 | Home > Puma (Tdci) > Designing a digitally controlled heater system, help needed. |
 
|
|
|
| B4Lamb Member Since: 21 Mar 2015 Location: Abergavenny, Wales Posts: 233   
|
 Click image to enlarge Click image to enlarge "You are never to old to learn something new" |
||
28th Dec 2018 5:40pm |
|
| mata Member Since: 24 Jan 2014 Location: Manchester Posts: 153  
|
Edo, if your Puma/TDCI is heating the cab when set to cold then its in need of adjustment. Mine did it for a while, you need to lubricate the end of the cable near the valve and the valve linkage. |
||
|
3rd Jan 2019 11:52am |
|
| B4Lamb Member Since: 21 Mar 2015 Location: Abergavenny, Wales Posts: 233
|
Hi Mata,
Thanks for replying to my post on wiper wheelboxes. Very useful. Regarding your project on climate control... I don't really understand your need to adjust heating when off motorway driving. The reason I say this is that the water temperature flowing into the heater radiator should be roughly the same in any driving conditions as it is controlled by the engine coolant thermostat. I would be interested to understand why you think there should be a change between motorway and country road or town driving? Regarding control... as you can see from my test measurements, the open loop (ie. Set a valve position and measure the response) is, as expected a first order system. ie. Set a valve position and the cab temperature will eventually to reach stable temperature following roughly an exponential curve. In an electrical system this is equivalent, roughly, to charging a capacitor via a resistor so the time constant defines the RC value. The added complexity is the losses which in a defender is significant compared to a modern well insulated car. As you may have observed from my test results the temperature measured in the back area of my hardback did not get anywhere near the front cab temperature. This is because a) there is no direct hot air feed to the rear area behind the bulkhead and b) the rear area has huge heat losses through the roof and side panels. My defender is fitted with plywood inner side panels and has a rubber floor mat and a carpet covering as well as a ceiling head lining yet depite all this losses at the rear are significant. This cold air mass behind the bulkhead needs to be constantly fed with heat otherwise it will result in loss of temperature in the front. In electrical terms this is very roughly like a parallel resistor across the capacitor. It means that air temperature from the heater vents in the front have to run at a much higher temperature than the required temperature in the cab just to compensate for the losses in the rear which are higher than the front. A temperature control system, designed optimally, will automatically control the heater radiator valve to provide the required hot air output to achieve the desired temperature in the cab where you locate the feedback temperature sensor. Thinking aloud your pwm controlled valve will provide a proportional control of the valve and hence air temperature output. Proportional only feedback control should be necessary to to provide a fairly stable cab environmental air temperature. Some small amount of integral additive feedback may be benificial to correct any final steady state error in temperature. In retrospect having measured the response time in a real world situation I don't think differential gain feedback is required so PI only control should be adequate. If you are not going to develop a mathematical model and design the PI gain values to optimise the system then I'm sure you can get there quite quickly by initialy providing controls to adjust the gain values under test conditions to get a good response. Once you achieve these you can hard code them in. In commercial climate control systems they normally also have control of the ventilation fan speed, I'm assuming you are not planning this? The effect of increasing fan speed reduces the time constant to get to a required temperature so if the fan speed is manually adjusted this can interfere with your optimum feedback control gain settings. This is worth considering but I can't at this moment predict the impact, it may be fairly minor and possibly can be ignored. Some more tests required perhaps? Where are you going to place the measured cab temperature? Are you going to modify the existing temperature control knob for the user to set a required temperature? "You are never to old to learn something new" |
||
|
6th Jan 2019 10:32am |
|
| mata Member Since: 24 Jan 2014 Location: Manchester Posts: 153
|
Hi John, no problem, the wiper mechanism isn't anything difficult but i learned the hard way not to use inferior components and completely removing the dash is probably well worth the small amount of extra work to make the job much less painful (trapped hands, scuffed knuckles sore fingers) overall.
