Universal line driver for differential and single ended usage?

[Manx]

I need a line driver that could be used for both differential and single ended operation. All I can find are RS485 compatible drivers, that are not good for single ended operation, and RS422 drivers that would be OK concerning voltage levels, but their current is limited to 20 mA, and I'd like to have at least 50 mA.

I'm thinking of simply using two single ended line drivers, non-inverting and inverting. This would probably be the cheapest and would save me from spending further days looking for a suitable IC. However, I don't know how bad it would be if I do it in such way.

Output from my device will go into servo motor drivers. Normally it will be just driving a led, but I'm wondering what would happen if the servo motor driver had a "proper" differential receiver; how bad my non-standard approach be in that case. Also, I'm wondering what using two ICs as a substitute for a differential driver would do to EMI.

[T3sl4co1l]

What's wrong about RS-422/485 drivers for single-ended operation?  Just leave one channel open.

Could also wire together a few logic buffers in parallel, or use a small gate driver.  Neither of these are slew-rate controlled for EMI, of course (preferably, the signal should be carried in coax).

What do you mean driving an LED?  Like an optoisolator?  Or just for indication?  How many?  How fast?  Is this just low frequency PWM for RC servos?  Do you really need much current drive?

What distance, and what kind of cabling?  Do you need a common ground or no?

Tim

[Manx]

RS485 gives too big output low voltage. On the other hand RS422 gives only 20 mA.

Yes, I mean driving one led in an optoisolator. One microsecond rising/falling edge should be enough, I think, although I don't mind having it faster. It's for industrial servo drivers. I think 50 mA would give me some headroom just in case. One servo driver I checked takes 20 mA on an input; I don't want to be at the limit.

Distance max 2 meters. Need for a common ground will depend on the servo driver. I want to provide an universal output.

[SiliconWizard]

Uh huh, not too sure about what the problem is exactly either.

As an obvious side note, you could just have two separate output drivers, one differential, the other not (and possibly with higher current for the single-ended one if I got it correctly), both fed by the same digital input signal? They could go to separate connectors as well, as using a single connector for both modes could be confusing for the end-user.

Except if your design is ultra cost-constrained, really... I wouldn't bother doing it another way. Especially if you intend on driving standard RS-485 receivers, which may be what you're implying with "proper differential receiver". Better use standard drivers so they are conforming.

Just a thought, we don't know squat about your application, so...

[Manx]

It's for a motion controller.

[langwadt]

RS485 gives too big output low voltage. On the other hand RS422 gives only 20 mA.

many drivers are spec'ed as RS-422/485 the difference just the conditions for the spec


 

[T3sl4co1l]

Well heck, if you're just driving LEDs, why not a unipolar output stage, like a good old fashioned transistor?  NPN or N-CH.  Or P, doesn't much matter.  Can still have a pull-up/down resistor to keep it clean, or a smaller driver in the passive direction to give better drive with less current consumption.

This is symmetrical (and suitable for gate driving purposes), but shows a possibility:



IC1A could be replaced by a single common-emitter/source transistor (and no follower) to give strong drive in one direction, while IC1C still delivers bipolar drive.

I'd suggest at least a >= 300R ferrite bead off of this, for cable of that sort of length; and maybe a 100p-1n to ground after that, so you get some mild LC filtering right at the source.  And an ESD clamp diode (BAT54S?) would be nice too, just for belts-and-suspenders.

Tim

[Manx]

Thanks. But the idea is to (1) have a differential line, and in addition, (2) have an option to use it as single ended. I would be nice to make it single ended, but these things are done as differential lines, and I think it's a good idea to follow that. At the same time, I want it to be possible to use this as a single ended output, for "backward compatibility".

I searched many RS422/485 drivers and came up with nothing which would fit my needs. I'm a bit tired of this. So that's why I have this idea to use single ended drivers. I thought about 74ACT541 plus some inverting counterpart.

Thanks for the hints on filtering and ESD protection, as it more or less confirms that my thinking was going in a right direction. I was planning to use a 1 nF cap and SRV05-4 diode array. And then I was wondering if I should use a series resistor or a ferrite bead (like 742792097 1500 ohm one). I thought resistor would make it a bit safer, but I don't know what it would do to the signal (?).

[NiHaoMike]

What frequency do you need to operate at? If it's less than a few MHz, some MOSFET gate drivers might work well.