Q: 24V Line driver IC

[fabhund]

Dear Forum.
I want to make a PCB for converting the single ended 5V output from the Par. Port to a 24V Differential line signal, and vice versa.

Is there any chip "sets" for this purpose. Looking for DIL packages, and with 4 drivers (Quad) in each package.

Best Regards, Bo Andersen.

[JohnHaine]

Hi Bo, when you say 24v differential, do you mean two complementary outputs which each swing 0 to 24v depending on the state of the 5v input? Try the L293 or SN754410. These are probably over specified for the purpose as they are designed for motor control type applications but should do the job. Fraid you only get two differential drivers per package though....

[Clive S]

Dear Forum.
I want to make a PCB for converting the single ended 5V output from the Par. Port to a 24V Differential line signal, and vice versa.

Is there any chip "sets" for this purpose. Looking for DIL packages, and with 4 drivers (Quad) in each package.

Best Regards, Bo Andersen.Most outputs from the p.p. are only about 3.3V nowadays ..Clive

[Robin Hewitt]

Do you have 24 Volts of do you want to squeak 5 up to 24?

If you have 24V how about and op-amp wired as a comparator. If it's a data signal go for a high slew rate.

[Jonathan]

Most outputs from the p.p. are only about 3.3V nowadays ..Clive

You can shift it to 5V with a small mosfet/transistor and a couple of resistors, like my impromptu 'breakout board' here:

http://www.mycncuk.com/attachment.php?attachmentid=14210&stc=1

Or you could go for something more sophisticated - e.g. add a schmitt gate. Or kill two birds with one stone and use and optocoupler and phototransistor instead of just a transistor.

[JohnHaine]

The SN754410 suggested above is compatible with low level CMOS voltages so should be ok with a modern pp. It also I think has in effect a Schmidt trigger input. I notice from the data sheet that it's going obsolete though RS have them in stock at about £1.40 each. It only does one way though, for 24v to pp I'd just use a single npn transistor with the 24v on the base through about 100k and a 4k7 pull-up to the pp Vcc in the collector. Remember it will invert the logic level.

[fabhund]

The SN754410 suggested above is compatible with low level CMOS voltages so should be ok with a modern pp. It also I think has in effect a Schmidt trigger input. I notice from the data sheet that it's going obsolete though RS have them in stock at about £1.40 each. It only does one way though, for 24v to pp I'd just use a single npn transistor with the 24v on the base through about 100k and a 4k7 pull-up to the pp Vcc in the collector. Remember it will invert the logic level.

Thanks for all the posts.

Case is that I already has a semi-working controlcabinet. A wood router based on stepmotors, controlled via puls/step from an ESS smoothstepper.

However I have a lot of noise (Due to improper shielding, and poor worksmanship, I know !)

Now I want to make it more correct and one of the things is to drive the stepper controller inputs (Optocouplers) from a line driver, using a shielded twisted pair cable.
Right now it's single ended connection (Common anode) and a straight, non shileded cable.

So from the ESS output, I want a Differential/Complementary Line driver, so 2 wires to each input.
In order to make it as immune as possible i want that to have 24V on the lines.
And the driver must be able to source/sink 20mA for the Photo-LED

I expect a min. of 500kHz pulse frequency.


For Limit switched i use Omron proximity sensors that work in 12-24VDC range. And is Sinking current (Ground) when activated.
Again this I want to drive in 2 wires as before, to a "reciever" chaninge the 24V line signal to the ESS input pin.

I have 5-12-24-48V available in my cabinet.

On top of the all the cable routing must be reworked, so stepper, sensor and spindle cables are routed separately.

