Jump to content
Sign in to follow this  
wrs

E36 Heater Conversion to Electric

Recommended Posts

Converting my heater over to electric started from the failure of the central locking system and electric windows on my coupe. During the diagnostics for the electrics I discovered 3 water leaks contributing to the water under the carpet that corroded out the +12V junction under the passenger seat.

1. One of the floor bungs had popped out allowing road water to enter when driving in the wet.

2. The seal between the heater and firewall had failed + the drain for the air entry into the heater was blocked.

3. My heater core is leaking (but only just).

Because the heater solenoids are also stuffed with one side is stuck on hot and the other cold I decided rather than try to fix / replace everything why not convert the heater to electric. The BMW solenoid spare part is around $650+GST... My battery is in the boot with the main 70mm² cable running under the drivers seat. Tomorrow I'm going to cut this cable and re-join it with a barrel hex-crimp + fit 2 x 6mm² cables in prior to crimping. These 2 cables will be double sleeved to make sure there's no chance of a short to chassis. They will supply the bulk power to 2 new electric heater elements.

I purchased a complete second hand heater from the wreckers. I also used this as an excuse to pull the dashboard out of the donor car so I could see how it's done and what old brittle plastic breaks in the process. It's remarkably easy, just time consuming. I've now stripped down the heater unit and removed the air-conditioning core and the heater core. My air-conditioning compressor died about 10 years ago and de-gassed itself in the process so I removed it back then - not worth fixing. When I did the engine upgrade a while ago I ripped out all the remaining air-conditioning garbage from the front of the car - got rid of around 60kg.

I've purchased 2 x 300W ceramic PTC heaters from Ebay. These should draw about 21-23A at 14V since they operate almost like constant power devices. This means a total load of 42-46A on the alternator - hopefully it can cope. If not then I'll upgrade it. The new heaters should arrive later this week.

I'm going to fit the new heater elements to an aluminum plate that will be glued into the position the old heater core lived. There's 3 openings that also need to be blocked up; where the air-conditioning pipes entered, where the coolant pipes entered through the firewall and where the coolant pipes enter the heater box.

Now for the controls for the heater elements. The heater element will not be allowed to run unless the engine is running and the battery voltage is above 13.5V. This will turn on 2 relays that supply 13.5-14.4V to one end of the heater elements. The heater element on-off control will use the original water solenoid control signals. This will operate the left and right heater elements independently just like it used to. I'll construct a simple dual switch using mosfets which is enabled by the heater solenoid signals. There will be a temperature sensor on each heater element to limit the element to 95'C (similar to coolant temperature). The mosfets will switch on and off to keep the elements at 95'C. There will be a secondary sensor on each element to switch it off if it reaches 110'C in case the primary fails (drop out the main relays). The mosfets will pull the other end of the heater element to ground.

The climate control system is expecting the heater core to run at about 90-95'C (coolant temperature) so I'm hoping setting the elements to this temperature means the temperature regulation will still work in a similar way to original. I suspect the 300W heater elements won't get to full temperature most of the time though. The airflow will likely be much higher than required to keep them below 95'C. It's also likely it won't create anywhere as much heat as the old heater core could have as 600W is not much power. It will be instant heat though as soon as the engine is running and it will still be enough to heat the small space - it just might take a minute or two longer to warm up.

I'll take a bunch of photo's once the heater elements arrive and start fitting everything to the donor heater case. I'll also include the fairly basic circuit for the control of the elements once this is knocked together.

Given there's a lot of E36 out there with failed heater solenoids and heater cores just waiting to leak (or already leaking) this mod might appeal to some people. The cost in dollars to do the conversion is relatively minor. The time is not. I'm guessing around 4 days to build the new heater, strip everything out of the car, bypass the heater pipework, fit the new heater and put everything back together.

My car is already stripped out except for the dashboard - no seats (except drivers seat), carpet, glovebox or console - about a days work so far to strip everything out, clean it up and rust-kill a couple of surface rust areas that have been sitting wet for a while.

  • Like 3

Share this post


Link to post
Share on other sites

The new cables are in.

New Hex Crimp Join + 2 x 6mm² Cables + Glue-Lined Heatshrink. The cables are protected by some 16mm beta-flex tubing.