Regarding the heater control, I do find the heater needs adjustment depending on how the vehicles being driven. I get a noticeably hotter experience when the engine works hard. Perhaps this is my vehicle or my driving style but i do tend to try and maintain vehicle speed going up a big motorway hill and am not averse to using a lot of throttle and working the engine hard. Perhaps if i were more sympathetic the heater output would have less noticeable fluctuations. I'm not a defender connoisseur, this is my first and only one, but a friend of mine has a 2.2 tdci and reports the same experience - although he is also prone to a heavy right foot if this is a factor... Thanks for your thoughts on feedback. As i've said i'm a novice at this but have been reading up on the fundamentals on PID control and also the PID arduino library. Given that some fairly crude heater knob twiddling can regulate the cabin temperature fairly reasonably i''m not imagining that the control/feedback will need to be particularly elaborate - as you have said. Having said that, once the basic system is working some minimal code tweeking of the arduino using the PID library *should* allow powerful control if done well. Regarding fan control, yes ive given this some thought but only so far as for now i'll be getting things up and running with the fan on a steady "1". Once this achieves desired results i plan to integrate fan speed into the controller also. Once i've studied how the existing control regulates fan speed i hope to be able to achieve a setting slower than "1" as i think its quite noisy and often slower would be adequate. Automatic control of this would be getting towards the end of the project but then there's always webasto control with remote triggering via SMS! Cheers, Mat |
||
|
7th Jan 2019 2:19pm |
|
| B4Lamb Member Since: 21 Mar 2015 Location: Abergavenny, Wales Posts: 233
|
Hi Mat,
Yes I've virtually fully removed the whole facia. Who ever added the alarm system though routed cables through places that full removal would involve cutting and connectorising or rejoining. Idiots. Anyway I've got it away far enough to work with ease with the upper ducting removed. Unfortunately I'd ordered the wheel box kit before you answered my query and the are all bearmach parts. They look well made to me. The existing outer tubes look perfect so I'm not going to get new ones of these, I'll just liberally grease the new drive cable on reassembly. The N'S wheelbox was seized fully and had stripped the the coiled drive wire at that position so it was a right mess. Anyway just have to put the lot back together now. Re Heater ie. Main subject, I'm thinking the reason for the heat difference between hard driving and slower is that although the engine coolant temperature should be about the same the flow rate could be a lot higher if the thermostat is having to open wider to maintain the same temperature under heavier loads. A higher flow through the heater matrix means the input to output temperature differential will be less meaning a higher air temperature for the same air flow rate . This is just a guess but the nearest reason I can come up with. Upping the fan speed when there is a larger error between desired temperature and measured temperature would speed up the time to get to the final value, it effectively reduces the time constant so optimal PID Gain settings for a particular fan setting will not be optimal for different settings. It may well be in the noise but it may need to he taken into account. When I get my landy back together ill repeat the measurements I did but with full fan speed. I'll try and pick a day when the outside ambient is the same or similar to last time. I think Fan speed is adjusted by selecting different resistor values in series with the motor, a very inefficient control method but common on all older designed cars. Ill take a look tomorrow. A digital switched PWM with MOSFET driver is likely to be the best speed control method but you will need to switch above the audible range or heavily filter the output so it doesn't buzz like a Banshee. Anyway I'm sure you can do it all from one small microcontroller, it will have  all else to do.
Cheers, John "You are never to old to learn something new" |
||
|
7th Jan 2019 11:12pm |
|
| mata Member Since: 24 Jan 2014 Location: Manchester Posts: 153
|
I was thinking PWM fan control would be the way to go, so simple with an arduino, but that's for another time.
Prototype kind of finished and ready to start testing in the defender but still waiting for the small piece of pipework to plumb up the valve! Click image to enlarge |
||
|
8th Jan 2019 11:12am |
|
| B4Lamb Member Since: 21 Mar 2015 Location: Abergavenny, Wales Posts: 233
|
Ideal.
Where will you be measuring temperature in the cab? How do you set the desired temperature?, are you modifying the existing dial that moves the old valve cable? I take it -4 is not the actual temperature in your house haha Is that valve sprung to normally closed for the heater matrix? How much current is required to open the valve? Sorry lots of questions  "You are never to old to learn something new" |
||
|
8th Jan 2019 6:59pm |
|
| mata Member Since: 24 Jan 2014 Location: Manchester Posts: 153
|
This is the proof of concept, no temperature measurement / control yet, just ten settings 0 - 9 varying the pulse width. Full hot = valve always open, full cold = valve always shut, half way = valve open for 1.5s then valve shut for 1.5s.
