3 Phase Sequence Indicator 220V

[Apofview]

Hi guys,

I have this old three phase sequence indicator.





Schematic is in attachment. (made with digikey online app)


Note: uploaded corrected schematic

Bonus photo:


Can someone please explain how this works, it is different from other online schematics. (solved)

I would like to replicate but with modern parts, and to try to miniaturize.
Can this be rebuilt with modern through hole  parts like  small neon's, small resistors 0.5w, modern capacitors?

[NiHaoMike]

Just some wizardry with RC time delays such that the voltages add with one phase order and subtract with the opposite phase order. It would be easier to understand if you consider the circuit for only one of the bulbs.

[Apofview]

Can I use ceramic disc capacitor 103 3kV instead of axial?

And if there is RC time constant, I probably couldnt use standard resistors with 10% tolerance...
That would explain 1% tolerance on resistors...

[TimFox]

According to your photo, the capacitors have (normal) tolerance of +/- 10%.  The time constant is the product of R x C.
However, 1% resistors are not expensive.
Ceramic capacitors vary drastically depending on material.
For good results in critical analog circuits, you should use only NP0 = C0G material.
Your 3 kV discs may well be X7R, Y5V, or (heaven forfend) Z5U, which are very nonlinear at large voltages.
Another useful choice (although physically larger) is polypropylene film, readily available at this capacitance and voltage rating.

[Apofview]

Thank You for advice, I did notice polypropylene caps in search results.

[Apofview]

PCB should be optimized for new components, searched for online auto-routing tools, not for electronic hobbyist...
So maybe someone will find this schematic interesting or maybe useful.
Thank You.

[Apofview]

Just some wizardry with RC time delays such that the voltages add with one phase order and subtract with the opposite phase order. It would be easier to understand if you consider the circuit for only one of the bulbs.

While analyzing noticed error in schematic, will correct and upload:

corrected version

[Apofview]

Now with correct scheme it is easier to understand how it works.
Also at my local electronic parts store there is no 200k and 400k resistors...

[TimFox]

What about 180 k, 220 k, or 390 k?
Do they have 100 k (put a few in series)?
Remember that the tolerance on the capacitors in your photo is 10%.

[Apofview]

What about 180 k, 220 k, or 390 k?
Do they have 100 k (put a few in series)?
Remember that the tolerance on the capacitors in your photo is 10%.

They had 100k 1/4w 10%, resistors from image are 1% if im not wrong?

Also did check time constant:
t=R*C
"RC time constant - this figure — which occurs in the equation describing the charging or discharging of a capacitor through a resistor — represents the time required for the voltage present across the capacitor to reach approximately 63.2% of its final value after a change in voltage is applied to such a circuit."
For 50Hz system full cycle duration is 20ms, half cycle is 10ms...

Right neon is "connected"  R and S, and left neon is "connected" to R and T line.

O, I get it :-)
only neon on 120 degrees lines works:
- if order is RST, neon on R-S sees 120 degrees and is lit, other neon is R-T so 240 degrees and not lit.
- if order is RTS, neon on R-S sees 240 degrees and is not lit, other neon is R-T so 120 degrees and is lit.

For RST following right neon schematic, between R-S there is connected  600kOhm and 20nF and time constant is 0.012s or 12ms.
For RST and time constant 12ms R-line is at peak and S-line is approximately at zero.

For RTS following right neon schematic and for given time constant of 12ms R-line is at peak and S-line is also at peak so right neon could not be lit for reverse rotation.

[CaptDon]

That circuit is very tolerant of absolute value drift!! Look at the ionization voltages of NE-51 or NE-2 lamps, they may have easily a 30% trigger point difference in just 3 different samples and probably up to 50% in total darkness! The coolest phase rotation meter I ever saw looked like a pocket watch with a sweep second hand. It had 3 leads marked of course A, B and C. Depending on phase rotation the hand would run C.W. or C.C.W.  Here is another cool thing about phase reference, One of the factories where I worked had a low power transmitter modulated by the sinewave of the 'A' phase. With a dual trace scope and the reference receiver driving channel 1 of the scope you could touch the tip of a 100X probe to any phase in the building and determine if it was in phase (A), lagging 120 degrees (B) or appearing to lead by 120 degrees (the prior C). We had a lot of three phase stuff marked with Brown, Orange and Yellow generally used on 480. 208 three phase was generally Red, Blue and Black. Red, Black and Orange was strictly reserved for 'Wild Leg' 240 where you had Red=120, White=N, Black=120, Orange=208 but you still got 240 across the corners.

[ajb]

The coolest phase rotation meter I ever saw looked like a pocket watch with a sweep second hand. It had 3 leads marked of course A, B and C. Depending on phase rotation the hand would run C.W. or C.C.W. 

On the other end of the complexity spectrum, we bought a rotary screw air compressor where the phase indicator was a piece of paper taped on one edge over the radiator exhaust.  Instructions advised to immediately turn the compressor off and correct the phasing if that piece of paper wasn't blown up by the radiator fan when it started up!

[TimFox]

What about 180 k, 220 k, or 390 k?
Do they have 100 k (put a few in series)?
Remember that the tolerance on the capacitors in your photo is 10%.

