Buffers or line drivers

[engr_joni_ee]

We have 8 x GPIO from Microcontroller which need to be connected at a remote location which is around 8 to 10 meters. The load current is something around 25 mA per GPIO. Kindly suggest line buffer/GPIO buffer that can be used with the Microcontroller board which we need to design. It is important that there should be no I2C. Just a simple parallel in parallel out.

 

[std_match]

Load voltage? High-side or low-side switching? Inductive load?

 

[engr_joni_ee]

0 V low
3.3 V high

 

[BradtheRad]

The setup sounds vaguely similar to Ethernet equipment, in the sense the cable has several wires, uses a few twisted pairs and a few not twisted. Ethernet carries data several feet and you need to carry data several feet.

I guess you don't need ethernet hookup nevertheless you can test whether twisted pairs (differential signaling) are any more reliable than single-ended shielded wiring. It depends on how much interference you're exposed to for 30 meters distance.

Buffer IC's include cmos 4050 (non-inverting) or else strings of 4049 4069 inverters. Attach a few in parallel if you need a wire to deliver higher current. Investigate 3.3 V devices to see what's available today.


Place the buffer IC's at each end. Why risk your microcontrollers? Send from a micro to a nearby buffer. Receive 30 meters away through a bit of waveform conditioning to a buffer.

I suppose differential signaling requires twice as many parts and twice as many tests for continuity.

--- Updated Jan 22, 2025 ---

Carrying load current is a job for wires of course. Do you wish to send data back from the remote location? Did you rule out home-brew wireless transmission via light or laser or fiber, or radio?

 

[barry]

The setup sounds vaguely similar to Ethernet equipment, in the sense the cable has several wires, uses a few twisted pairs and a few not twisted. Ethernet carries data several feet and you need to carry data several feet.

I guess you don't need ethernet hookup nevertheless you can test whether twisted pairs (differential signaling) are any more reliable than single-ended shielded wiring. It depends on how much interference you're exposed to for 30 meters distance.

Buffer IC's include cmos 4050 (non-inverting) or else strings of 4049 4069 inverters. Attach a few in parallel if you need a wire to deliver higher current. Investigate 3.3 V devices to see what's available today.


Place the buffer IC's at each end. Why risk your microcontrollers? Send from a micro to a nearby buffer. Receive 30 meters away through a bit of waveform conditioning to a buffer.

I suppose differential signaling requires twice as many parts and twice as many tests for continuity.

--- Updated Jan 22, 2025 ---

Carrying load current is a job for wires of course. Do you wish to send data back from the remote location? Did you rule out home-brew wireless transmission via light or laser or fiber, or radio?

CD4050 is a terrible choice, it’s well over 30 years old. There are MUCH better options. Further, the CD4050 can only source 3 mA. You’d need 8 of them in parallel (1.3 packages) per signal.

A 74ACT244 has 8 buffers with +/-75mA output current with a 5V supply. 74LVC244 will provide 24mA, if you need a 3.3v supply.

 

[BradtheRad]

Very well, scratch the CD4050 suggestion. There's nothing wrong with using modern IC's assuming one can solder those tiny tabs. Or unsolder them in case one has to replace the IC.

 

[FvM]

Signal speed wasn't yet mentioned, but it's important. Also if cable uses impedance matched, individually screened wires. If not, you'll expect reflections and crosstalk between individual signals, requiring bandwidth limited drivers or filters.

Generally speaking low voltage, single ended logic signals aren't well suited to be sent over 10 m cable, common mode noise can easily disturb logic levels.

 

[engr_joni_ee]

The application is that we need to control Relays which are located remotely on Relay board. The switching of Relay is slow. The Relay boards are located around 8 to 10 meters away from Microcontroller. Therefore, we are looking for single ended I/O buffer IC.

I will look at 74ACT244 and 74LVC244.

 

[KlausST]

Hi,

and the relay is driven by 0V LOW and 3.3V HIGH?

I´ve rarely seen relays for 3.3V ... and they usually are not driven push AND pull.

***
And no LOW level is specified by a single value "0V" ... but usually with a valid RANGE. The same is true for HIGH levels.

