Driving multiple MAX22191 industrial inputs.

Thread Starter

G0JLB

Joined Sep 23, 2007
13
Hi,

I am looking at using the MAX22191 device as a digital binary input on a circuit / PCB Module
The MAX22191 is a Ultra-Low-Power, Parasitically Powered Digital Input which typically accepts a 24V input. I want to use this device to then drive an opto coupler and then in turn a MCU. I have a single prototype PCB made which is working perfectly on the bench.

But my question or concern is that I want to drive 32 of these devices from one 24V signal. So basically, 32 of these boards all about 5M apart all driven by on 24V signal. Each input has a TVS diode and should draw less than 3mA. So when the 24V signal switches on all devices or boards on the 'network' will see a logic high and when the 24V is switched off (0V) then the inputs will see a low.

Will driving 32 (Maybe more) of these devices over maybe 100M plus of 1.5mm2 cable be ok, or will the effect of having these all in parallel cause a problem. These IC's are a digital industrial input which I assume would normally only be driven by one sensor or industrial digital output such as a PLC output.

The next question is what is the best way to drive these. A simply relay switching a 24V supply may work but what about using a dedicated push / pull industrial digital output of some description or even a high side driver. But it would really need to switch to 0V for a low rather than just OC.

Any thoughts on this would be really apricated.

Thanks
 
Approximately at what max frequency will the 24V input be switching at?
Is your 24V input just a "turn on " switch or is there data?
I assume there isn't any data as you are talking about using a relay?
Is this an actual industrial controller i.e. what is its application?
 
Last edited:

Thread Starter

G0JLB

Joined Sep 23, 2007
13
Approximately at what max frequency will the 24V input be switching at?
Is your 24V input just a "turn on " switch or is there data?
I assume there isn't any data as you are talking about using a relay?
Is this an actual industrial controller i.e. what is its application?

Hi,
For the initial design their is no data, it is simply to turn on and off options to multiple modules remotely. Which may change once or twice each day. In fact we have 16 options so each remote module will have four MAX22191 circuits (4 inputs) then the master control box will simply output four 24V/0V signals depending on the selected options, This would mean that the control cable would be 5 core (4x signal and 1 ground). Ideally, what would be nice is a proper data signal of some sort but as time is limited its needs to be fairly simple, quick and robust. I was also looking at an isolated RS485 network as each module does not need to have a unique address and I can simply send a broadcast message to set the option. But it will take much to long implement and test. The selected option (1 of 16) on the master control box simply needs to tell all modules on the 'network' what option to set to and all modules will be set to the same option. But this is needs to be a reliable setup. I also have been looking at using industrial digital output chips to drive the network of inputs. Maxim have a high side / push pull output chip that looks interesting, but I am not sure if it could drive all 32+ inputs at the same time (MAX14900E) without problems, which is why I was looking at a simple relay output switching either 24V or 0V but this is not really idea.

So, really all need is an isolated method to send 1 of 16 options to multiple modules over a 100M plus cable. They need to be isolated as the PCB modules also contain a 3 phase supply. I was looking at isolated RS485 for this reason but all to stop ground loop issues.

Does that all make sense? Or have have I just confused the issue :)

Thanks
 

crutschow

Joined Mar 14, 2008
34,285
If all the outputs are isolated opto couplers, then I see no problem with driving all of them from one source, if that is your question.
I see no need for further isolation.
 

Thread Starter

G0JLB

Joined Sep 23, 2007
13
Thanks for your replies.

The inputs for each module will be a MAX22191 as in the diagram above, and its output will drive a optocoupler into a microcontroller. But of course the MAX22191 (Inc TVS) inputs will all be in parallel and connected by the long cable going from one to the other and drive by the master control box. So really my two questions are:

1. Will having all the MAX22191's inputs connected in parallel and all being driven together by the same source over a 100M+ cable (from the master control box) cause an issue. I assume the MAX22191 is designed to be used as one single industrial input.

2. Apart from a relay switching between 24v and 0v is their a better solution to driving all the multiple inputs together with some sort of digital industrial output from the master control box? Using relays seems a bit old tech. it would be nice to have a digital solution like the MAX14900E for example controlled from a microcontroller to drive the cable and all the inputs.

Thanks

John
 
Hi,

I am looking at using the MAX22191 device as a digital binary input on a circuit / PCB Module
The MAX22191 is a Ultra-Low-Power, Parasitically Powered Digital Input which typically accepts a 24V input. I want to use this device to then drive an opto coupler and then in turn a MCU. I have a single prototype PCB made which is working perfectly on the bench.