They had 100k 1/4w 10%, resistors from image are 1% if im not wrong?

Also did check time constant:
t=R*C
for 10000pF -> 10nF and for:
- 200kO time constant is 0.002s or 2 ms
- 400kO time constant is 0.004s or 4 ms
- 600kO time constant is 0.006s or 6 ms
ever*circuit.*om is not free anymore...

Full cycle is 20ms, half cycle is 10ms...

Right neon is "connected"  R and S, and left neon is connected to R and T line.

O, I get it :-)
only neon on 120 degrees lines works:
- if order is RST, neon on R-S sees 120 degrees and is lit, other neon is R-T so 240 degrees and not lit.
- if order is RTS, neon on R-S sees 240 degrees and is not lit, other neon is R-T so 120 degrees and is lit.

Note that, at 50 Hz, the reactance of 10 nF is -318 k , which is comparable to the resistance value at these nodes.
I didn't know they still made 1/4 W 10% resistors:  I haven't seen anything worse than 5% recently.
You didn't indicate your country, but Mouser has cheap Yageo 200 k, 1/4W, 1% metal film axial-lead resistors for $0.075 (USD) each in quantity 10.  158 V rms across 200 k is 1/8 W dissipation.

[Apofview]

I got it all figured out, all sums up, so I must be wrong 😁
But if I'm right, any RC combination with time constant of 12ms will work.
Use of multiple resistors is probably to increase voltage tolerance and reliability of resistors.

[Apofview]

Note that, at 50 Hz, the reactance of 10 nF is -318 k , which is comparable to the resistance value at these nodes.
I didn't know they still made 1/4 W 10% resistors:  I haven't seen anything worse than 5% recently.
You didn't indicate your country, but Mouser has cheap Yageo 200 k, 1/4W, 1% metal film axial-lead resistors for $0.075 (USD) each in quantity 10.  158 V rms across 200 k is 1/8 W dissipation.

I'm from Montenegro, didn't noted that I did not set up country, sorry, corrected that.
We have only one electronic store left, assortment is fairly limited. My hobby collection is almost comparable - have sets of resistor, capacitors, smd... But always something is missing. I'm still thinkin is it worth waiting, and will this end up on pile of "next" project, but really had fun with this device.
Would like to make some smaller footprint board, but no online autorouters, and KiCAD is another topic...

[Apofview]

That circuit is very tolerant of absolute value drift!! Look at the ionization voltages of NE-51 or NE-2 lamps, they may have easily a 30% trigger point difference in just 3 different samples and probably up to 50% in total darkness! ...

Hm, original circuit total resistance between lines are 3x200kOhm or 600kOhm, and with 2x10nF cap RC is 12ms,
if i change that to 3x220kOhm, with same value for caps of 2x10nF, time constant changes  to 13.2ms, maybe it will work...
Other connection is 400+200+400=1MOhm, if I replace 400kO with 390kO new resistance would be 390+220+390=1MOhm.
This could work...

1/4W and 1/2W resistor rating is 200V, will this be ok for resistors, there will be two in series but still it will be 400VAC across this two...

[TimFox]

Note that those changes due to resistor values are probably smaller than those due to the 10% tolerance on the capacitors in the final mathematical products.

[floobydust]

The Philips indicator tubes appear to be gas discharge "surge arresters" which is why they are quite hefty.
I don't know OP's part number but 4349 is the lowest voltage 130-180VDC ignition, 100VDC extinguish voltage.
Ref. Philips Electronic Tube Division PocketBook 1954, pdf pg. 231/234 (but slowww 45MB download)

[Apofview]

The Philips indicator tubes appear to be gas discharge "surge arresters" which is why they are quite hefty.
I don't know OP's part number but 4349 is the lowest voltage 130-180VDC ignition, 100VDC extinguish voltage.
Ref. Philips Electronic Tube Division PocketBook 1954, pdf pg. 231/234 (but slowww 45MB download)

Original marking on neon tubes is "Phillips Holland Z11T".
Found this: "Most small neon (indicator-sized) lamps, such as the common NE-2, have a break-down voltage of around 90 volts."
Also, in random NE-2 datasheet for 220-250V and series resistor 540kOhm average usuful life 25000h.
So this make it a little bit more interesting.
Voltage across neon in this config might be 400VAC (line-line 230*1.73~400VAC), but on schematic there is some configuration which looks like voltage divider?

[Apofview]

First 'rats nest' ready, testing tomorrow.
My test box supports up to 3x115VAC, so first test will be with that voltages.

Update:
tested 3x115V AC, works as intended.
From below image Yellow-R, Red-S, and Black-T line
upper NE-2 right direction
lower NE-2 left dirrection (NE-2 salvaged from standard switch, resistors 0.5W, 220kO, 390kO)

[floobydust]

Neon lamp firing voltage is imprecise, it depends on ambient light, background radiation, electric field etc.  They are terribly moody.
The circuit reminds me Radio Shack "Goofy Lite" that had a sequential neon lamp blinker. If you brush touched two neon lamps, you could get the flash sequence to reverse.
http://www.netzener.net/index.php/project-articles/11-build-a-neon-goofy-lite
https://threeneurons.wordpress.com/nixie-power-supply/neon-bulb-chase-lite-kit/