As does the alternator output. And spark plug returns. Just have a look with a scope some day.

FWIW, if you read the MS forum, vast numbers of problems are caused by not following the MS wiring diagram and thinking grounding doesn't matter. And using a dirty supply.

My advice is to do the job correctly. And save tears later on.

stevieturbo · Post by **stevieturbo** » Mon Oct 08, 2012 4:17 pm

Can't think why you'd ever need to tell the ecu battery voltage. If its turned on, its always connected to battery voltage so it should know without you telling it.

And as above, proper grounding and isolating wires from spurious high voltage signals is very important with any ecu. Some are fussier than others though. Do it right and you can at least rule out potential problems.

Power grounds will often originate at the engine block. Sensor grounds should be isolated and originate from the ecu only.

DaveEFI · Post by **DaveEFI** » Mon Oct 08, 2012 4:24 pm

stevieturbo wrote:Why on earth does MS make such a simple task as calibrating a sensor so difficult ?

And only 3 points for calibration ? It will never be accurate over a wide range as most sensors are not linear.

It may be a budget ecu, but that really wouldn't inspire confidence.

How wide a range to you want it to be 'accurate' over? Most would be happy with the minimum and maximum likely to be encountered. And the software will be aware of the characteristics of a thermistor, and interpolate the correct curve from the figures entered. It's hardly rocket science.

r2d2hp · Post by **r2d2hp** » Mon Oct 08, 2012 4:42 pm

As Stevie said the sensors are not linear so it needs more than 2 points.

Thermistor characteristics can be calculated using a formula that can be found here.

http://thermistor.sourceforge.net/

This is probably built into the ECU and used by setting of the 3 parameters, well it is on mine. You don’t need more that 3 to do this but it would help

Hi Mike, MS is for people who want to learn how everything works. It’s an educational system and Phil clearly says this on his website probably for legal reasons. There are plenty of other systems out there where it’s a damm site easier to get the things working but don’t give you flexibility/complexity of MS.

The sensor grounds go back to the ECU but the ECU gets its ground from the Block as stated here >> http://msextra.com/doc/general/grounding.html. This is to prevent multiple earth loops but don’t see how this isolates it from any noise that’s seen with a scope on the block

Quagmire · Post by **Quagmire** » Tue Oct 09, 2012 8:45 am

To answer questions of accuracy- Easytherm (used to generate resistance curves in ms1) uses the steinhart-hart equation to generate the non-linear curve for the measured three values and so should be nearly spot on. This was (as far as I know) incorporated directly into later Ms versions, removing the need for easytherm and simplifying the process.

DaveEFI · Post by **DaveEFI** » Tue Oct 09, 2012 9:24 am

The correct star ground point is the MS itself. Same as any electronics. In practice, this means its connector. The ground lead from the MS to the block is carrying relatively high currents - injector grounds - so it is best to ground the sensors at the MS connector and avoid those currents.

In practice it may not make a difference - but why risk it doing so?

kiwicar · Post by **kiwicar** » Tue Oct 09, 2012 9:34 am

Hi Reg,
that is how I saw it in the diagrams, using the block as a star point and taking everything (including the ECU Earth) from there. As you know I have fed the coil supply and the ECU separately after the main isolator switch and Kept the Starter solenoid on a separate feed. Having no alternator on there and a separate power feed for the starter the only way I can see of keeping noise any lower is to avoid earth loops and I will try running a separate earth to a 2 wire temperature sensor.
I appreciate the "educational" aspect to the MSQ, I have been following the forums on an off for about 10 years, and I built a version2 about 7 years ago, what I hadn't taken account of was the sheer quantity of options on the V3 set up and the proliferation of information but how very little structure to how that information is accessed.
I spent some time trying to find a genuine alternative, I wanted mapped wasted spark using 4 coils and a trigger wheel and Tachometer drive, I wanted a 2 stage rev limiter, I wanted data logging and I wanted to be able to activate various solenoids off engine revs and ancillary triggers. There are actually very few ignition options that will give you this in one box, yes your emerald might , but I think it may struggle to do all the above. The other approach would be to buy separate ignition drive, logging and controller boxes, but then I would be into a world of woe trying to get everything to “talk” to everything else, this is a lot of what I do at work and it is a real pain.
If I could get t working it would be highly educational to me being able to log and tune the fuel and ignition (and form my early reading of MSQ this was the intended educational area) I don’t think it is educational to be struggling with badly ordered information and multiple sources trying to get the thing set up and running.
Best regards
Mike

DaveEFI · Post by **DaveEFI** » Tue Oct 09, 2012 9:38 am

r2d2hp wrote:
[snip]

The sensor grounds go back to the ECU but the ECU gets its ground from the Block as stated here >> http://msextra.com/doc/general/grounding.html. This is to prevent multiple earth loops but don’t see how this isolates it from any noise that’s seen with a scope on the block

The sensors run from the 5 volt PS which has its own smoothing/filtering.
Keeping the grounds separate from high current devices (apart from at the star grounding point) means there are no such currents flowing in them. For a current flow, you need both ends of a circuit connected.

It's exactly the same principle as audio with unbalanced connectors - ie most domestic stuff. The mains ground (if there is one) is only applied at one point, even although all the interconnect grounds are physically connected to it. Usually the earth is at the amplifier. Add another mains earth connected to the screens at say a CD player, and you'll get a ground loop which results in hum. That is tiny current flow along the interconnect screens.

Ian Anderson · Post by **Ian Anderson** » Tue Oct 09, 2012 11:56 am

Mike

Don't you sometimes wish you just used a leaky bucket?

Ian