Thursday, September 11, 2014

A Failed Water Pump and Building a Circuit

Sometime in January, I think it was, I was driving the Z3, having a good time, minding my own business, when the  Zilla controller's light started flashing at me, indicating a fault with the controller.  The car kept moving and responding as normal, and shortly afterwards the light stopped.  It's done this in the past, during summer time when cooling system wasn't up to keeping the controller cool.  But this was January, so that couldn't be it.  Right?  This happened a couple more times, completely baffling me.  Each time the fault would clear on its own by the time I got home.  I hooked the laptop up to the controller... No faults.  Hmm.

Shortly after that I took the car off the road for some mechanical upgrades and a few small repairs: new steering rack, and replace the damaged front bumper.  Because I'm careful with my work, and took my time (read lazy), the car was off the road for about 2.5 months.  I got it back on the road in late May.  The first drive back in the car is always glorious.  I love driving the car! 

On that first drive, I was on my way home and the warning light started flasing again.  Then it started flashing faster.  It was at that time I noticed that the car was barely responding to the throttle.  At that point, the light came on in my head and I realized what had been happening all along.  But how could the car be overheating?  A hot day in May, sure I understand, but not to this degree.  And certainly not back in January when it was 60°F outside!  What could possibly be going on?

I popped the hood and immediately notice a small bubble in the line that leads from the Zilla to the radiator.  It pains me to say it, but that is the high point in the system.  Anyway, I notice it's barely moving.  When I'd filled the system, the pump was able to push all that air through the system down to the radiator where I bled it out.  I climbed under the car, touch the little pump and I found it was cool to the touch, and didn't seem to be running.  I tapped it, and it started running, but it didn't sound right.  After much testing and what not, I was able to determine that my little Laing D5 Strong pump had given up the ghost.  That was a big surprise; these things are known to last forever.  Fortunately I had a replacement pump waiting in the wings, but not the same make or model.

Sometime last February, Jack Rickard at EVTV announced that they had some Pierburg water pumps for sale that they'd acquired in the Azure Dynamics bankruptcy.  These are the same pumps used by multiple OEMs including BMW.  They normally retail for nearly $500, and he was selling them for $100.  What a deal!  Maybe it was premonition, but I decided to buy one in case I needed it.  Truthfully, it was probably a stupid purchase at the time, but fate intervened to make me look like a genius. 

The only thing is this pump puts out 8 gallons of water a minutes, but he Zilla only requires 2.  The thing is, more water is not better.  Your first thought might be that more water will likely cool the system better (I know that was mine).  But it turns out that if you push water through a heat transfer system too fast, it simply doesn't have the necessary amount of time when it's in contact with the surfaces to transfer heat efficiently.  Fortunately, this pump has a PWM input signal line that you can use to control the pump, dialing it from 10% duty cycle up to 100%.  The problem is, I know nothing about circuits other than if pressed I can identify a circuit versus say a can of fruit. 

I decided to do a little Googling to see what I could find, and wouldn't you know it, I found a YouTube video (since taken down) posted by a man named Klaus Wolter about how to control the pump.  He subsequently posted a design for a PWM circuit he'd designed to control the pump.  I'm reprinting the circuit here, though I have no way to contact Klaus for permission.  Klaus, post a comment if you object.

For those of you who don't know, a PWM signal, or Pulse Width Modulated signal, is simply a series of On-Off signals used to control a motor.  Now some PWM controllers actually send current to a motor to control it's speed, like for instance the Zilla controller in Z3.  Others simply produce a PWM signal with no current.  That's what this pump requires.  It reads that signal and the circuitry in the motor interprets it and then controls the motor's speed, meaning that this circuit doesn't directly control the motor.  It became clear that I was going to need that circuit if I was going to make this pump work in my car, so I decided to give it a go and try and build it.

