Understanding the VDO Cruise Control Amp

[Tim Polzin]

Made some progress this weekend on the cruise brains. I've started a new thread instead of adding it to 3 plus pages of the last one.

First off, thanks to James Mahaffey, a MB affectionato, we now have a description of a similar, although not identical VDO cruise circuit.

The 10-pin connector is actually a 12-pin connector with pins 1 and 2 blanked off.

The speed-sensing portion of the cruise control circuit is shown in Figure 1. Test points are indicated by boxed numbers. This wave is observable at point 1, is then half-wave rectified (point 2). The signal is then amplified and limited by Q5 and Q6, giving a
rectangular wave of 6 volts at point 3. At point 4 the wave is a 0.1 to 0.2-volt ripple. The signal is converted to a DC voltage, varying linearly with respect to the frequency of the speed sense signal, observable at point 5. R69 trims the sensitivity.


Voltage Comparison

The voltage generated in the speed-sensor circuit is captured on capacitor C19 when you set the cruise control by engaging the Accelerate Set or Decelerate Set controls (see Figure 2). In the engaged condition, the output of Q9, a dual-gate FET riding C19, becomes the standard voltage against which the car's throttle is controlled through a servo-amplifier. An Accelerate Set signal latches down relay 2 (K1 in Figure 5), overriding the controller, opening the throttle and connecting C19 directly to the speed sensing-circuit. A Decelerate Set command does the same except that the output current is turned off, closing the throttle. A Cancel command disengages the controller by dropping the latch of relay 1 (K1 in figure 5), but the charge on C19 is preserved. Setting Resume relatches K1, and pressing the brake pedal unlatches KI. The standard DC voltage is amplified by a section of the quad comparator, U1, connected as an op amp. R76 trims the servo gain.

Another part of U1 is connected as a multivibrator, as shown in Figure 3. The amplified DC voltage controls the pulse length of this oscillator, observable as a rectangular pulse train 7 volts high at point 6. The oscillator output is converted to a current of 100-300 mA, on pins 7 and 3, by Q2 and the final driver Q1. The current is used to drive a pilot valve controlling a vacuum-diaphragm actuator connected to the throttle. Voltage indicative of the average current produced by Q1 is observable at point 7. Q3 is the current-to-voltage converter, providing a negative feedback stabilizer to the multivibrator. R27 trims the multivibrator switching decay rate, and R11 trims the negative feedback.

A third section of U1 is connected as a Schmitt trigger, as shown in Figure 4. If the car's speed drops to 10 MPH below the "stored" speed (standard voltage), the Schmitt trigger fires, giving a negative pulse observable at point 8. The pulse, amplified and inverted by the
fourth section of U1 and Q4, causes K1 to unlatch, dropping control of the throttle. This condition is perfectly normal and is particularly noticeable in a diesel on a steep climb. Another normal drop-out condition is a speed below about 12 MPH. When the standard voltage (charge on C17) drops to 1.0 volt, Q8 conducts, grounding the negative input to the latchout amplifier, as would the Schmitt-triggered negative pulse. R74 trims the Schmitt trigger threshold, and R36 trims the latchout amplifier gain.

Figure 5 shows the relays and the diode logic array for the four cruise control switches; all four are momentary switches. A Decelerate Set command connects pin 4 to +12 volts. Accelerate Set connects pin 9 to + 12 volts. Pin 10 is connected to +12 volts upon Resume, and pin 8 is always at +12 volts except momentarily upon the Off command. Pin 6 is connected to the brake light, with K1 grounded through the light unless the brake pedal is pressed.

Fig 1

Fig 2

Fig 3

Fig 4

Fig 5

[dshepp806]

Yeah: Now THAT's what I'm talking about!!!!!

WELL DONE....and THANKS!!!!!!!!!!

Best,

[Tim Polzin]

The next step was building a test bench to fake the amp into thinking it was running.