But my question or concern is that I want to drive 32 of these devices from one 24V signal. So basically, 32 of these boards all about 5M apart all driven by on 24V signal. Each input has a TVS diode and should draw less than 3mA. So when the 24V signal switches on all devices or boards on the 'network' will see a logic high and when the 24V is switched off (0V) then the inputs will see a low.

Will driving 32 (Maybe more) of these devices over maybe 100M plus of 1.5mm2 cable be ok, or will the effect of having these all in parallel cause a problem. These IC's are a digital industrial input which I assume would normally only be driven by one sensor or industrial digital output such as a PLC output.

The next question is what is the best way to drive these. A simply relay switching a 24V supply may work but what about using a dedicated push / pull industrial digital output of some description or even a high side driver. But it would really need to switch to 0V for a low rather than just OC.

Any thoughts on this would be really apricated.

Thanks
From what I see the Maxim is not designed with what you have in mind(The long cable) The 100 metre cable concerns me on several fronts. Cross talk, RFI and reflections. You are going to need signal conditioning before it reaches the Maxim end of the cable as there is going to be a bucketful of spikes during transition and RFI spikes if it's near a 3 phase installation. Again, I do not know what EM environment the cable is running through. At best I would suggest using individually screened twisted pair for each receiver. There are probably off the shelf Tx Rx combos, with fault tolerance, that can be used to do this stuff and you will only probably need a single twisted pair. I know it's only DC signalling at the moment, but can your processor end deal with the spikes it's going to receive from the optocoupler? You maybe alright, but normally ironing out problems in this sort of scenario will take weeks not days, but I do not have enough information to advise you and do not know what the safety issues are.
 
Last edited:
Ignore the relay shorting out the supply, but hopefully you can see what I mean!
Sorry if I'm being overly pessimistic. Whenever I hear the words 100m cable, I think yikes, common/differential mode RF interference reflections, crosstalk. I worked for a company designing, building and installing industrial controllers. Whenever we came across installations where we needed communication over a line, even 10 meters we would use a module designed for the job because it just wasn't worth the hassle. You have all sorts of unknowns, which these units are designed to cope with.
I hope you succeed with your endeavour and I'm very sorry I couldn't be more encouraging.
 

Thread Starter

G0JLB

Joined Sep 23, 2007
13
Hi,

Yes, these sort of things are my concerns. But as its not data, and just a switched dc signal and with the MAX22191 being fairly robust and able to handle +/- 60V inputs and all 4 inputs on each module will be stable at the same time I was hoping in may be ok. The MAX22191 are used for PLC input circuits and the like. This is also where sending data over a network may be better, like isolated 485 or something else altogether. I don't mind using a module on each PCB that is designed for his sort of thing but finding something simple (as in not lots of unrequired features) and low cost is an issue. Another option (Maybe) is to simply have 4 relays as the inputs with 230V AC coils and a 230V switched 4 core cable. I don't like that idea buts its a thought. If I had time I could develop and use something like a MODBUS or even DALI2 setup, which may work but it would just take too long to develop the code, porotype the PCBs and test in a real environment. At the end of the day all I need to do is select 1 of 16 options on 32 remote units all at the same time from a control box somewhere and maybe change the selection a few times each day. Back to the drawing board maybe.
 

crutschow

Joined Mar 14, 2008
34,285
Since you are just switching DC, I see no problem with using a MAX14900E or similar to drive the lines.
I don't see noise or ringing on the line as being of concern.
You could also use a DC output SSR for that purpose if you wanted to avoid mechanical relays.
 
Hi,

Yes, these sort of things are my concerns. But as its not data, and just a switched dc signal and with the MAX22191 being fairly robust and able to handle +/- 60V inputs and all 4 inputs on each module will be stable at the same time I was hoping in may be ok. The MAX22191 are used for PLC input circuits and the like. This is also where sending data over a network may be better, like isolated 485 or something else altogether. I don't mind using a module on each PCB that is designed for his sort of thing but finding something simple (as in not lots of unrequired features) and low cost is an issue. Another option (Maybe) is to simply have 4 relays as the inputs with 230V AC coils and a 230V switched 4 core cable. I don't like that idea buts its a thought. If I had time I could develop and use something like a MODBUS or even DALI2 setup, which may work but it would just take too long to develop the code, porotype the PCBs and test in a real environment. At the end of the day all I need to do is select 1 of 16 options on 32 remote units all at the same time from a control box somewhere and maybe change the selection a few times each day. Back to the drawing board maybe.
If they were ALL switching at the same time then all you would have to do is deal with the reflections. But if you have some active and some inactive then you are definitely going to get crosstalk with cheap cables.
Then that is something else you are going to have to deal with.
I did not work with PLCs, but if you could set it up to ignore small duration logic 1's then you maybe be OK. The other option is to use current sensing, which is what is used in the industry(4-20mA), which is less susceptible to RFI compared to voltage sensing.
 