33kgdoj.jpg

Bung Holes Sealed + Surface Rust Killed & Painted

25817r7.jpg

Cover Refitted

11jcj74.jpg

I've started work on the control board. The circuit is attached as a PDF. Decided to use a small 8 bit Arduino based microcontroller as it makes the board really simple. I'll complete the PCB design in the next couple of days then order from China - should be here by next weekend.

BMW Heater ControllerRev A Schematic.pdf

Share this post


Link to post
Share on other sites

A little more progress today:

Heater PCB Designed and ordered from China.

246vuxv.jpg

My alternator appears to be 140A so should not need to be upgraded.

30xaemf.jpg

New heater cores arrived today

20h4qpu.jpg

Will start fitting up the cores tomorrow.

Share this post


Link to post
Share on other sites

What software did you draw up the pcbs in. Interested to see how much current it actually draws. 
 

 

Share this post


Link to post
Share on other sites

Protel99se - been using the Protel products for 25+ years. Haven't bothered to update it as it does everything I need it to do. I've even done an A3 sized 8-layer board and it coped ok.

I did a couple of load checks on the heater element last night and it appear to initially have a negative temperature coefficient. I was expecting this but it almost looks like the minimum resistance is going to be about where my operating point is. The label on the packaging says the Curie temperature is around 200'C which means it likely to be negative up to just before then. I didn't have enough current available from my power supply to see what happens up round 14V. I've also designed the control round using semiconductor temperature sensors so 125'C max on the element. If I change to PT100's or similar I could let the element get hotter.

The net result might be more power than expected. That wouldn't be too bad as long as the current isn't too high.

The PCB is almost complete at the manufacturer (up to silk-screen stage) - may even ship today since it's only 1:30pm in Hong Kong!!

Share this post


Link to post
Share on other sites

I might be wrong but my understanding is that the heater solenoids in the e36 cycle  open and closed for varying lengths of time depending on the temperature selction ( in the analogue a/c, probably a temp sensor switched thing in automatic versions) which will mean you will have to leave the temp set to max on the original controls to have a constant signal at the old solenoid plug. Would you just be able to run arelays directly from those signal wires to power the heaters and let the original system switch them on /off?

Share this post


Link to post
Share on other sites

Thanks - had already considered that. I was counting on this as part of the control. I'm only wrapping a temperature limit control loop round the heater core temperature so stop it hitting 200'C+.

It's also why I'm switching the main control current with mosfets as a relay would wear out quite quickly being modulated.

Inside the heater there are 2 temperature sensors, one for each side. Looking at the guts of the control board these sensors appear to feed back the air temperature after the heater cores to the uP. I suspect the water solenoids are modulated to control the air temperature coming out of the heater to match the user control setpoint. My only concern is if they've implemented a PI or PID loop to control the solenoids based on the time constant of the heater energy/mass which I stuff up by putting in a heater core with a much faster response due to much smaller mass. Worst case I can try to emulate the original heater time constant in software...

That was also another reason to do the control loops in software vs opamps and comparators - easy to change.

 

Edited by wrs

Share this post


Link to post
Share on other sites

The PCB arrived today. The parts I didn't have in my left-overs box aren't going to arrive until next week...

jsccup.jpg

The new elements are mounted. Now the wires need to be properly terminated and a baffle added for the cables to pass through.

2e55zlx.jpg

Share this post


Link to post
Share on other sites

Where do you get heater valves for $280-300? The local BMW shop quoted $650+GST.

I'm not charging the battery then taking the energy out - the heater won't turn on until the battery is above 13.5V which means the alternator is active. All the energy for the heater comes from the alternator which is about 95%+ efficient. The total fossil fuel to heat conversion is more likely 30% - still pretty rubbish + isn't using the already waste heat from the engine as you say.

However, the project is in part about just doing a conversion with the added bonus of instant heat once the engine is running + no risk of a water leak from the heater core ever again.

My solar PV on the house more than makes up for the fossil fuel use the electric heater element will use...

Thinking about it though, the heater is actually 0% efficient overall. It makes heat which leaks out of the car so it's a total loss either way it's done.

  • Like 1

Share this post


Link to post
Share on other sites

Can't understand how different dealerships can have such differences (>200%) in price...