I used a dpdt spring loaded switch to operate the set/resume function, and a spst switch for the cancel function. The common connector of each was tied to +12 v. The set line runs to pin 9, the resume to pin 10 and cancel to pin 8. To cancel, 12 v is momentarily disconnected from pin 10. The brake switch pin (pin 6) is grounded.

Pins 3 and 7 provide output current to the vacuum actuator. I connected a low value resistor between these lines to monitor voltage.

Ground to pin 5, power to pin 12 and signal to in 11.

Speed is sensed by a line from the speedometer. Since the 911 has 8 magnets activating the speed sensor, one revolution of the rear wheels results in 8 pulses. A typical tire/wheel combination will rotate roughly 810 times in a mile, resulting in 6480 pulses. A speed of 60 MPH gives a 108hz pulse rate inserted into the cruise control module at pin 11. I simply fed a reed relay with low voltage AC and generated a pulse rate of 60 hz, which was enough to make the amp function. The reed relay was connected between pin 11 and ground.

This basically gave me a test bed to run the cruise amp.

One interesting observation is the output to the vacuum actuator is a series of what appears to be non-uniform pulses. The duty cycle of the pulses seem to determine the operation of the actuator.

Briefly, what I found on two of my amps was poor solder joins. Running the amp and using a scope to trace the output signal, I was able to pinpoint the exact faulty joins.

The third unit lost speed after setting. Reading the output voltage of Q9, I could watch the voltage decay with time. I replaced the sample/hold capacitor (c19, figure 2). That seemed to cure the problem. One word of caution. Q9 is very sensitive to static. Be very careful in testing or replacing any part of that circuit.

All three of my boxes now appear to be operating. Once the snow is gone and the streets are dry, I'll try them out in the car.

Whats next? Figure how to set sensitivity to cut down on speed variation that I recall one of the amps exhibited.

Tim

[Tim Polzin]

For clarity, here are the locations of Q9, C19, and relays K1 and K2 (actually reversed from the numbers shown).



Tim

[imcarthur]

Great work, Tim. Were the nfg solder joints in a specific place or random?

Ian

[Tim Polzin]

Originally Posted by imcarthur

Were the nfg solder joints in a specific place or random?

Ian

On one board, (the one that never worked) the faulty join was in the output stage before it got to the output transistor (Q1). The join was simply poorly soldered, and was nicely covered with all the crap they dipped the board in. For the second amp, the problem was at the connector where it was soldered to the board.

Tim

[rick-l]

Subscribed

Any chance you can see a part number on Q9 or is the part in a heat sink?

Do you know where voltages A+, B+ and C+ derived?

[dshepp806]

Keep it coming.,.,.,.I'm in heaven!!!!

[Tim Polzin]

Originally Posted by rick-l

Subscribed

Any chance you can see a part number on Q9 or is the part in a heat sink?

Do you know where voltages A+, B+ and C+ derived?

BSV81 n-channel mfet 30v, 0.05a - T0-72 case.

Don't know about the A,B@C voltage source.

Tim

[Lorenfb]

"First off, thanks to James Mahaffey, a MB affectionato, we now have a description of a similar, although not identical VDO cruise circuit."

So, it's the schematic of the early Mercedes Benz cruise control, as the Porsche
cruise control uses 2 quad comparators, i.e. 8 comparators and figures 1-5 only
indicates 5 comparators. Thus, a questionable and incomplete schematic, right?

[rick-l]

Tim Polzin] BSV81 n-channel mfet 30v, 0.05a - T0-72 case.

That part must be ancient.

The only reference I can find to it is from NXP (Phillips) go to page 6 of selector guide

It says it is a depletion mode MOSFET. I am not that familiar with that type of part and wonder if that is really the part in the cruise control box. I haven't seen a symbol like the one on the schematic (no arrow). I would assume the four leads are Drain, Source, Gate and Body. A negative voltage on the gate with respect to the drain turns it off.

[dshepp806]

Geeze,..can't we get our hands on the real deal (ie: Porsche)?