Hope this draws a more clear picture of what I'm trying to achieve here.
Again thanks for all the replies, I have something to google now :-)

-Bo-

[m_c]

If you are going for differential pairs, then I wouldn't worry too much about what voltage the differential pair is running at. 5V is prefectly good for differential systems, as that's voltage what most modern CNC encoders operate at.
The key feature of differential pairs, is any noise that gets picked up, should be cancelled out by it affecting both wires by a similar amount and the differential voltage remaining reasonably constant. (This is a very basic explanation, which I'm sure Jonathon will criticise me for it!)
.
If you're running single ended, then yes, a higher voltage is better, which is why industrial machines nearly always run 24V control systems.

[Jonathan]

The key feature of differential pairs, is any noise that gets picked up, should be cancelled out by it affecting both wires by a similar amount and the differential voltage remaining reasonably constant. (This is a very basic explanation, which I'm sure Jonathon will criticise me for it!)

It's a very basic concept, so your explanation is good :)

There are plenty of differential driver ICs to choose from, I used these (specifically MAX490) in my servo motor drive and they would likely be fine for the application at hand:

http://www.maximintegrated.com/en/datasheet/index.mvp/id/1111

[Robin Hewitt]

If you are using opto couplers then your signal is current not voltage dependant and you are effectively proof against electrical interference.

If it isn't working you are barking up the wrong tree if you think screening and more volts is going to help.

[fabhund]

If you are using opto couplers then your signal is current not voltage dependant and you are effectively proof against electrical interference.

If it isn't working you are barking up the wrong tree if you think screening and more volts is going to help.

Guess I didn't pay too much attention to the cabling, screening etc. Fact is that I get faulty signals from endstops if I run them at 5V without opto. Now I changed that to 24V proximity sensors and an opto near the BOB to ground the input pin. Now that's working.

So the stepper drivers 5V opto inputs Enable/Pulse/Direction. Those are wired single ended with common anode.
Just figured to go all in and use Line drivers where I can, and make it as stable a possible.

The MAX490 chip looks perfect, if I read the specs. right. Then it can drive an opto directly (20mA) ?

Thanks again, Bo Andersen.

[irving2008]

How far from ESS to drivers? Unless it's more than a metre or so this sounds more like a ground loop issue than induced noise. Not convinced you're addressing the right problem.

[fabhund]

How far from ESS to drivers? Unless it's more than a metre or so this sounds more like a ground loop issue than induced noise. Not convinced you're addressing the right problem.

Thank you for your reply.
From the ESS to the BOB, i would say <75cm, and from the BOB to the driver <50cm

So the ESS drives the Opto-LED on the BOB, which again drives the Opto-LED in the drivers. (This is coming from converting, from Par-Port, to the ESS)
Not sure if I can skip the BOB, and use the ESS directly to the Driver Opto-LED. But for convenience I kept them......

When you say "Ground-Loop" is that from the PE (Earth symbol) so all units PSU's, PC, VFD etc. to be grounded.
And/or
All the 0V pins on all PSU's: PC-PSU, 5-12-24-48V to have a common 0V
Right now all the PSU's are running without common 0V.

Best Regards, Bo Andersen.

[Robin Hewitt]

It is very easy to forget that as soon as you add a metal enclosure you are not just wiring electrically, you are also wiring inductively.
Do you remember the old trigger operated soldering irons? They had transformer with lots of turns on one side and about 2 or 3 turns on the other shorted out by a thermocouple. The thermocouple received hardly any volts but massive current causing it to reach solder melting temperatures in a second or two and then miraculously stop on account of the fact that it was a thermo couple :)
Anyway the point is, you don't need a lot of turns to make a transformer, one will do. The easiest way to make a single turn transformer coil (earth loop) is to drill a hole in a piece of metal then put a wire through it. Pass an alternating current through that wire and current will start to flow around the rim of hole. Drill another hole nearby with another wire through it and you have a circuit. No nearby hole and it will merely start to warm up.
If you want to put wires through a metal plate I would suggest you knock one ruddy great hole, fill it with a Tufnol sheet then drill holes through that.
If you make sure that everything going in or out goes through the same hole it will all cancel out. Simples.