Megasquirt giving misleading readings....

General Chat About Cooling & Overheating

Moderator: phpBB2 - Administrators

Post Reply
scudderfish
Site Admin
Site Admin
Posts: 486
Joined: Sat Feb 28, 2009 6:53 pm
Location: Harpenden, Hertfordshire

Megasquirt giving misleading readings....

Post by scudderfish »

For a while now, I've been occasionally seeing mad temps on my RV8, up to 118C. Towards the end of last year I spent a fair degree of time faffing around with a Wipy microcontroller and a bunch of DS18B20 temperature sensors. These seemed to agree with my suspicions, but as I had them strapped to the outside of hoses they weren't really giving me the true story. I had originally planned to put a couple of sensors directly in the radiator, but I had shied away from that in case it went wrong and I broke the rad.

It then dawned on me that I had a couple of steel pipes helpfully in the pipework on both sides of the rad, so I JBWelded a couple of sensors in. If it all went wrong I could just replace the pipes or drill out the sensors and weld the holes back up.

Image

Anyway, I put them in, topped up the coolant and went for a drive, knowing that I probably had a chunk of air still in the system. I did about 8 miles, and the CLT reading on the megasquirt did this :-

Image

Temperature goes gently up, thermostat opens and then it oscilates until it rapidly climbs to 118C. I got about 400m and then bottled it and pulled over. As suspected, a lot of air in the system and I used the bike valve to depressurise and topped off with a bottle of coolant. I got home and pullled the log file for the sensors actually in the coolant flow and they showed this :-

Image

Blue is the coolant coming out of the engine, red is the flow back from the rad, temps are in hundreths of a degree, so 8023 is 80.23C. This looks much more like I'd expect (the downward spikes are just a bit of noise in the readings) I've also highlighted on both of them 960 seconds into the log where the MS is seeing 118C where as the top hose temp is a much more reasonable 80.12C.

I've got an Offenhauser 360 manifold, and to get the sensor that came with my MS to fit, I used an adaptor. I think to length of this adapter means that the sensor is not actually in the flow and more prone to bubbles.
Image

The car seems to be running fine when the MS reports mad temps so I'm going to keep an eye on it for a while but I have more faith in the temps I'm measuring directly in the hoses.

Thoughts anyone?

Regards,
David



stevieturbo
Forum Contributor
Forum Contributor
Posts: 3979
Joined: Sat Nov 18, 2006 6:22 pm
Location: Northern Ireland

Post by stevieturbo »

Exactly what sensor are you using ?

And why cant you just screw ( or a similar one ) it into the manifold etc directly ?

ie in the engine, before the stat.

Even worst case.....the "hot" heater hose might even be an option as this will always see hot engine flow, unaffected by anything....unless the water pump breaks.
9.85 @ 144.75mph
202mph standing mile
http://www.youtube.com/watch?v=XgWRCDtiTQ0

DaveEFI
Gold Member
Gold Member
Posts: 4603
Joined: Fri Feb 26, 2010 4:27 pm
Location: SW London, UK

Post by DaveEFI »

Your warm up graph makes no sense to me. It is showing the return temperature is higher than the flow during warm up, then goes to what you'd expect later - return lower than flow. Which would suggest the calibration of the sensors is out. And I don't like those spikes.
They both appear to be showing ambient on a cold engine, though, which is a good start.

I'd be inclined to measure their resistance at known spot temperatures. Say 0, 50 and 100C. MS can be calibrated like this to an unknown sensor.

But as has been said, the MS CTS really needs to see the temperature of the cylinder heads - not rad.
Dave
London SW
Rover SD1 VDP EFI
MegaSquirt2 V3
EDIS8
Tech Edge 2Y

User avatar
Darkspeed
Guru
Guru
Posts: 913
Joined: Sun Sep 02, 2007 4:43 pm
Location: Shropshire
Contact:

Post by Darkspeed »

Graph kind of makes sense

That funny route of the return hose in the rad will have water from the returns to the pump back thermo cycling up to that sensor until the stat starts to open and rad flow is allowed to flow past it.
4.5L V8 Ginetta G27

scudderfish
Site Admin
Site Admin
Posts: 486
Joined: Sat Feb 28, 2009 6:53 pm
Location: Harpenden, Hertfordshire

Post by scudderfish »

