Current regulation based on frequency

[hotepmichael]

I need to design a current control current based on frequency.
I have a flywheel spinning at 60 cycles/sec. If that flywheel changes frequency, I want to feed that change to a current regulator circuit.
If the frequency go below 60 cycles/sec, I want the current to increase, and if it goes above 60 cycles/sec, I want the current to decrease. I have a good understanding of current regulators that use Zener diodes as the control element, but do not know how to get started with frequency as the control element. If someone could direct me to some information, I would be in your debt.

Thanks,

Hotepmichael

 

[MikeMl]

Do you want a digital or analog circuit?
What are you planning to use as the 60Hz reference?

 

[hotepmichael]

Thank you sir,

The 60Hz reference will be a spinning flywhell, and I want it to be an analog circuit at this time.

Hotepmichael

 

[MikeMl]

The 60Hz reference will be a spinning flywhell, and I want it to be an analog circuit at this time.

No, the reference will be an LC oscillator, the 60Hz power line, a 555 oscillator, etc. (something that is always at 60Hz, regardless of how fast the flywheel is spinning).

The frequency of the the signal from the flywheel is what you need to compare to the 60Hz reference frequency, using a digital frequency comparator, ratio detector, or frequency discriminator.

 

[marcbarker]

Sounds like a job for a CD4046. Make up an 'opamp' circuit to drive the motor speed control, but utilise 'frequency feedback' rather than voltage!

Feed a 60 Hz reference (i.e. power line frequency) pulses into one PC input. Feed a one-pulse-per-rev signal from your flywheel into the other PC input. The PC output drives a low pass filter, say 100 K ohm pot and 10 uF to ground. Insert a 1 K pot in series with the 10 uF and ground. The voltage developed on the cap is unity-gain buffered and the buffer output (via a resistor to reduce the control loop gain) becomes part of the speed feedback to the motor speed control. Easy!

The motor will acellerate the flywheel to 60 rps, may overshoot it, then settle nicely at 60 rps. The 4046's VCO is not used.

If the speed of the flywheel is 'hunting', adjust the 1K higher. To increase the control loop gain, reduce the 100K pot setting.

 

[Hero999]

The original poster hasn't filled his location in so we don't know whether he's in the US (60Hz) or EU (50Hz).

If you're unluky enough to have a 50Hz power line then you'll have to make your own frequency reference.

 

[tcmtech]

Best bet would be to use the LM2907 Frequency to voltage converter IC. It can be set up with either voltage or current output relative to its input frequency. They are simple and easy to use.

Its in my top 10 favorite ICs lists too!

Heres the full specs sheet for it.

 

[crutschow]

Best bet would be to use the LM2907 Frequency to voltage converter IC. It can be set up with either voltage or current output relative to its input frequency. They are simple and easy to use.

A F/V converter will not work properly for this application. You need a circuit to compare the reference frequency with the generator frequency and output an analog voltage such as the 4047 suggested by marcbarker, or a exclusive-or gate used as a phase-detector.

 

[Mr RB]

Why won't tcm's freq->voltage work?

That's how I would do it. I would use a 555 as a monostable pulse generator, then filter the output to give a DC voltage linearly proportional to freq. Then use an opamp to compare that voltage to a reference voltage and drive the motor current accordingly.

Ideally you would generate more than 1 pulse per revolution, to speed response time and one payoff is the control system is not dependant on any freq reference, you could just adjust the ref voltage and use it for 50Hz or 60Hz etc with a twiddle of a trimpot.

The OP never said he had a reference frequency or needed very high accuracy, he's talking about zeners etc as control elements, I would assume its for controlling a generator or something simple like that.

 

[marcbarker]

Best bet would be to use the LM2907 Frequency to voltage converter IC. It can be set up with either voltage or current output relative to its input frequency. They are simple and easy to use.

Its in my top 10 favorite ICs lists too!

Heres the full specs sheet for it.

That's an idea, I used this actual F-to-V conv once in a vehicle cruise control system, which worked very well.

You can't just generate the motor control current signal directly with it though, needs some more analogue signal processing, otherwise the speed will 'hunt' or surge continually around set point. Plus you also need a error amplifier and reference voltage of some kind.

My opinion is if you want to work with voltages during development, go for a F-V chip plus analogue signals solution. If you want something much simpler and don't mind working with frequencies, use a 4046 solution.

 

[tcmtech]

As typical we run around with different ideas but until the OP tells us what power he is using and any other detail we can know for sure.
So what are you doing exactly with a flywheel now that produces a constant RPM and references it to a current change.

The LM2907 is easy to use as a voltage reference to a current amplifier circuit. By using a voltage rise through a fixed Resistance you will get a proportional current reference. Its how a basic frequency to 4 -20 Ma signal converter works.

I have often thought about using a LM2907 IC and a power op amp to drive the throttle butterfly on an electronic drive by wire throttle body. Just by using a simple pot and a few other basic components I could make a full drive by wire governor for a gen set that would have a very tight speed control range. One far finer than a standard mechanical governor has. Its what this thread reminds me of in a way.

 

[marcbarker]

I have often thought about using a LM2907 IC and a power op amp to drive the throttle butterfly on an electronic drive by wire throttle body.

Been there, got the t-shirt.

It won't work the way you're currently thinking it will, ....because I know something you don't know Bet you can't guess what it is?

(clue: 'something special' needed in the feedback loop, to counter the 'hunting' effect)

 

[hotepmichael]

Gentlemen thank you very much for the ideas. I am looking at the opamp right now.

 

[Hero999]

(clue: 'something special' needed in the feedback loop, to counter the 'hunting' effect)

Phase compensation to prevent oscillation?

 

[marcbarker]

Phase compensation to prevent oscillation?

Near enough... I'll give you it.

In a process with a lot of intertia, such as a cruise control for a vehicle, or a heavy mass flywheel, no amount of proportional control alone will be right. The speed will hunt, because of a lack of 'damping'. So in the control loop you need to add a very good helping of 'D' term into the mix. Feed back the 'rate-of-change' (of the process) as an error.

In an analogue circuit with opamps, it'll literally be a differentiator. If it's a 4046 circuit, it'll be some ohmage inserted into the LPF capacitor.

You don't have to bother with any 'I' term at all, because the process itself is a big integrator.

 

[tcmtech]

Been there, got the t-shirt.

It won't work the way you're currently thinking it will, ....because I know something you don't know Bet you can't guess what it is?

(clue: 'something special' needed in the feedback loop, to counter the 'hunting' effect)
__________________

So you just figured I would hook the LM2907 right to the Power amp IC? You should know me better than that. It would have taken a few paragraphs to describe the whole system.

But at least your paying attention and thinking!
I've worked with commercial and industrial stuff for years and am fully aware of the need for dead band/hysteresis on gen set governor systems.

I just though it would be a fun thing to build on a slow day. Thus that why its never been pursued in some years now!