GetDeviceInfo

Joined Jun 7, 2009
2,192
as long as your 24v can source the current, you shouldn't have any problems. the MAX22191 will shunt current until the switch point is reached, which should act as a buffer against noise, more so as the number grows.
 
Since you are just switching DC, I see no problem with using a MAX14900E or similar to drive the lines.
I don't see noise or ringing on the line as being of concern.
You could also use a DC output SSR for that purpose if you wanted to avoid mechanical relays.
Hi,

Yes, these sort of things are my concerns. But as its not data, and just a switched dc signal and with the MAX22191 being fairly robust and able to handle +/- 60V inputs and all 4 inputs on each module will be stable at the same time I was hoping in may be ok. The MAX22191 are used for PLC input circuits and the like. This is also where sending data over a network may be better, like isolated 485 or something else altogether. I don't mind using a module on each PCB that is designed for his sort of thing but finding something simple (as in not lots of unrequired features) and low cost is an issue. Another option (Maybe) is to simply have 4 relays as the inputs with 230V AC coils and a 230V switched 4 core cable. I don't like that idea buts its a thought. If I had time I could develop and use something like a MODBUS or even DALI2 setup, which may work but it would just take too long to develop the code, porotype the PCBs and test in a real environment. At the end of the day all I need to do is select 1 of 16 options on 32 remote units all at the same time from a control box somewhere and maybe change the selection a few times each day. Back to the drawing board maybe.
I'll bow out at this point. The other guys disagree, perhaps I'm being over cautious. When I worked on mobiles the radiation was enough to make opamps go beserk from metres away. I was the guy that had to deal with the demodulation of the RF into the phones audio and deal with radiated and conductive immunity. I'm just saying that you're building a 100m antenna and I have no idea what the EM environment is like nor do I know how safety critical your application is.
Try this, power the max22191 from a 5m lead with 24V supply.
Make a call on your mobile and place it next to the cable. Watch the output with the scope. Make sure the scope and lead is well away from the mobile! If it's ok then it's a pretty robust device.If you have an old GSM phone then that would be better as it's TDMA instead of the present CDMA. GSM is finally fading out! Best of luck and I hope your circuit works.
 

Thread Starter

G0JLB

Joined Sep 23, 2007
13
Morning all,

Sorry, more waffle.........

Thanks for your opinions and views on this. I will give the MAX22191 a go with maybe 10 modules all connected to long lengths of cable and see how I get on. The MCU on each module can look for a change on the four inputs and only act when they have been stable for x amount of time.

One other question which some of you may be able advice on is this.

What would be better a 5 core (Screened connected to earth at the source end), 4 core for the 4 switched 24V signals and one 5th core for overall signal ground (I prefer this due to cable cost and connector sizes) or 8 core with one signal and one ground for each output / input line going to all 32 modules?

I like the idea of the MAX14900E, but is their a lower cost or less complex alternative which can operate with push/pull outputs may be either single or quad channel but with some sort of smart control like the MAX14900E? I also like the idea of a SSD and I have used a high side smart SSD in a TO220 package before but of course a push pull o/p would be nice so when the 24V signal is off it actual switches to ground rather than left floating, so all of the 32 MAX22191 inputs have a definite 24V or 0V.
A simple MCU control (SPI, I2C, single I/O pins) that has protected push / pull outputs would be nice. If the the MAX22191 draws no more than around 3mA from the signal line (Parasitic power) then that's about 100mA max supply. With some margin maybe 250mA would be nice. The 24V power in the remote box would come from a high quality medical PCB PSU.

Lots to think about here,maybe I will get some test modules made and do some real testing over100M of cable.

I will also have a look around at 4-20mA control maybe. I am aware of this method but never used it for any projects.

Thanks again to everyone, these discussions are really helpful,and can steer projects in all sorts of directions :)
 
Morning all,

Sorry, more waffle.........

Thanks for your opinions and views on this. I will give the MAX22191 a go with maybe 10 modules all connected to long lengths of cable and see how I get on. The MCU on each module can look for a change on the four inputs and only act when they have been stable for x amount of time.

One other question which some of you may be able advice on is this.

