(RJ11 vs RJ45) or Telephone cables verses Ethernet 100 base T

Re: (RJ11 vs RJ45) or Telephone cables verses Ethernet 100 base T

stephen george — Mon, 24 May 2010 21:58:05 -0400

Hi Thomas

RJ11 and RJ45 cables have a different amount of wires in the cable.
So generally no you cannot mix and match them.

Generally.

RJ45 is used for networking
RJ11 is used for telephones

Makerbots electronics have used the RJ45 for it's own ends outside of the ususal networking area. In their later designs they dropped this idea.

regards

Stephen

Re: (RJ11 vs RJ45) or Telephone cables verses Ethernet 100 base T

Thomas — Mon, 24 May 2010 12:49:49 -0400

Can I use an ordinary telephone with RJ11 port to a network and telephone system which have only RJ45 cable.If possible how

Re: (RJ11 vs RJ45) or Telephone cables verses Ethernet 100 base T

BodgeIt — Fri, 30 Oct 2009 21:05:12 -0400

On my Arduino all ~ I did was to remove the power source selector link completly.

Thus the Audrion is not using power over USB or its own onboard 7805 5v regulator.

The Audrion is then powerd using the 5v and ground connection points using Zacks break out board ~ thus it has its very own connection to the ATX power supply.

This also removes the need to put any dummy loads on the ATX power supply's to maintain stability of the 5v line.

Re: (RJ11 vs RJ45) or Telephone cables verses Ethernet 100 base T

mccoyn — Thu, 22 Oct 2009 13:22:31 -0400

> Motors and heaters take significant amounts of current, so they can change "ground" voltage by a significant enough amount for motor noise to be interpreted as valid logic transitions, causing all sorts of problems.

The power molexes contain two grounds for just this reason. You run the steppers and heaters from the ground wire next to the +12 V line and run the digital logic stuff from the ground wire next to the +5 V line. All we have to do is make sure the Arduino is grounded to the reprap supply and everything should work.

Unfortunately, I don't think the existing boards are designed this way. Either the two grounds are shorted in the daughter boards or only one is used.

Re: Power over ethernet = 25 Watts

Triffid_Hunter — Wed, 21 Oct 2009 20:52:44 -0400

in Gen2 land, looks like every module that takes significant current has its own power molex, so there will never be significant current between them and the arduino. You do have to connect arduino ground to reprap psu ground, but a thin wire should suffice as it should only ever carry the miniscule return currents from signals.

If your computer and the reprap PSU are connected to different circuits in your house, that wire may carry large currents during certain conditions (usually electrical faults but not always), so it's wise to ensure your computer and reprap are both plugged into the same power board, double adapter or wall socket.

Re: Power over ethernet = 25 Watts

stephen george — Wed, 21 Oct 2009 19:44:37 -0400

I love it when you talk equations.

Grin

Thanks for the explanation on star point for ground. Much appreciated.
Yes you are quite right when I opened up the psu to hot wire it to always come on
I noticed all the grounds, 12 volt and 5 volt power came from the same solder points.

In GEN 2 land

1) So We have a PSU for the Reprap powering the daughter boards.
2) And we have another PSU in the computer powering the arduino/Microcontroler via usb.

And then we have small thin ethernet wires connecting the GRN of the
microcontroler to the GRN of the daughter boards.

Therefore

Would it be a good idea to use a nice big fat GRN cable (which would have previously been used on the motherboard connector of the PSU) from the star point
of the reprap psu to the GRN of the arduinio/ Microcontroler which is powered by
the computer PSU?

Thanks agin for taking the time to reply.

Stephen

Re: Power over ethernet = 25 Watts

Triffid_Hunter — Wed, 21 Oct 2009 07:04:44 -0400

yeah, POE is sent at 48v too, so only 500mA per pair.

Quote
Stephen George
Sorry don't quite know what "star point is inside the PSU" as far as I know if you are connected to ground you are connected to ground that ground.

(and other grounding-related comments)

wire has resistance, so when current flows through it, a voltage develops across it (V=IR). Motors and heaters take significant amounts of current, so they can change "ground" voltage by a significant enough amount for motor noise to be interpreted as valid logic transitions, causing all sorts of problems.