I found a couple electronic component stores near by and went and collected all the parts I'd need. I bought at least two of each figuring that I was sure to foul up the first attempt.  Next, a friend encouraged me to buy a bread board so that I could build it out first without soldering any of it.  The only problem is that there were a couple aspects of the circuit that I didn't understand, particularly how that diode fit in.  Fortunately my bother-in-law Mark is an EE and he talked me through it, showing a great deal of patience I might add.  I put the circuit together on the bread board and sat there staring quite proudly at my accomplishment.  It was at that time that I realized I had no way to test it.  I don't own an oscilloscope.  But, I had an idea.  I know that motors controlled by a PWM circuit see an average voltage depending on duty cycle.  If you have a 100 volt system, and you're driving the motor with a 50% duty PWM signal, the motor "sees" 50 volts.  OK, really it sees 100 volts for 50% of the time, but since it's switching on and off so fast it averages out to 50 volts.  I'm sure every EE reading this is going to be shaking his head for some reason, but that's how I understand it.

With that understanding, I thought I might be able to throw my multi meter on the circuit's output and I should be able to measure a difference as I twisted the potentiometer.  Sure enough, I could!  OK, time to solder it up.  Here's what I ended up with.

Not terrible for a first ever attempt.  But would it work?  I hooked it up to the pump, ran some hoses to a bucket of water and turned it on.  Nothing.  I could feel the pump was on, but it felt like it was barely running.  If I twisted the pot I could tell the motor spun a bit faster, but still the pump moved no water.  Well, time to troubleshoot the circuit.  I spent hours going over it, checking and re-checking, but it all looked good.  I finally decided I was going to EVCCON in a week or two, I would take it there and ask some friends there if they could take a look at it. 

I explained the situation to a friend, one of the attendees, Doug Ingraham and he happily offered to help.  We borrowed Jack's oscilloscope, hooked it up, and to my great surprise, the circuit worked perfectly!  I was thrilled, but puzzled as to what could have been the problem when I was testing it.  It took a quick conversation with Jack to discover the problem.  Once again, it was my surprisingly great ability to assume that was my downfall.  I had assumed that if there was a problem, it MUST be the circuit I'd just built.  In my defense, that was more than 99% likely.  In this case however, it was the fact that these pumps are not self priming.  There was no way the pump was ever going to suck water up out of a bucket.  *Hand on forehead, shakes head slowly.*

I got back from EVCCON, and the next day tested the pump with a gravity fed system and it worked perfectly.  But it was a good thing I spoke with a number of people about this while at EVCCON because I learned I was going to need to make a couple other modifications to the car to accommodate the larger diameter inlet and outlet from the pump.  Particularly I was going to need to make sure that the pump had enough water feeding to it, or the pump would cavitate, ultimately ruining the pump.  The old pump had a 3/8" input, but new one, 3/4".  The tiny reservoir that supplied coolant to the old pump, with its 1.5 cup capacity and 3/8" outlet would not do.  I replaced it with a 1 quart model that has a full 3/4" outlet and inlet. 

I had the whole system together a couple days later, and the car was back on the road running cooler and better than ever.  That is until I turned on the A/C for the first time and found that it had lost it's charge.  I think I might have hit a hose or something while working in there, perhaps I broke a seal and allowed the gas to escape.  *Sigh*  It never ends.  But winter is coming.  I don't need to attend to that right away. 


Klaus W said...

Hello, Tim. I'm the Klaus that posted the PWM pump circuit you built. I just wanted to say I'm glad it worked for you and you are welcome to post the circuit for anyone to use. You are also welcome to contact me if you have any further questions. I love your build!

Tim Catellier said...

Klaus, Thanks for the comment, and thanks for putting together, and sharing that great little circuit. Believe it or not, I've been contacted be a couple individuals asking about it, and I shared with them the PDF that you put together. It's been immensely helpful.

Jacques Le Roux said...

Klaus, I just want to know if I can use this same circuit for the cooling fan? I need a circuit that receives an input from the temp sensor and then supply the pwm signal to the fan