What would be better a 5 core (Screened connected to earth at the source end), 4 core for the 4 switched 24V signals and one 5th core for overall signal ground (I prefer this due to cable cost and connector sizes) or 8 core with one signal and one ground for each output / input line going to all 32 modules?

I like the idea of the MAX14900E, but is their a lower cost or less complex alternative which can operate with push/pull outputs may be either single or quad channel but with some sort of smart control like the MAX14900E? I also like the idea of a SSD and I have used a high side smart SSD in a TO220 package before but of course a push pull o/p would be nice so when the 24V signal is off it actual switches to ground rather than left floating, so all of the 32 MAX22191 inputs have a definite 24V or 0V.
A simple MCU control (SPI, I2C, single I/O pins) that has protected push / pull outputs would be nice. If the the MAX22191 draws no more than around 3mA from the signal line (Parasitic power) then that's about 100mA max supply. With some margin maybe 250mA would be nice. The 24V power in the remote box would come from a high quality medical PCB PSU.

Lots to think about here,maybe I will get some test modules made and do some real testing over100M of cable.

I will also have a look around at 4-20mA control maybe. I am aware of this method but never used it for any projects.

Thanks again to everyone, these discussions are really helpful,and can steer projects in all sorts of directions :)
Check this document on cable configs.
https://www.emcstandards.co.uk/files/part_2_text_and_graphics_21_may_09.pdf
My opinion is to use Cat 5 with a twisted pair for each input.
I know, at the moment, you are running it as all on or all off, but that will change with your sixteen configs, hence a pair for each.
Make sure you slug the input as you don't want to have a problem with radiated emissions. Use an am radio to detect this.
Also the RC config mentioned, at the receiver end would probably be a good way of dealing with transitions.
Setting the PLC to ignore spikes hopefully will make this reliable.
 

Thread Starter

G0JLB

Joined Sep 23, 2007
13
Thanks, I thought of CAT5 or CAT6 (Shielded) but the connectors need to be IP66 rated (Maybe IP65).
I have been looking but they seem hard to come by. I have looked a Plenum CAT5 or 6 which looks more suitable.
 
Thanks, I thought of CAT5 or CAT6 (Shielded) but the connectors need to be IP66 rated (Maybe IP65).
I have been looking but they seem hard to come by. I have looked a Plenum CAT5 or 6 which looks more suitable.
I remember sitting in a lecture room when the lecturer, an actual humorous chap who would talk about data sausages, brought in transmitting data along twisted pairs.
We had a couple of students who were very close friends and sat at the front because they were very attentive. The lecturer turned to face them and said, "Talking of twisted pairs, how are you two doing".
They went very red. Yes, I know it's an old one, but it's still a good one in my book!!! I'll be very interested to hear what happens with your project, so, if you have time, tell us how it went.
 

Thread Starter

G0JLB

Joined Sep 23, 2007
13
Ah, I remember those days. I always sat at the back and was led astray by the the ones that didn't really want to be there. Turn out good in the end though.

Anyway,another question on this regarding Industrial Digital inputs (24V) using a device such as the MA22191 (Type 1 & 2 Inputs) is which is the best way to drive them. Either a high side or low side line driver (Industrial output chip) or one that has a push-pull output. I understand the difference, but with a high side driver, driving the MAX22191 at the end of a 100m cable run, the input to the MAX22191 will effectively be floating when the driving output is off. Whereas with a push pull line driver the MAX22191 will always see a one (24V) or 0v, but if its zero on the input, then potentiality could that result in a zero or ground loop that may then cause further issues, or would a high side driver be a better option?

Also, another thought on all of this is, and that my original question was about, driving multiple MAX22191 input devices from one line driver, so effectively 32 inputs spread over 100m of cables but be driven by one 24V signal. So another thought would be to only driver the first module (MAX22191) with the control box output over a fairly short cable run then the first module would 're-transmit' the signal via another line driver to the next module and so on. This way the total current would only be around 2.4mA and hence the volt drop would be less and the cables could be lower cost and smaller gauge (Like CAT5 for example). Plus in theory their would be no limit on the number of modules on the length of cable as each module would simply be driving the next module over 5m of cable., and possible have each line output 24V supply isolated.

Only downside is that if one module fails then the remainder on the network would not work, but a failed module could easily be linked out or bypassed.

Any thoughts? At the end of the day all I want to do is transmit 4 binary bits (Not data, just on/off) to multiple units (Maybe 32) over approx 100m of cable, but in a reliable and robust way.

Thanks
 
Top