So, run 2 amperes through two 0.2 ohm runs of wire, and 4 0.2 ohm connector junctions, and suddenly your "ground" is 2.4 volts above what everything else thinks is ground! that's more than enough to register as logic 1 in many logic chips, and will send others into SCR latchup, and it's not always the chips near the current-hungry device that are affected. That's some pretty crappy wire and some pretty crappy junctions, but consider corrosion, loosening connectors, poor crimping, flex damage and occasional current peaks, then add induced noise into the mix and you get pretty much the same effect.

Now imagine what happens if you chain a whole series of circuits in series with a long power bus - even low current circuits will be deeply affected by others on the bus.

A star point establishes one central point in the wiring loom to call ground. Each circuit has its own wire to that point, eliminating multiple connections in series. Running thick wire to high current devices keeps their local grounds within a small tolerance, and signal processing circuits' grounds simply don't get affected by other things drawing lots of current.

High current circuits may still see their own motor noise via their local ground vs signals incoming from other places, which is where differential signalling helps immensely - ground plane noise is common mode, and so eliminated before digitising because only the difference between two input lines is considered.

there is tons of information about the importance of star point grounding and differential signalling, especially in high fidelity audio circles, and it applies to high current digital projects for all the same reasons.

If ever you happen to open up an ATX PSU (or just look at this image), you will see that all the black wires clump together. This is an ideal star point, and saves us the hassles of making another one outside the PSU.

As for power over phone cable, it all depends on the current. one result of ohm's law is that P=I^2.R, or power (dissipated as heat) is equal to resistance times the *square* of the current. In other words, doubling the current through a resistor makes it *four* times hotter. Phone cable is designed for devices that take at most 20mA. 2A is 100x this current, so the cable would dissipate 10,000x more heat than when your phone is hanging off it! Whether or not that's enough to melt the plastic, short the wires and set the molten plastic on fire really depends on the wire's resistance per meter, the ability for it to move heat from the copper to the ambient air, and the type of plastic used. Ethernet cable is superior on every one of these fronts, and has more copper as well. Dedicated power cables would be even better.

The extruder heater takes about that much from what I've read, and the stepper motors (and hence the stepper controller board) could probably take that much in certain conditions, so while phone cable can be used to power logic such as the reprap motherboard and may be perfect for the end-stops, it's simply not suitable for powering any of the more power hungry peripherals. Even ethernet cables cut it rather fine even with the current shared across most of the wires.

The other problem with phone cables is that they're totally inconsistent with wiring. Some have 1 pair, others have 2 pairs, and just within those half will swap each pairs' wires and the other half won't. Regular phones will work perfectly with any of these variations but anything polarised will fry if it's wrong. The handyboard project got seriously ass-bitten by this, as they use rj11 (phone) jacks for signal and power.

More modern standards such as ethernet and usb rigorously define pairs, polarities, wire types and temperature ratings, even colour codes so if a cable is wrong or ambiguous in any way, the blame falls squarely on the manufacturer.

Power over ethernet = 25 Watts

stephen george — Wed, 21 Oct 2009 00:02:49 -0400

Power over ethernet
[en.wikipedia.org]

Apparantly the most power for IEEE 802.3at PoE is 25 watts
or just enough to run a light bulb

Interesting read

Stephen

Re: (RJ11 vs RJ45) or Telephone cables verses Ethernet 100 base T

stephen george — Tue, 20 Oct 2009 23:11:29 -0400

< answer from previous post
> reply (mine)

< Its important to remember that there are outside EMI that may affect devices that require a strong signal to pass through. Regular telephone cable doesn't have this, instead they rely on twisted pairs to reduce EMI. [en.wikipedia.org]

> neither does ethernet cables - generally
(source [en.wikipedia.org])

< ethernet cables can carry more power than phone cables simply by stint of having more wires in them.

> Okay so would a 2x telephone cables (2x 4 pair) carry the same power as 1 ethernet? (1 x 8 pair)

Phones are only supposed to use 10mA*10v = about 100mW, not nearly enough to run a heater or motor.
> I don't quite see your point. we are talking cable not telephone.

Sending that much power over a cable designed for signal sounds pretty scary.
Sounds like the recipe for a fire. I am talking cables between the micro controler and daughter boards (stepper, dc motor driver etc) not from stepper board to stepper mottor.

