What is the phase detection capabilities of a CD74HC7046AE?

Question

So I have a CD74HC7046AE IC and I also have 2 separate square waves where 1 wave is 90 degrees ahead of the other in phase. Both waves have a 50% duty cycle and a 1 Hz Frequency. They BOTH also are 5 Vpp with a +2.5V offset. If Wave A leads wave B (90*) the motor is spinning CLOCKWISE, but if B leads A (90*) the motor is spinning COUNTER-CLOCKWISE.

How can I use this phase detector to easily differentiate between directions? For example: if A leads B then the output is 5V, but if B leads A the output is 0V.

I was looking at the datasheet and down on page 5 a condition is shown, but this is not very descriptive in my opinion...

From what i gather from this the output would be the following:

If A leads B, a square wave would be outputted between the High Impedance off state (~2.5V??) up to 5V.

If B leads A then a square wave would be outputted ranging from 0V up to the high impedence off state (again 2.5V??).

I figure if this is the case I can then throw the signal through a low pass filter to average them out (The time lag is ok).

Is this what the outputs are? If not then what are they? How does a phase comparitor actually allow someone to differentiate between phase if it doesnt output distinctly different voltages for each condition?

NOTE: Wave A=CompIN, while Wave B=SigIN and the output is PC2out.

Answer 1

Hi Jeff, still working on it, eh?  Yes, it looks like you are correct re the pulse outputs under the 2 conditions.  And let's say you attach a 50/50 divider, say of 2 10k resistors between Vcc and COM, and the tristate output connected to the junction of the 2.  The divider eliminates the vageries of what level the output is at when neither output transistor is on, a case that happens every cycle and is liable to cause logic probs w/ the input of whatever this will go to.  And yes, you could low-pass it, but you still need a comparator with some positive feedback after the low-pass.  The comparator will also need a 2,5V souce to compare with.  Alternatively, I suggested instead the pair of inverters, still with the positive feedback resistor, and an input cap.  What I like better about this approach is that no low-pass filter is needed, and there will be no delays other than the propagation delays of the inverters.  Of course, the component count is a little lower, especially if the 2 inverters are 1/3 of a hex inverter IC

But remember I mentioned that someone probably makes a quadrature decoder IC?

http://literature.agilent.com/litweb/pdf...

http://findarticles.com/p/articles/mi_m0...

Both command and feedback counters plus DIR outputs all in one package!  Google "quadrature decoder" for other possibilities.  It may be that the entire loop has a 3 or even 2 IC solution, including bridge drivers, especially if you use a micro.

BOL-Gary

PS - OOOOOPPPPSSS!!!!!  I just realized that the trailing edge of either switch will look just like the leading edge of the other if you do the circuit as I described, generating a false direction change - c**p!  It's out, but the lowpass still works, if you want to go that way.  Just make sure that it will be impossible for the feedback counter to change state on a direction change BEFORE the DIR is correct.  Sorry.