Process Control of Car Cooling System

Engine coolant temperature has as many variables as there are keys on your keyboard. Ambient air temp, under hood air temp, humidity, radiator size, radiator construction, radiator color, radiator location, radiator inlet and outlet size, radiator design, coolant used, coolant routing, engine design, thermostat setting, cooling system corrosion, heater size, heater operation, air conditioning presence, heat load on the air conditioning system, vehicle speed, air flow through the radiator, grill design sealing between radiator and radiator support-hood, water pump design , water pump speed, cooling fan(s) size, sealing between fan(s) and radiator, fan shroud design, sealing between shroud and radiator, engine size, engine design, engine speed, engine efficiency, engine tune, engine load, transmission design, transmission efficiency, torque converter design, torque converter efficiency, transmission cooler design, transmission cooler efficiency, Tire size, axle ratio, tire air pressure, body shape, driving style, road surface, road elevation, road incline, just off the top of my head. If you can quantify these variables, we might be able to answer your question. Good luck, and happy motoring.

I think this is the subject of the question.

My opinion is that it doesn't matter.

Nice ride!

Lyn, you've been checking up on me! Thanks for the compliment. 20 years of tinkering and my kids inheritance. It was worth it.

Obviously a labor of love.

OH! Well, that clears things up some. To the list you can add manifold pressure, and fuel mixture variances. You may also want to consider the location you are picking up the coolant temperature at. Your SBC should have three different locations for temp sensors. Between cylinders 1 & 3 or 6 & 8, in the cylinder heads, and in the intake manifold front water passage. The cylinder heads tend to be hotter than the manifold, and the rear location(6&8) tend to be hotter than the front (1&3). But, the best location for your use, (in my opinion) would be the temperature of the water going into the water pump after it has been cooled by the radiator, or whatever else is cooling it. If your coolant is leaving the manifold at 195, but returning at 160, there is no need to turn on the fans. But if the coolant is leaving at 195, and only cooling to 190, then you need the fans on.

I would suggest that you monitor coolant at as many places as you can until you are satisfied that the cooling system is protecting your Cavalier sorry, roadster. There are heat crayons that have specific melt temperatures. If that was in my driveway, I would rub a bit of different crayons on the heads between each pair of the exhaust ports, as well as the points where the coolant heads to the back, and the water pump inlet. If the engine temp gauge shows that your coolant is below 240 at all times, but one of the cylinder heads is showing that it melts a 2 60 crayon, I would think that you would want to drop that coolant set point some to keep that head temp lower. BTW, if you need to store that somewhere for the winter, let me know always room for it in my driveway.

Bob, I can watch the temp at 5 places now! The fan sensor is in the block between 6 and 8, there are Moon gauges each side of the intake manifold, the dash gauge sensor is between 1 and 3, and the EFI sensor is in the radiator.

Cooling was marginal when I first built the car due to the low coolant flow rate (my electric pump is a marine bilge pump, about all there was 20 years ago), but I studied up on coolants and system pressure, and now it runs year round without problem, even when I lived in Death Valley, TX. Here's a white paper some wise guy wrote about it: https://tbucketeers.com/threads/cooling-secrets.12176/

OK, so you have some fabricating and electrical skills that may not be commonplace. Most of what I wrote are from your history book I see. I have digested your posts on Bucketeers, and see where you have assured all of the temp sensors are quite close. Do you see any differences in the temperatures between the two heads? Based on the reverse flow, would a temp sensor installed in one or both of the coolant drain plugs be a better source of the coolant temp after the coolant has picked up the engine heat. As a test, use two additional sensors to match the existing dash gauge sensor. Run temporary wires to a three way switch that would allow you to switch between the three sensors, and read on the dash. Once enough information has been gathered, just make the desired one permanent.

BTW. I love the work you have done on the car, especially the challenges you deliberately took.

The heads run at the same temp. I kinda figured they would, but I had the two holes to plug and thought Moon gauges would look cool there. For the real oldtimers there is a tie-in: Moon bought Potvin, who made the blower drive, in 1963. And in 1993 Moon was kind enough to raid the warehouse for a mold and cast a piece I was missing.

I can read the EFI temp sensor on the handheld control unit and it matches the dash gauge. I like the 3-way switch, and I'm redoing the dash this winter...

Thanks for the nice words. I built the car 20 years ago partly as a test bed to try new stuff (lotsa folks along the way: "that won't work!"). I've been thru 2 inductions, 4 distributors, 2 motors, 3 EFI ECUs, 2 sets of injectors, replaced all the lights with LEDs, 4 fuel pumps and 2 regulators, and rewired 'til I dream schematics. But damn it's been a fun ride.

Don't forget the third control, the thermostat...Usual temp will be specific to your car, average is 185-210° F....should hover around 190°...Just shooting from the hip, I would set the first fan at ~198° and the second at ~205°....

Oh, no thermostat on my car. Coolant is piped by electric pump from the motor to the radiator located in rear of car. I've used 3 or 4 different systems to control the fans by coolant temp. Simple on/off of both fans by a common temp sensor, individual on/off of each fan by separate sensors, and this latest scheme. All have worked.