I don't think so. After all we were rationed to 3 when there was a shortage.

re: grounding. while a star ground is generally a good idea, it makes wiring things up that bit more complex, plus the star point is inside the PSU, so using the drive molexes gives you a star ground automatically.

> each of the sub boards are already connected to the psu ground. Job done.
Why do we have to have another ground from ardiono/micro controler to the sub boards? don't we just need to connect the microcontroler to the PSU ground?

Sorry don't quite know what "star point is inside the PSU" as far as I know if you are connected to ground you are connected to ground that ground.

Still think that telephone cable is better

5) It's more flexible than ethernet.
6) Can be scavinged from old telephones
7) Will scretch nicely due to the coils

1) They are designed to carry power. I.E. your phone hand set
2) It's really old technology
3) It's very wide spread so everyone can get the connectors
4) It's cheaper than ethernet

I especally like the telephone wire From the phone base to the handset which is designed to stretch (due to it's coils) and copes well with continuous movement and being wrapped round fingers while you are on the phone. - good stress testing.
I don't think ethernet is so durable.

Stephen

P.s. Thanks for all the replies to my newbie question.

Re: (RJ11 vs RJ45) or Telephone cables verses Ethernet 100 base T

Triffid_Hunter — Tue, 20 Oct 2009 02:58:30 -0400

actually, telephone cables are simply parallel wires, and only certain uncommon types of ethernet cables (STP rather than UTP) have a shield.

the twisted pair itself regales most noise to common-mode, while the signal of interest is transmitted differentially, making them fairly easy to differentiate at the receiving end.

ethernet cables can carry more power than phone cables simply by stint of having more wires in them. phones are only supposed to use 10mA*10v = about 100mW, not nearly enough to run a heater or motor. Ethernet cables are rated to carry about 300mA per wire, and will carry an amp if they're hanging in air with good ventilation, and 6 of 8 wires are used to carry power to the extruder.

the stepper boards also need end-stop signals

re: grounding. while a star ground is generally a good idea, it makes wiring things up that bit more complex, plus the star point is inside the PSU, so using the drive molexes gives you a star ground automatically.

Basically, for all the electrical concerns you raise, I'd make the steppers run over rj45 as well, and maybe add a power molex to the extruder controller

Re: (RJ11 vs RJ45) or Telephone cables verses Ethernet 100 base T

Grogyan — Tue, 20 Oct 2009 02:02:59 -0400

Its important to remember that there are outside EMI that may affect devices that require a strong signal to pass through.
Regular telephone cable doesn't have this, instead they rely on twisted pairs to reduce EMI.
[en.wikipedia.org]

RJ45 or ethernet cables have an EMC shield that wraps around the whole length as well as having twisted pairs, but unlike straight twisted pairs it doesn't require an equal and opposite signal to auto correct for data loss, ideal for digital systems.

One thing that all the RepRap boards fail to do is screen out as much of the EMI as possible through shielded cable, and it may only take your cell phone to bleep the tower to cause an error in your build. The opposite is true too, to prevent the RepRap from exerting EMI to nearby devices that may be of critical nature eg a dialysis machine

(RJ11 vs RJ45) or Telephone cables verses Ethernet 100 base T

stephen george — Tue, 20 Oct 2009 00:30:49 -0400

Sorry to ask a newbie question

Why are the new gen 3 boards using 4 pair ethernet (8 cables).
rather than 2 pair telephone cable (4 cables).

I would have thought good old telephone cable would be the way to go.

1) They are designed to carry power. I.E. your phone hand set
2) It's really old technology
3) It's very wide spread so everyone can get the connectors
4) It's cheaper than ethernet
5) Each Gen 2 board uses 4 cables -

Stepper - Ground,Step,dir,min
Opto 3 - Ground,signal, +5
PMW driver - Ground,pmw1,pmw2,pmw3
DC motor drv - Ground,Dira,PWMa,Dirb,PWMb, (okay you got me there)

And while we are at it why do we have a Ground anyway?
The PSU will have a proper ground. Use that.
If anything we should be connecting the Arduino/microcontroler
to that ground. Not all the sub boards.

Or am I missing something?

Grin

Stephen