Simple phase sequence detector using NAND gate

[vis5254]

I was able to get a circuit for this application searching a lot .
Does this circuit looks legit .
All I required was to have LED on when phase sequence was RYB and Led off when any another phase sequence arrives !
PNP - 2N4121
GATE IC - CD4011B

[soldar]

A very quick look. Gates D2 and D3 form a flip-flop bistable which is set by phases 2 and 3. If then comes phase 1 gate D4 changes to state but if before phase 1 comes the flip flop is reset by L2 then D4 remains. In principle I see how it could work and it is easy enough to breadboard.

Be carefull with the three inputs. What voltages are we talking about?

[Ian.M]

That circuit wont do what you want.  It is a phase sequence detector if you build it as far as the gates, that will output a low going pulse for approx 60 deg of the input cycle if all phases are present and L2 lags L1 and L3 lags L2, and if any phase is missing or the sequence is different, outputs a steady high level, but the transistor and LEDs part of the circuit is entirely bogus, with VL2 steadily illuminated and VL1 varying in intensity slightly due to the gate output (and passing far too much current for a 30mA LED).

LTspice sim attached.

There's also the issue that it uses Neutral as Vss for the logic circuit, and its input phases aren't isolated.  That's acceptable if its *JUST* a box on the wall with LEDs protected by lensed lampholders so there is no risk of contact, and its own isolated internal 12V supply, but is unsafe if its output goes to other low voltage circuits.

[rstofer]

If you look around, you can find circuits that use only passive components and don't require an external source of power.

https://www.edn.com/design/power-management/4318736/Phase-sequence-indicator-uses-few-passive-components
https://m.eet.com/media/1129096/12890-figure_3.pdf

I used to do this stuff for a living.  I use the Knopp K-3 Phase Sequence Indicator

https://www.amazon.com/Knopp-K-3-Phase-Sequence-Indicator/dp/B07M687W4Y
 

[soldar]

That circuit wont do what you want. 

A tangent if I may. I have no formed opinion on that circuit and do not want to spend any time analyzing it but I am always very hesitant and very careful before I would say something posted by someone doesn't work because I figure whoever posted it had probably tested it and maybe it was just that I was not getting the idea.

Anecdote: Some years ago I was a newly arrived poster at some electronics forum and I posted a circuit. A certain old time member promptly dismissed it saying it could never work and he gave his analysis proving it. Even though I explained I had built it and knew many people who had built it and it worked fine the entire board supported the old member telling me, the newbie, how much the old man knew and how I could not hope to be at his level. The old member was adamant and nobody bothered to do their own analysis. This went on for a while until, I forget for what reason, maybe someone said they had built it, the case is that the old member admitted he was mistaken and the circuit I had posted was good.

At that point, everybody, and I mean everybody, turned to the old man and grovelled saying how he was "a class act" for admitting he was mistaken, etc. Not a word to me, the newbie, who had put up with so much shit.

I forget the forum and it does not matter anyway but it had a very cliquish atmosphere and I guess I did not last long. Newbies were definitely not welcome.

[Ian.M]

The O.P specifically asked "Does this circuit looks legit"?

Its the absence of a base resistor for the PNP transistor that should clue you in that the O.P's found circuit may have problems.  Combine that with the 100R load on the CD4011B final output, and looking at datasheet figs. 5 & 6 'Output High (source) characteristics', its obvious that if you want the transistor off when the CD4011 output is high, it must source well under 2mA.   That means the Vf of VL1 must be at least 11.2V (12V supply - (0.6V Vbe + 0.2V drop across R7)).   VL1 must therefore be a compound LED consisting of a series string of dies, which is typically only done where high intensity is required, and running it at under 2mA is unlikely to produce sufficient illumination.

Also it is immediately obvious that the output must pulse if it indicates anything, as the NAND gate after the flipflop has one input fed from phase L1, forcing the output high for a minimum of 50% of the period.


A circuit can be bad even if one working unit has been built.  e.g. if it takes no account of component tolerances so will only work if you are lucky or if you select for particular tolerance extremes.

If you can at any time prove me wrong, I'll thank you for doing so.

[soldar]

If you can at any time prove me wrong, I'll thank you for doing so.

No, as I said, I have not spent any time looking at that circuit.

I was making just a very general comment about why I am hesitant to say something won't work even if I am pretty sure. That's all.

[ArthurDent]

...."I use the Knopp K-3 Phase Sequence Indicator

https://www.amazon.com/Knopp-K-3-Phase-Sequence-Indicator/dp/B07M687W4Y"

Here's the same type of phase sequence indicator on eBay new for 1/6 the price.  It looks like they were made by the same company. These have what amounts to a very small 3-phase motor inside so are very simple. Which way the 'motor' tries to turn indicates the phase sequence. This type indicator should last forever. Note that older ones like this used a different color code for the phase lines.   

https://www.ebay.com/itm/MILITARY-GENERATOR-QVS-1022-60HZ-3-PHASE-SEQUENCE-INDICATOR-TESTER-110-550-VAC/202682630574

https://www.quora.com/In-a-three-phase-system-three-wires-of-colors-red-yellow-and-blue-are-used-to-discriminate-each-phase-Now-how-will-you-identify-each-phase-wire-without-knowing-the-color-and-without-any-pre-information

[soldar]

Here's the same type of phase sequence indicator on eBay new for 1/6 the price.  It looks like they were made by the same company. These have what amounts to a very small 3-phase motor inside so are very simple. Which way the 'motor' tries to turn indicates the phase sequence. This type indicator should last forever. Note that older ones like this used a different color code for the phase lines.   

Sounds like a pre-electronics solution. Today it should be easy to make something electronic for little cost. 

[schmitt trigger]

I am subscribing to this thread, as I built a functional phase sequence detector many moons ago,  but can’t find the schematic right now.

If I find it I will post it, if I recall correctly I had to use a retriggerable monostable in addition to the gates to get rid of the output pulses as mentioned above by Ian.

And it was optocoupler isolated, too

[soldar]

Yes, it should not be difficult to design one.

Say the phases are A, B and C and each one generates a short pulse say at the top of the positive half cycle.

Pulse from A sets a flip-flop that enables pulse B to light an LED. Pulse C resets the flipflop.

If after pulse A comes pulse B it will light the LED.

If after pulse A comes pulse C the flip flop will be reset when pulse B comes around and the LED will not light.

I think this basic idea would work.

[Zero999]

The circuit posted at the start of the thread can be easily fixed with the addition of an emitter follower, a base resistor and higher value current limiting resistors.

[soldar]

I believe the common point of R4, R5, R6 should be also connected to GND.

In the diagram it is connected to the Neutral which is not strictly necessary but doesn't hurt either.

[Zero999]

I believe the common point of R4, R5, R6 should be also connected to GND.

In the diagram it is connected to the Neutral which is not strictly necessary but doesn't hurt either.

Yes, you're right. The neutral should be the same as the circuit ground.

[schmitt trigger]

I found the circuit. see attached.

It is essentially identical to the circuit above, but it is optocoupled for safety. Resistor R, is selected for a peak current of 10 mA.
As the optocoupler inverts the signal, a second transistor is required.

EDIT: OOOOps! I forgot to include the antiparallel diode across the optocoupler input for phase A.

[soldar]

Good find! The optocouplers are definitely a good idea although not needed if the device is properly isolated. But better safe than sorry.

OTOH, you could probably easily design it to take the power from the wires themselves.

I really have no need for this because I do not use three phase but if I did I would definitely put this on my list of projects.