hi all, some of you have helped greatly with my ongoing project of a complex water level switch. With this help I have managed to add a variable delay on the relay de-energise circuit, add extra indicator leds, as well as work on EMC preparation.

There is one thing that would finish off the unit perfectly. A remote display option.

At present 4 daughterboards feed a local display board with 3 leds per daughter board. The latest arrangement has 1 green led for water, 1 red led for steam and 1 amber led for fault. If a daughterboard is set to be normally sensing water and it is sensing water, the green led will be lit only. If the board is set to be normally sensing water but is sensing steam then the red led will be lit and the amber led will be lit.

So there is a total of 12 leds. Each daughterboard connection to the local display has 12v+ connection and one 0v connection per led. This gives 16 connections to local display. I am trying to figure out the best way to replicate the display remotely without having to run a 16 core cable. The remote could be 500m away. The solution would ideally be cheap to implement and only use 4 to 6 core cable as well as being EMC friendly.

Any ideas would be great.

 

Food for thought. DTMF. The same thing used for touch tones. You have 16 digits available. Probably messy.
Any way: A start DTMF, the x-y or LED to be on, a STOP DTMF and a previous/current state sort of thing.

A serial link like RS485.

Take a look at www.picaxe.com.

 

+1 on the RS485 you could incorporate a Modbus type protocol.
Max.

 

Modbus seems like a lot of work and complexity to move one byte at ultra low speed.

Since the red and green LEDs are mutually exclusive, that is one bit. The fault LED is the second bit. Times four is only 8 bits for 12 LEDs. This can be encoded onto 4 wires plus grounds with a few gates, so the cable run would be a single piece of CAT-3 or better. No microcontrollers, no coding.

ak

 

I had thought about the RS485 but was not sure what was involved both cost and space wise.

I do like AnalogKid's idea

I would be very grateful if you could give me more of an idea of how this would work and with what components.

Thanks guys

 

Is there ever a case where both the red and green LEDs are off at the same time?

ak

 

Is there ever a case where both the red and green LEDs are off at the same time?

ak

Hi, no never, or on at the same time (that is per daughter board) there will only ever be red on, red and amber on, green on or green and amber on. At any time either the green or red will be on.

 

In that case, post #4 should work.

 

In that case, post #4 should work.

Thanks. Can you give me some pointers on where to start please.

 

I have been looking at this today and have come up with the following general circuit per set of 3 leds. It only needs one wire from green led and one from Amber per set. When the green led is off on the local display then there is 12V at the connector, This 12V feeds a NOT gate on the remote, the green led on remote will be off. The same feed from the local display also if feed to another N channel which switches the red led 0V. The amber LED uses another NOT gate.

All seems to work well. The 1K resistors are just for test and values will be refined.

 

Using a FET to "short out" an LED will work, but is not an efficient design. In fact, the system current increases when the green and amber LEDs are off. Also, you don't need the 1K resistors in series with the green and amber LEDs.

Here is an alternative red/green driver that uses only one transistor and does not waste current. This is from another thread; for your application replace the 555 with the green LED signal input. The combined Vf of D2 and D3 is more than that of D1, enough so that when Q1 is on, D1 "shorts out" D2D3. Rearrange the LEDs for correct signal polarity.

ak

 

Thanks..... I will give it a try.

You sa

Using a FET to "short out" an LED will work, but is not an efficient design. In fact, the system current increases when the green and amber LEDs are off. Also, you don't need the 1K resistors in series with the green and amber LEDs.

Here is an alternative red/green driver that uses only one transistor and does not waste current. This is from another thread; for your application replace the 555 with the green LED signal input. The combined Vf of D2 and D3 is more than that of D1, enough so that when Q1 is on, D1 "shorts out" D2D3. Rearrange the LEDs for correct signal polarity.

ak
View attachment 118152

Thanks ..... I will give it a try. Before you said that the 12 leds could be encoded over just 4 cores. I just don't see how this could be done. All I can work out is 2 cores per 3 led set plus one core for 12v and one for ground, 10 in total.

 

Before you said that the 12 leds could be encoded over just 4 cores. I just don't see how this could be done. All I can work out is 2 cores per 3 led set plus one core for 12v and one for ground, 10 in total.