This latest control seems to hold a constant temp the best so far, maybe because the others were just on/off strategies.

I wonder how a conventional PID controller with setpoint would respond. I worked in process control many years ago; do they now have self-tuning controllers that can optimize their parameters?

Thanks for those great links. Oh damn, I'm starting to get the itch again...

The quaint terminology "process control of car cooling system" right from the get-go removed the topic from reasoned discussion.

Oh I wouldn't say that....

Whatever that was, I got tired of downloading it after about 30 seconds, and thus I missed it.

SolarEagle, you found my rebuild thread! Can't hide on the net. Comments and questions on that thread are most welcome, too.

Now if you want to go way overkill, you could install a 12 VDC to DC PWM speed control for your electric water pump. Some models are available that can accept a 0-10 V analog signal from a PID process controller. This would allow control of the coolant flow through the engine to maintain a consistent coolant temperature. In effect, this would do, at a higher level of sophistication and price, what a regular engine thermostat does.

Necessary components to do this are available from Automation Direct, to mention one source that I am familiar with.

I have entertained that idea for some time. Since the present pump has such a low flow rate, there is really nothing to be gained. Now when this pump (20 years old!) ever dies, I'll go with a new more powerful one and revisit that idea...if I outlive the pump.

Those old pumps don't like to be frozen with water in them ask me how I know. Big block Chevy with fan controls temp sensor in top radiator hose. A last trip of the season to the drag strip and forgot to drain car after last run, on the way home it got real cold and could not get all of the water out. Even running the car to warm it up would not help as the bilge pump would not turn the rotor to move the water. Hated to see the old girl go but just did not have the time to find another before the next race. Went back to alum Moroso pump and belt drive. And just never found the time to retrofit it back since someone wanted the old girl more then I did now I wish I had her back 66 Chevelle station wagon.

I'd be careful about reducing the water flow, especially with your induction system. Reduce it too much and you get a hot spot due to stagnant area or steam/air bubble trapped. What you have now is basically proportional control of the fan speed. PID might not make much improvement, and could start oscillating since you probably have significant dead time with that radiator mounted in the rear.

Is there a reason to adjust (or just monitor) so many variables if the goal is to maintain an optimum temperature? Fans and radiators with thermostats work to keep my car and my motorcycle amazingly consistent in any ambient condition. (Porsche and Ninja)

Of course, if you are paranoid enough, you can put 6 sensors on each variable. Then, just like gymnastics, you can throw out the highest and lowest values, and average the rest. The ultimate in SQC...

Well reasoned, Codemaster. Gave you a GA.

The range for T1 and T2 is 160 to 235. Here are two extreme examples of settings:

I wonder how the operating temp would behave in each case.

This control strategy, with no setpoint, fascinates me. I suspect that by fiddling with T1 and T2 one could attain a desired "pseudo-setpoint", a temp that would only vary by a few degrees due to air temp and engine load. I wonder what adding a little I to the P would do. I don't remember ever using D in the process control systems I did many years ago. Guess I need to get a real PID controller and do the experiment. No rest for the weary.

Thanks for the GA!

I'm not an expert on control systems, but I believe it's usual to enter a setpoint and a proportional band. Entering your T1 and T2 is doing the same thing by a slightly different route. Reducing the width of the control band gives tighter control but can cause instability.

If it's proportional only control there is an offset if the inputs change, as that's the only way the system "knows" to change the output. Integral control avoids the offset by in effect adjusting the setpoint (internally) till the offset is zero. As I'm sure you know if you have a background in control. Differential action I believe can improve stability. Of course you need to set up the various time constants to give best results. Be an interesting project.

Nice car BTW

Been decades since I did that stuff, but it was fun. Need to bone up on the latest wonders. It's slightly crazy to put an auto-tuning PID controller in a hot rod, but I love to experiment. And my sanity has always been questionable.

Partlow Controls has (or had) some really good literature on setting up and tuning PID controllers. (Basically "bump" the process, see how it reacts, do some simple calcs, and then reenter the revised parameters.)

Or you could try fuzzy logic. (I've never done it yet, but it sounds promising, and maybe by now well proven.)

I have firearms that I've never fired. I bought them because I wanted them. Not because I needed them.

Are you doing this because you feel you need to, or because you want to?

Want to!

While the vehicle is in motion is any air directed through the radiator ? I'd expect the fans to be off most of the time when just cruising with the enough fresh air being ducted I guess in your rear mount setup through the radiator(s).

Do you have any aircon in this? Auto or manual? Dry sump oil cooler?

I'd really like to see what you have created. That shiny blower at the front is sweet. No intercooler?

Beautiful work of technical art. Don't even try to sell it, the prices you will be offered will break your heart.

What portion of your life has gone into this? That's its real value.

Yes, there are scoops on the sides of the bed that direct air to the radiator, and the low pressure area aft of the radiator helps. The fans rarely come on when cruising in nice weather. Blower just pumps air and no intercooler.

Thanks for kind words. I've put 20 years into it, and can't imagine life without it.

Pics:

https://plus.google.com/photos/104971212334307682869/albums/5552767776876190417