Yup, waste engine heat conversion into heat inside the cabin is close to 100% but then it all leaks out so effectively 0%. Otherwise you could turn your heater on for 5 minutes and then turn it off again and you'd stay warm for ever...

Yup again to the alternator but the conversion of fossil fuel to rotational energy is around 30% then the alternator 95% so close to 30% effective efficiency conversion from fossil fuel to heat (not 10%)...

We're also only taking about 0.6kW of energy, it's bugger-all.

It's a project being done for the fun of it - I don't care if it's not particularly efficient.

  • Like 2

Share this post


Link to post
Share on other sites

The PCB is now partly loaded with the parts I had in stock. The remaining bits won't be here until mid next week.

At least the software testing can begin...

w6v5hj.jpg

  • Like 1

Share this post


Link to post
Share on other sites

It's been a while - had to work over labour weekend so no progress for 2 weeks. I now have a week off work as days in lieu... Since HB day was on the Friday I got 4 days so took a days annual as well.

The only thing left to do is test the software. I'll start this tomorrow and see how badly wrong it is. Following is some progress photo's. The new PCB fits where the old aluminum water pipes came out and headed towards the firewall. I've bunged up the two openings where the water pipes and aircon pipes came through. Where the old heater core used to sit also needed a plate fitted which worked out well because it's also where the cables now come out. A plate was fabricated to fit and then holes drilled for the cable entry. The cables that go out to the water solenoid on X35 and X10054 have been clipped off near the connectors and joined into the control cables for the new board. This means the original solenoid plug can be cut off or just tied off (at least it's not powered now).

Side view showing how the board fits where the pipes came out.

2z7048m.jpg

Original Plugs and new main power relays

24ew8c5.jpg

New Power Relays - the main power connection goes onto the unused tabs.

24y8thl.jpg

Another view, looking from the front.

300zbtg.jpg

Heater cores with all the wiring installed. The blue cables are shielded signal types which connect to the LT9701 temperature sensors clamped under the rectangular blocks of ally.

294t840.jpg

 

  • Like 4

Share this post


Link to post
Share on other sites

Interesting stuff, my initial thoughts (which you've almost certainly covered) are around worst case failures, will a failed mosfet cause a fuse to trip or will the PTC nature of it means it just applies full heat and could start melting your dash from inside out if the fan isn't going?

And can this be easily adapted for the EV guys? series up the elements so it can run at 72V or 144V?

Share this post


Link to post
Share on other sites
2 hours ago, 3pedals said:

Epic fail in the logic and fact departments - but enjoy your project 

Eh?? I don't think you understand basic physics...

Do you use your aircon for cooling - this is not really any better than electric heating? The aircon uses rotational energy just the same as the alternator... I don't really care what you think anyway. It's installed now so no going back...

  • Like 1

Share this post


Link to post
Share on other sites
49 minutes ago, tawa said:

Interesting stuff, my initial thoughts (which you've almost certainly covered) are around worst case failures, will a failed mosfet cause a fuse to trip or will the PTC nature of it means it just applies full heat and could start melting your dash from inside out if the fan isn't going?

And can this be easily adapted for the EV guys? series up the elements so it can run at 72V or 144V?

It would be easy to modify for EV and higher voltages. The PTC elements are available in different voltage ranges. They don't series well being a PTC - the first one to go higher in resistance gets it. The FET's used in this design are only 60V rated but really low ESR. Higher voltage won't need as much current so using a higher voltage FET's with a higher ESR will give similar losses. The biggest problem with the current PCB is the voltage clearance between the tracks. For higher voltage you'd really want to use bigger clearances. For high reliability DC circuits I always use 'reinforced' creapage and clearance rules (4x the standard).

Dead FET's are detected in the control logic and the relays are opened (a fault LED comes on). There's also redundant temperature sensors which will open the relays. The PTC elements will also max out at 210'C and power limit. Not that I'd want this temperature sitting in the plastic heater box - most likely would start melting things. 3 levels of safety should make it safe.

It's all tested and working ok. The only problem I have is the control uP is active all the time, even with the ignition off. For now I've pulled the power plug from the PCB. It appears the 12V supply I've used from the heater solenoid is not switched. I'll have to find a switched source that comes on with the heater power. I assumed the heater solenoid power would be switched with the ignition - should have checked it first. This is the only error though. The heater control logic is working perfectly and heat comes out the vents. The heat is nowhere near as strong as the original water based radiator but I'd expected that. It's hard to tell now when it's already warm if it will be enough heat in winter. It's going to be much better than it was though - it didn't used to work at all.