The rad return has to go over my suspension (no space under the rocker arm because of the steering rack/column) and so whilst the thermostat is closed and there is no flow through the radiator, hot water will congregate there. Once the stat opens, water flows as expected and the temperatures do their thing. The spikes are single reading misreadings that I haven't filtered out fully. I'm running the unshielded signal lines from the sensors past my fuel injectors and coil packs, I'm surprised I get anything sensible out of them :) The sensors are DS18B20s on a Dallas OneWire interface talking to a jury-rigged microcontroller buried behind my dash (one of these https://www.pycom.io/product/wipy-expansion-board/) which logs the readings of any sensors attached to a CSV file on an SD card. They're not analogue so the MS won't talk to them.

I believe the MS sensor is fine. It was supplied with the MS by Phil from ExtraEFI. A couple of months ago I took it out and dunked it in hot water along with one of the DS18B20s and measured them within 0.5C of each other as the water cooled from near boiling down.

I think my problem is a combination of the MS sensor not being in the full flow of the coolant and it being susceptible to air/steam pockets whilst I burp the system.

DaveEFI
Gold Member
Gold Member
Posts: 4603
Joined: Fri Feb 26, 2010 4:27 pm
Location: SW London, UK

Post by DaveEFI »

Where is the MS CTS situated? Mine is in the normal place in the inlet manifold, and generally shows a more consistent temp reading than the car gauge, with its sensor beside the thermostat.
Dave
London SW
Rover SD1 VDP EFI
MegaSquirt2 V3
EDIS8
Tech Edge 2Y

scudderfish
Site Admin
Site Admin
Posts: 486
Joined: Sat Feb 28, 2009 6:53 pm
Location: Harpenden, Hertfordshire

Post by scudderfish »

You can just about see it in the last photo, the hexagonal brass thing. This is on the front of an Offenhauser 360.

Regards,
David

User avatar
Darkspeed
Guru
Guru
Posts: 913
Joined: Sun Sep 02, 2007 4:43 pm
Location: Shropshire
Contact:

Post by Darkspeed »

If you make an assumption that the M/S sensor is actually correct - what you have is a high reading of water temperature from the left hand bank - Your next reading is where the water from both heads is then mixed in the manifold out to the top hose.

You may well have limited water flow in that head - The next check I would make is to take a surface temperature reading of each head. Water will take the path of least resistance until the flow paths are balanced so if there is a restriction in the block or head on one side the flow will increase in the other until the pressure is balanced.

Then again without knowing exactly how the cooling system pipework is fully configured its just a guessing game.

I would take surface mounting or IR thermo reading on the same place on each head - or with IR take a number of spot readings on the heads and the block to see what the comparative temps are.
4.5L V8 Ginetta G27

SuperV8
Guru
Guru
Posts: 947
Joined: Tue Jul 07, 2009 11:25 pm
Location: West midlands

Post by SuperV8 »

How have you set up your earth wiring for megasquirt?
This can adversely effect sensor readings.
Where do you have the sensor earths go to?
The earths for all sensors should go back to megasquirt, not to the engine, or chassis.

Tom.
Dax Rush 4.6 supercharged V8 MSII

DaveEFI
Gold Member
Gold Member
Posts: 4603
Joined: Fri Feb 26, 2010 4:27 pm
Location: SW London, UK

Post by DaveEFI »

You could do an easy check Replace the MS CTS with a fixed resistor. Something like 180 ohms will simulate a fully hot engine, if it's the common type of sensor, and it should also run OK when warmed up. See if you get any odd readings then - it should remain the same regardless.
Dave
London SW
Rover SD1 VDP EFI
MegaSquirt2 V3
EDIS8
Tech Edge 2Y

scudderfish
Site Admin
Site Admin
Posts: 486
Joined: Sat Feb 28, 2009 6:53 pm
Location: Harpenden, Hertfordshire

Post by scudderfish »

I've been fiddling fixing some leaks and replacing a corroded thermostat housing. As I've drained the manifold, I thought I'd take the temp sensor and it's housing out and this is what I found.

Image

The sensor is completely shrouded by the adapter so I suspect I get a steam bubble caught in there and the MS says the temperature has gone through the roof when in reality it's OK.

This weekend, I'll drill and tap (or replace) the adapter so the sensor fits further into it.

Image

Post Reply

Return to “Cooling Area”