Multi-level bit coding. If red/green is 0 V / 1 V, and amber is 0 V / 2 V, and those are summed in an opamp acting as a wire driver, the output is 0, 1, 2, or 3 V. One LM358 dual opamp as a receiver recovers both bits from 1 wire plus GND.

a.n.a.l.o.g

ak

 

One LM358 dual opamp as a receiver recovers both bits from 1 wire plus GND.

I puzzled over this for a bit, but can't see how to decode 3 levels with just two op amps.

 

Here's an LTspice simulation of a decoder for 3 analog levels as proposed by AK, using one LM324 quad op amp package and one CD4070 CMOS XOR package, that I think will do what you want.
Since the supply is 12V, I used trip values of 3V, 6V, and 9V for the three levels to minimize noise sensitivity over the 500m cable.
If noise is still a problem, you can add an RC low-pass filter at the Input to this circuit.

At the transmit end you can use an op amp to generate levels of 4.5V, 7.5V, and 10.5V for each state to be in the middle of the detection windows.

 

Here's an LTspice simulation of a decoder for 3 analog levels as proposed by AK, using one LM324 quad op amp package and one CD4070 CMOS XOR package, that I think will do what you want.
Since the supply is 12V, I used trip values of 3V, 6V, and 9V for the three levels to minimize noise sensitivity over the 500m cable.
If noise is still a problem, you can add an RC low-pass filter at the Input to this circuit.

At the transmit end you can use an op amp to generate levels of 4.5V, 7.5V, and 10.5V for each state to be in the middle of the detection windows.

View attachment 118234

Thankyou very much for your efforts on this. I will weigh up the options and space required.

 

First pass at a receiver. I'd use an LM358 but there isn't one in the LTS library. I'm not the LTS wiz Wally is, but I think this captures the idea. The red trace is low for red LED on, high for green LED on. The blue trace is low for amber LED off, high for amber LED on. Here are the outputs based on the input (green) voltage. Of course, all of this can be scaled for a different supply voltage, different transition levels, etc.

0.0 V - 0.7 V - red on, amber off
0.7 V, 1.4 V - green on, amber off
1.4 V - 2.1 V - red on, amber on
2.1 V and up - green on, amber on

ak

 

First pass at a receiver.

Nice simple circuit and you did do it with just two op amps.

 

First pass at a receiver. I'd use an LM358 but there isn't one in the LTS library. I'm not the LTS wiz Wally is, but I think this captures the idea. The red trace is low for red LED on, high for green LED on. The blue trace is low for amber LED off, high for amber LED on. Here are the outputs based on the input (green) voltage. Of course, all of this can be scaled for a different supply voltage, different transition levels, etc.

0.0 V - 0.7 V - red on, amber off
0.7 V, 1.4 V - green on, amber off
1.4 V - 2.1 V - red on, amber on
2.1 V and up - green on, amber on

ak
View attachment 118256

Thanks so much for the help. Please stick with me in trying to get it going. At the moment I have a setup working that uses 10 cable cores. You said it could be done with less. The circuit above is the receiver, so on the remote display.

I am not sure how to implement the circuit, I guess firstly how do I get the 0v or 12v signals from the local display to the remote? Secondly, how to get the outputs to power the leds? I assume that I would need to use fets and the outputs from your circuit a tivate them.

Sorry if I am being dumb but this is a bit learning curve for me.

 

I am not sure how to implement the circuit, I guess firstly how do I get the 0v or 12v signals from the local display to the remote? Secondly, how to get the outputs to power the leds?

Below is AK's circuit slightly modified to give higher voltage trip points for greater noise immunity and with added LEDs to show how they would be illuminated.

The first half of the simulation shows the trip points of 3V, 6V, and 9V.
The second half of the simulation shows the LEDs lighting for the nominal signal levels of 0V(Red), 4.5V(Green), 7.5V(Amber and Red) and 10.5V(Amber and Green) from the transmitter.

This In signal (simulated here by V2 and V3) would be transmitted over two wires (signal and common) from the transmitter.

If you need more LED current than about 2mA as shown (which should be sufficient for high brightness type LEDs), you will need to add transistor drivers to the op amp outputs.

I show the quad package LM324 op amp but, since only two op amps are needed, you could use the dual LM358 op amp instead.