I can't check the current draw because I don't have a DC clamp meter anymore. I'll have to borrow one from work to test what the power draw is. I suspect it will be a bit more than 300W each side because the PTC elements actually have an initial negative coefficient so the wattage goes up until they hit about 160'C and then the resistance rapidly starts to rise. Since I'm operating them at 95'C the resistance is about 20% lower so I'm likely drawing 360W+...

  • Like 1

Share this post


Link to post
Share on other sites
3 hours ago, 3pedals said:

Epic fail in the logic and fact departments - but enjoy your project 

Can you stop arguing on every single thread? People wont want to post projects if they know they will be attacked.

  • Like 3

Share this post


Link to post
Share on other sites

Here's a wrap-up of what I learned during the software testing and then in-car hardware testing.

The PCB worked perfectly first time so no issues. The power supply was originally going to come from the heater solenoid +12V but this line is always live even with the ignition switched off. The power on this line also gets switched off when the drivers-side temperature selector is wound to maximum. The next issue which was minor was the logic levels for the solenoid on/off control were inverted. I'd assumed they were pulled low to activate the solenoid to let water flow into the heater core. However, the heater solenoid valves are always open and close when the solenoid is activated. This means the solenoid drive signals go low to turn off the heater instead of on. This was an easy fix - the advantage of doing all the control in software, the input on-state level was inverted.

I ended up taking the 12V supply for the new PCB from the heater fan switch. The easiest access to this was on the resistor block that plugs into the back of the heater box. There's 4 wires for the 1,2,3,4 values on the heater fan speed selector switch. I added a diode from each wire to a common point so no matter what the switch setting was there's always 12V. Now the heater controls only come on with the fan switch and car ignition also has to be on first. If the engine isn't running the main power relays don't pull in because the battery voltage is too low.

At idle only one heater can be used as the battery voltage sags too much due to the alternator not producing enough output at low RPM. From 1500RPM both can be run. You can certainly hear the alternator loading up though - will have to see if it toughs it out. Usually the heater will only get used if it's cold so the alternator will also get colder cooling air so hopefully it won't be a problem. I couldn't find much information about the alternator and what it's continuous rating is.

If the voltage gets too low the FET's switch off and the relays open. There's then a 15 second time delay before the relays are allowed to close again and FET's turn back on. This is to stop rapid cycling of the relays.

I did some more tests today to see how much heat is generated. With the fan on 3 there was a noticeable difference in the air temperature coming from the vents. It might be better than previously stated. The real test will be how long it takes to warm up the cabin in winter.

 

Edited by wrs
  • Like 1

Share this post


Link to post
Share on other sites

Could you look at a bigger alternator - I know some beemers had some pretty grunty ones?

Also could you change all your lights to led to lower the load when driving on a cold night?

  • Like 2

Share this post


Link to post
Share on other sites

I already have a 140A alternator which is the biggest one available as a standard part for the E36. To go bigger might require a custom fit. The electric elements should use about 60A which is well within the 140A rating of the alternator. What I can't find out is if the 140A rating is continuous or short-term for peak loads with a somewhat lower continuous rating. With the heater running + headlights on high-beam + heater fan + other minor loads there could be close to 100A required. Knowing the continuous rating of the alternator would help give some certainty it will be ok or could fail.

LED lights is a good option as long as they have the gadget to trick the blown bulb detection circuit. Hadn't thought of this as an option to lower power use...

  • Like 1

Share this post


Link to post
Share on other sites

Here's a couple of shots of the heater fitted in-car. The two red wires in a loop need an in-line fuse fitted as they're direct wired to the battery, currently with no protection. Ok for testing but not running live. Once fuses are fitted they'll be tied near the back of the loom and accessible when the kick panel below the glovebox is removed. The side of the glovebox clears the tallest parts on the PCB by about 5mm. It all just squeezes in where the old water pipes used to go.

2z7igqb.jpg

2eoyvja.jpg

  • Like 1

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
Sign in to follow this  

×
×
  • Create New...