Still Trying to come up with a cruise control that works on an off road vehicle

[kinarfi]

Me plans are to make a cruise control by monitoring the engine speed via the tach output, 2 cycles per revolution, and convert that frequency to voltage via LM2907-8 and use that voltage to control the throttle via a servo.
I would like 1800 rpm to produce a servo pulse of 2 ms and 4500 rpm to produce a pulse of 1 ms at a frequency of 50 Hz, or vice versa, so as the engine slows down, the servo increase the throttle and as the engine speed gets faster, the throttle backs off.
I tried using vacuum - it sucked
Tried using a DC motor, that doesn't work very well either.
Now I'm trying to use a servo of the remote control variety from model air planes. Just need to figure out how to create the set point that the servo will try to maintain.
Anyone have any suggestions or experience with this?
I am running out of ideas to try.
Thanks
Kinarfi

 

[ronv]

Sounds like fun. Difficult, but fun.

Just 1 gear?

 

[ronv]

Pulse generator

I think something like this might work to generate the servo pulse. If I understand the hobby servos, driving them more often than 50 ms is ok - in fact better.
We need to figure out compensation of the tach and scaling to the CV input on the 555.

 

[kinarfi]

5 gears forward, 1 backward, seems like I'm stuck in backwards most of the time, except when driving. I want to monitor the engine so that if I'm heading up a rough hill in a lower gear, the cruise control will hold the speed steadier than I can with my handicaps. The commercial ones won't stay engaged, the dc motor one works fairly well but it overshoots and as long as there is any power applied at all, it won't back off, then it backs off too much.
Need to figure out how set the correct pulse width with a pot, and then have the engine speed adjust it accordingly. Maybe a voltage adder controlling the pulse width.

 

[ronv]

Yea, I think that's going to be the hard part- compensating the response to the engine. Do you have a spice model for the frequency to voltage converter. If not how about the schematic - then we can get on the same page.

 

[kinarfi]

Here is what I have on my plate so far, notes are in the schematics
Thanks for looking!
Kinarfi

 

[alec_t]

According to your notes

L1 is a servo motor with the guts removed, so it is just a geared motor,
it is driven forward by the pwm and it was my hopes that the pull from
the trottle spring would unwind it, but if there is any voltage applied, it
won't unwind

If you've removed the feedback pot then the servo will no longer be responsive to pulse width to maintain position. So it's no longer a servo, just a geared motor, and a PWM drive for it seems unnecessary. Wouldn't a simple comparator circuit do the job?

Edit: @ronv

If I understand the hobby servos, driving them more often than 50 ms is ok - in fact better

That's maybe true within reason, but isn't your PWM rate about 250kHz? I'm guessing that may be well beyond the servo IC's range.

 

[kinarfi]

According to your notes
If you've removed the feedback pot then the servo will no longer be responsive to pulse width to maintain position. So it's no longer a servo, just a geared motor, and a PWM drive for it seems unnecessary. Wouldn't a simple comparator circuit do the job?

Part 03 ALL 040712 mod1.asc is just what you suggest and state, except L1, servo motor with all of it's guts removed, gear motor, is driven by a FET via a comparator.
Another trick for servos is to replace the pot with 2 resistors of the same value, then it will stop rotating only when a pulse width of 1.5 ms is applied.
Kinarfi

Going to go try this, next.

 

[alec_t]

servo motor with all of it's guts removed, gear motor, is driven by a FET via a comparator.

I see that, but why use pulses rather than DC to drive the motor? With the guts removed the 'servo' can't care less what the pulse width is and can only drive in one direction. Using a PWM signal would only be useful if you needed speed/torque control (is that what you're doing?) or, as in an unmodified servo, position control. IMHO position control is what you do need, rather than relying on a spring return.

 

[3v0]

I would add a rev limiter to the circuit.

If you can not get to a reasonable solution it may be time to think about using a micro controller.

 

[strantor]

the dc motor one works fairly well but it overshoots and as long as there is any power applied at all, it won't back off, then it backs off too much.

That's because of :

as the engine slows down, the servo increase the throttle and as the engine speed gets faster, the throttle backs off.

That's the nature of the beast, it's going oscillate without some kind of intelligent control. You need PID or some other kind of control algorithm.

 

[ronv]

Here is what I'm thinking:

Frequency to voltage converter to compemsation stage to 555 control voltage to normal servo (driven in both directions)

Desired RPM set with a pot.

Attached is a block diagram and a sim of a first pass at the compensation. I'm not very good at this so it may take some cut and try.

The downside I see is that it will lock on utill you turn it off.

The voltage in the sim is the control pot, but you can play with it to see the response of the circuit. I just guessed at the response of the engine.

I think the output of the 555 goes between 1 and 2 ms. Maybe I posted the wrong one.

 

[alec_t]

I would also be concerned about safety over-rides to prevent the throttle getting stuck open, and insurance/liability issues. Here in the UK insurance companies would refuse cover for vehicles with mods of this sort.

 

[MikahB]

Cool project! Seems like a PIC would be a great way to execute this - including the safety part (brake light on disengages cruise). I'm a newbie but would be happy to work on the PIC/code part of it (P-I Control) which I need to do anyway.

 

[strantor]

I really would recommend looking into PID. When you're dealing with a latent variable like the speed of a vehicle, no matter how sophisticated your circuit, if it's purely proportional, you are going to have oscillations and your cruise control is going to do all kinds of wacky stuff if you're not on perfectly level ground. Cruise controls found in cars are PID controls.
How Cruise Control Systems Work (PID)
More info about PID

 

[ronv]

Stantor,

Maybe you could take a look at this compensation and see what you think.

It needs some mods but is the same as the simulation attached above.

It does beg for a micro --

 

[kinarfi]

A few answers, I was using pwm to drive a dc motor hoping that when the pulses were vary narrow, the throttle tension would unwind the motor some and as the engine slowed down the pulses would widen and the motor would wind up a little, generally, the motor would be stalled all the time and pulse width would vary and thus vary the torque of the motor and the output of the "gear motor" would be in the area of 30 to 180 degrees.
rev limiter is in the ECU and on this engine, it will only rev to 6500 rpm.
Engagement, disengage and brake disengagement is handled by a different part of the circuit.
Also remember, this is on an off road vehicle and one of it's purposes it the help control speed during rough climbing terrain where I have a tendency to end up full throttle because of my disabilities.
Thanks,
Kinarfi

 

[Mikebits]

I don't see how cruise control would be practical in a off road situation. With unpredictable terrain, unexpected turns, whoops and dews out of nowhere. With all the obstacles presented to the off roader, they must have total control of the vehicle at all times. Off road cruise control would be nothing but dangerous, and a bad idea in my opinion, but whatever I guess.

 

[strantor]

Stantor,

Maybe you could take a look at this compensation and see what you think.

It needs some mods but is the same as the simulation attached above.

It does beg for a micro --

Ronv, I don't have the circuits expertise to critique that. I have experience with off the shelf control devices, but as far as the circuits themselves, I'm in the dark. If I were going to design this device, I would go straight to a micro, because I know I can figure out how to do it in code much faster than trying to figure out a circuit.

 

[3v0]

Maybe what is needed here is more of a fly by wire setup rather then a cruise control ? Some sort of improved interface between the driver and the machine.