# Servo Seems Absent in Automotive HVAC

Analog engineers know what a servo system is. More specifically, they understand (probably intuitively) how a closed-loop servo system is defined and how it operates.

There are three parameters or constants that are used to define the loop: the proportional, integral, and derivative gain terms. One, two, or all three of these are used in a typical system. The system will compare the setpoint (the desired result) to some measurable process variable. The system will generate an error based on the difference between setpoint and process variable. This simplified description shows how straightforward the system design should be.

This drawing, taken from the Wikipedia page for PID controllers, shows such a straightforward system:

A PID Closed-Loop Controller

This is the block diagram for the classic closed-loop controller that is used everywhere.
Except, apparently, in my cars.

HVAC (heating, ventilation, air conditioning) systems in cars have come a long way since the days of Henry Ford. Over the past 20 years, we’ve progressed from a slider knob that has its extremes labeled “Cold” and “Hot” to a digital readout with associated “Up” and “Down” buttons. This setup gives the impression to the user that there is a servo system hiding behind the dash that accurately and tightly controls the temperature in the cabin so that it matches the setting displayed on the readout.

Is this really a calibrated, closed-loop servo system? Or is this (as someone, somewhere else said) a design made by monkeys?

My experience on several vehicles indicates this is a system that is closed loop only if I am included as part of the loop. I need to characterize and quantify my annoyance, create a numerical value for the annoyance parameter, and write the corresponding loop equations for the control system. The reason for my annoyance? I have experienced systems that seem to have little regard to the outside temperature, the engine temperature, the cabin temperature, or the digital readout. Sometimes they seem to function as an on-off switch. Below a set-point of, say, 68, the heater is off; above 68, it’s on. I had better performance from my ’64 Falcon that used vacuum actuated baffles in the heater ducts.

Perhaps my cars use baffles that get stuck. This possibility points to a poor design. I don’t care to crawl under my dash and experiment with the control mechanisms to find out. Perhaps my cars use a simple proportional control to adjust the flow of hot water through the heater core based on temp-setting. Perhaps cabin temperature is not monitored at all. This seems like a distinct possibility.

Regardless, I just want better designs and better implementation. Comments welcome.

## 6 comments on “Servo Seems Absent in Automotive HVAC”

1. Michael Dunn
January 23, 2013

Back when my AC still worked, anything other than the coldest temperature setting was terribly inefficient. The AC was always on full, but warmer teperature settings would mix in warm air. I bet most cars still do this – at least the cheaper ones.

2. eafpres
January 23, 2013

Hi Brad–I share your view.  My guess is some systems simply change the position of the valve which regulates the amount of coolant that flows through the interior heat exchanger.  That what the knobs & sliders did–the motion moved a rigid cable wire which was connected to a lever which operated the valve.  Most of those systems I've had “featured” an inflection point right about where you wanted to stabilize the temp.  So you constantly adjusted back and forth.

I know there are systems with actual temperature sensors and controllers, but if the valves are very non-linear, and the air-dampers are too, and the sensors are in bad places (all in the dash?) then it still won't work well.

I always enjoy getting a rental car with 2 front zones and seeing if I can drive it crazy setting one way hot and the other side cold so they fight it out.

January 24, 2013

I find my newer HVAC automatic system vastly superior to the older ones.  Back in the '70s GM used electronic control over vacuum actuators.  They were servo systems, but performed rather poorly.  When the vacuum supply hose melted on the exhaust manifold, I got full fan full heat on one of the hottest days of the year.  I suppose you could call that the fail safe mode.  Better than freezing in winter.  Even when working, I was always adjusting the temperature knob on the older systems.  In fact I found simple manual controls better thorough the years.  But I must admit, my 2007 Buck Lucerne has excellent temperature control with a separate setting for the passenger.

I believe the problem in many systems is the sensor part of the loop.  You cannot just stick a thermistor on a black dash and expect it to read your perceived temperature.  Sunshine and shadow on the sensor and defrost mode made the older systems freeze and fry.  Another fly in the ointment is the sun's InfraRed hitting the person on one side of the car and not the other and making him uncomfortable.  Even a perfect system would have to  allow for the clothing worn in summer and winter.  Somehow the 2007 version works very well and I only adjust it a couple of degrees as I get up to temperature to get the temperature just right.  Another complicating factor of course are heated and cooled seats as well as heated steering wheels.  There's more to the design than meets the eye.

January 24, 2013

CJ – admittedly, the system is a bit more complex than my description. However, since some cars have systems that work fine, I know it can be done. Thanks.