Tim, your work is still appreciated and well done.

[Tim Polzin]

Originally Posted by rick-l

That part must be ancient.

I am not that familiar with that type of part and wonder if that is really the part in the cruise control box.

Yup, right off the part. State-of-the-art 35 years ago.

Tim

[Lorenfb]

"It says it is a depletion mode MOSFET."

Correct! Most/all were depletion mode back then, i.e. few if any were enhancement mode.

Those doing the "shotgun approach" (as most DIYers) might start replacing identifible parts,
e.g. capacitors, I.C.s, transistors, based on guesses and end up damaging the ECU.
Kinda like the DIY understanding and repair of the Bosch CDIs or the DME ECM
which are much simpler to diagnose than the cruise control ECU.

[imcarthur]

Wow. What else can I say? Just Wow!

Loren, as usual, your contribution to this thread & the other Cruise Control thread & in fact, all of your contributions to this forum & Rennlist are just so . . . so . . . electric, for lack of a better word. Just the fact that you would take time out of your busy, busy day to spread your words of infinite wisdom to us lowly, mere-mortal DIYers struggling with our “shotgun approach” in the wilderness of our garages. I know, we are all secretly, horribly ashamed that we take money – no let’s call it what it is – we STEAL money from bonafide, knowledgeable electronics ‘experts’ like yourself & starving mechanics around the globe - every time we pick up a wrench or a voltage meter to work on our car. And you are so right, we are bumbling idiots & just WHO do we think we are, posting our silly questions on this forum & sharing information that takes bread right OFF YOUR TABLE! If we had a shred of human decency & intellect, we would flock to your business & throw bags of cash at your door!

We are all waiting – with bated breath – for the next nuggets of information that will trickle from your keyboard whenever you deem us worthy of more of your scintillating guidance.

Ian

[rick-l]

Originally Posted by rick-l

Do you know where voltages A+, B+ and C+ derived?

Ooops.... I didn't see +12 volts and Cancel are the same pin (Pin 8). The unit must be supplied power from the cancel switch.

So to answer my own question

1. A is whatever voltage zener D32 is
2. B is a filtered 12 volts (V+ to the comparators ...headroom)
3. C is raw 12 volts to servo

[Phantomflyer]

I just tried the "shotgun approach" for my cruise control that always set 10mph to slow. I replaced C19 but nothing changed.

[LM3929]

subscribed

[scarceller]

Originally Posted by Tim Polzin

Made some progress this weekend on the cruise brains. I've started a new thread instead of adding it to 3 plus pages of the last one.

First off, thanks to James Mahaffey, a MB affectionato, we now have a description of a similar, although not identical VDO cruise circuit.

The 10-pin connector is actually a 12-pin connector with pins 1 and 2 blanked off.

The speed-sensing portion of the cruise control circuit is shown in Figure 1. Test points are indicated by boxed numbers. This wave is observable at point 1, is then half-wave rectified (point 2). The signal is then amplified and limited by Q5 and Q6, giving a
rectangular wave of 6 volts at point 3.

Tim,

I have been hunting for a good source of Vehicle Speed signal for my WideBandO2 system. I wish to record speed with the system and have it all setup to do this. But I need a nice 0-5Vdc square wave as the trigger to the TTL input in the controller. I see you mention that at point #3 in your schematic that it has 0-6V wave.

Is this always available here even when not using SpeedControl?
Can you sort of explain what the wave looks like? Does it swing + DC only? or does it also swing - DC as well?

I don't mind tapping into the speed control module to obtain a good signal, I never use Speed Control anyway.

Thanks for the great write up!

[Tim Polzin]

Sal:

Off hand, I don't know the answer to your question. When I had it set up on the bench, I faked the input signal line with a pretty noisy and spiky source.

If you just need a TTL signal, that should be pretty easy to come up with using the trigger line from the speedo, but wouldn't you actually be looking for engine RPM for your application?

Tim