up and down buttons change setting from 0 to 9, second line on the lcd shows the valve opening and closing for debugging valve is open (full hot) with no current on soleniod solenoid draws 0.6 amps with 13.5v across it. As its a car part theres no available datasheet but i might try and get a meter on one in situ in its intnded vehicle to see if theirs any post-activation current limiting going on. I saw a scope trace someone published that showed this wasn't the case but i'd like to verify. 0.6 amps at 13.5 v is a fair bit of wasted power and the solenoid heats up to (and stabilizes at) a toasty 75 degrees celcius when left on for a few hours. This is inside the house (19 degrees ambient), im not sure how being under a hot bonnet will increase this...  |
||
|
8th Jan 2019 7:28pm |
|
| B4Lamb Member Since: 21 Mar 2015 Location: Abergavenny, Wales Posts: 233
|
Hi Mat,
Okay I understand, good plan for the first test outing. Pity it has to be powered to keep it off as that will be in the summer when its hotter under the bonnet. If you have measured 75C as the final steady state temperature it reaches indoors then for every degree the final environment location is above your home room temperature is it what it will increase by, so if your room is 19C and it is say 40C under the bonnet then it will get to 96C. It's basic physics. Don't forget when 'off' one side of the valve will be in contact with the engine coolant which will be around 80 to 90C. Your room test does not account for that heat flow input. If it's designed for this purpose though I'm sure it will be perfectly fine. It's only consuming 8 watts according to your data so I'm surprised it gets that hot. I think it will get to and stay close to the coolant remperature anyway whether 'on' or 'off' as the water will be a good thermal conduction path. I dont know the construction but if it's steel or aluminium it will conduct some heat through the mounting. If it's stainless then it will not conduct very well as SS is a poor conductor. Once shut, you may be able to back off the current and still assure it is still shut however you would need a switch mode inverter driver to capitalise on the energy saving as a simple linear regulator or resistor in series will just waste the power at the driver instead of the valve. I can't see that it will current limit much. The solenoid coil will heat a bit when you apply current and it's resistance will rise a bit which will result in a slight drop in current but I cannot see this will be much. John "You are never to old to learn something new" |
||
|
8th Jan 2019 9:09pm |
|
| mata Member Since: 24 Jan 2014 Location: Manchester Posts: 153
|
Valve body material is some type of plastic. The cover over the solenoid is zinc coated steel and this is the bit that was eating up to 75 decrees. I don’t think it’s an issue but it was a suprise it got so hot.
Installed the valve today, no cutting or modification to existing pipes was needed, I left the original valve connected to the cable and tucked it out the way. The new valve has plenty of space and is supported pretty well by the heating pipes for now. I think I’ll remove the old valves bracket and make a new one for the new valve in due time. Went for a short test drive and had control over the temperature of the air out the vents. I plan tomorrow to go for a longer drive and get the car fully up to temperature and take some readings with some steady driving to see what kind of ‘resolution’ I have varying the duty cycle. Originally I had off, on and 8 settings in between but I’ve upped this to 18. All being well I’ll get a temperature sensor ordered and start working on some PID code to begin getting things automated. |
||
|
11th Jan 2019 8:05pm |
|
| B4Lamb Member Since: 21 Mar 2015 Location: Abergavenny, Wales Posts: 233
|
It all sounds good Mat. Do you have a data logger to record the temperature measurements on your tests?
For you next stage I would recommend the use of DS18B20 digital temperature sensors. You can buy pre terminated sensors on eBay quite cheap with a range of cable lengths. Or you can just buy the 3 pin chips that come in the form of a plastic signal transistor. John. "You are never to old to learn something new" |
||
|
11th Jan 2019 8:31pm |
|
| mata Member Since: 24 Jan 2014 Location: Manchester Posts: 153
|
Data logger.... not quite, more a meat thermometer, pad and pen!
That temp sensor is pretty similar to the LM35DZ I was considering. Regarding your question about sensor position I’m hoping that on the dashboard where the existing heating controls are will work. It appears to be out of any direct airflow from the vents so that is my aim.. |
||
|
12th Jan 2019 11:24am |
|
| dgardel Member Since: 30 Nov 2008 Location: Veneto (Heart & Head) Posts: 3586 
|
Remember that premium cars has a "flap" inside the cabin heater that mix cold with hot air for a quick temp modulation.
Puma Defender doesn't have the flap, so temp modulation it's done only by the 3 ways valve and this implies long modulation times with risks of the oscillation effects between hot and cold Discovery 5 td6 HSE Stornoway Gray Outback Engineering Limited Edition
IID Pro MV License |
||
|
12th Jan 2019 11:40am |
|
| dgardel Member Since: 30 Nov 2008 Location: Veneto (Heart & Head) Posts: 3586
|
I would add a frequency modulation of the fan motor so as to have a continuous regulation of the fan speed to help Discovery 5 td6 HSE Stornoway Gray Outback Engineering Limited Edition
IID Pro MV License |
||
|
12th Jan 2019 11:41am |
|
|
|
| All times are GMT + 1 Hour |
< Previous Topic | Next Topic > |
Posting Rules
|
Site Copyright © 2006-2025 Futuranet Ltd & Martin Lewis