*****
And why don´t you go the most straight forward way:
If you want to drive relays ... why not simply do a search for a relay driver IC?
And if that is not what you are looking for ... then tell us why not.

Is it really asking too much?

Klaus

 

[barry]

Hi,

and the relay is driven by 0V LOW and 3.3V HIGH?

I´ve rarely seen relays for 3.3V ... and they usually are not driven push AND pull.

***
And no LOW level is specified by a single value "0V" ... but usually with a valid RANGE. The same is true for HIGH levels.

*****
And why don´t you go the most straight forward way:
If you want to drive relays ... why not simply do a search for a relay driver IC?
And if that is not what you are looking for ... then tell us why not.

Is it really asking too much?

Klaus

Right.

Or, maybe you'd want to use an open-collector/open-drain driver or transistor array.

OP, if you had initially stated that you were trying to drive relays, you'd save a lot of wasted time.

 

[engr_joni_ee]

Right.

Or, maybe you'd want to use an open-collector/open-drain driver or transistor array.

OP, if you had initially stated that you were trying to drive relays, you'd save a lot of wasted time.

I realise it. Thanks.

 

[BradtheRad]

3.3V could droop unacceptably over wire runs of 60 meters round trip. Consider using voltage multipliers to increase voltage to relays or specifically, relay coils. Efficient is better by pushing greater volts at fewer Amperes.

 

[dick_freebird]

I think that line length wants controlled impedance
(like shared CAT5 cables could be good for RS-422,
RS-485 or LVDS). Noise environment and channel
switching period would drive selection of standard.
RS-422 parts for 3.3V supply are few or none,
although a bipolar type would work OK albeit
reduced level / noise margin. You'd need a
corresponding RX at the other end if you want
the noise / common mode rejection attributes.
Which give 30' distance, might well be significant.

 

[barry]

I think that line length wants controlled impedance
(like shared CAT5 cables could be good for RS-422,
RS-485 or LVDS). Noise environment and channel
switching period would drive selection of standard.
RS-422 parts for 3.3V supply are few or none,
although a bipolar type would work OK albeit
reduced level / noise margin. You'd need a
corresponding RX at the other end if you want
the noise / common mode rejection attributes.
Which give 30' distance, might well be significant.

It's a relay.

It doesn't need controlled impedance. It doesn't need differential signaling. It doesn't need noise margin. It doesn't need common mode rejection. It doesn't need a receive at the other end.

 

[D.A.(Tony)Stewart]

A good choice is opto-isolated 8x relay card online that can be driven from low current and voltage with 3.3V logic and adding 5V isolated supply
for coils.

or the non-isolated with TTL level to Darlington 8x ULN2803 with 12V supply for coils. https://www.electronics-lab.com/project/8-channel-relay-board/

The main requirement is to isolate the flyback pulse from the uC circuit with a common diode to +12 and the relays isolate from the load but the +12 common is shared with 3.3V so routing must be done to avoid ground shift from shared currents.

 

[FvM]

It's a relay.

It doesn't need controlled impedance. It doesn't need differential signaling. It doesn't need noise margin. It doesn't need common mode rejection. It doesn't need a receive at the other end.

Yes, looks like. But we didn't yet hear clearly if it's a 3.3V relay directly powered through 10 m cable. Stated 25 mA load current suggests this, although 80 mW coil power refers to a tiny reed relays only. Could be so easy by starting the thread with a clear description.

 

[KlausST]

It's a relay.

It doesn't need controlled impedance. It doesn't need differential signaling. It doesn't need noise margin. It doesn't need common mode rejection. It doesn't need a receive at the other end.

I agree ...

but with the lack of information .. I would not be surprised if at the receiver point there is a power supply for the relays .. in a way that the "signals" don´t need to carry the relay current.

But what do we know?

Sorry for the honest words ... but in my eyes the OP is too lazy to post clear relay specifications, exact relay type, or a link to it´s datasheet.
And a simple hand drawn sketch of what he wants to build. I guess it would take less than a minute.
It would speed up overall solution, avoid lengthy threads .. we don´t need to guess.

Klaus