# 0-10V on RJ45 jack - bad practice?

I'm developing a control unit for ventilation systems. Beside switching those fans with a relay they are often controlled with a 0-10V analog signal. Now I found a product that uses a RJ45-plug to supply a 0-10V analog signal.

As the controller's going to have network features as well, chances are that a user will mix up the fan-dimming-RJ45 plug with an Ethernet plug.

So I wonder:

• might connecting this 0-10V output to a network harm a router/other network devices?
• isn't it generally bad practice in electronics to (ab)use a plug that belongs to a different standard for this purpose?

I'd like to support this interface only if it is safe if accidentally connected to a network.

The 4 lines involved are:

• 10V power supply for internal circuit
• 0-10V dimming signal
• ground
• tachometer signal

Remaining lines are NC.

Pin Assignment

Fan's Circuitry

• google Power Over Ethernet – jsotola Aug 10 '19 at 15:16
• @jsotola yes, I read about PoE. From my understanding there usually is a splitter involved, seperating the power lines (which indeed seem to be at 48V) from data lines. E.g. a Raspberry Pi has no more than 5V onboard and a CPU running at 3.3V. My guess is, that the RasPi doesn't like 10V at its RJ45 jack which would allready a disqualifying circumstance. – Sim Son Aug 10 '19 at 15:23

You could make your system compatible (not damaging) to ordinary networking equipment using the same method as used in power over Ethernet (PoE):

You'd apply the 0-10 V signal to both lines of one signalling pair, and a reference ground to both lines of another pair.

Then the transformer at the receiving end would block the 10 V from reaching the low-voltage circuits of the networking equipment. But your circuit could pick off the 0-10 V signal from the center taps of the primary sides of the isolation transformers and use it as you wish.

You'd probably want to design your equipment not to be damaged if 48 V were applied where you expect 0-10 V, in case somebody connects a PoE source to your circuit.

• I really appreciate your efforts! Please see my latest comment on DaveTweed's answer: I might have misinterpreted how the fan in question is meant to be controlled. I now think that the fan's +10V pin 3 actually provides 10V. Therefore wouldn't it be save and the easiest way to divide those 10V with a digital potentiometer an give this signal to the fan's pin 2 (0-10V). I think due to my misunderstanding I caused some confusion... – Sim Son Aug 10 '19 at 17:13
• @SimSon, in any case, my answer only works if you're free to re-design both ends of the connection to use the pins you want to use. – The Photon Aug 10 '19 at 17:35
• Unfortunately that's not the case. The fan in question is an existing product I'd like to control (among others) with my board. So my actual question should have been: how to control such a device while beeing compatible (non-damaging and tolerant) with ordinary networking equipment including PoE devices. I'll accept your answer and might open another one once I've worked out a solution that I would consider suitable. – Sim Son Aug 10 '19 at 17:45

General rule of thumb is that 0-10V signals stay inside the electrical cabinet, and only 4-20mA signals leave the cabinet.
This ensures signal integrity, since a low impedance current loop is more robust, and you know when there is no recipient anymore due to cable loss or unplugging. Since the minimum of 4 mA is not reached.
Although it seems you are stuck with a 0-10V input. This is probably so users can wire up a potentiometer to easily control the speed.

Power over Ethernet is designed to be compatible. This is due to Ethernet using transformers, and POE only using multiple pairs to transfer power, if you stay within a pair you'll not receive more than only the ethernet signals. But you need 4 wires, which is two pairs. Meaning you could receive PoE and fry it.

If you are afraid of user mistaking the 8P6C of signals vs ethernet then maybe don't use 8P6C at all. Or at least not for anything that isn't ethernet or other a balanced differential signalling.
Instead use RJ12 (6P6C) or DB9, or a plain terminal block.

You can get DB9 so it is still possible to wire them on a standard UTP cable by a terminal block in the connector. For example Phoenix Contact SUBCON-PLUS-M/AX 9 - 2904467. Though expensive, I'm sure there are cheaper brands available locally.

• Thanks for your answer! I think I have misunderstood the purpose of this +10V Pin (3). Instead of a voltage-supply input this actually is a 10V output for e.g. a potentiometer, isn't it? – Sim Son Aug 10 '19 at 16:35
• Unfortunately I seem to be stuck to 8P6C as that's the standard this specific manufacturer is trying to establish for horticulture applications. – Sim Son Aug 10 '19 at 16:37
• @SimSon you can leave the 10V out pin if you are self powered. – Jeroen3 Aug 10 '19 at 18:26
• Right, but by actively providing 10V my board will probably fry ordinary network equipment... – Sim Son Aug 10 '19 at 19:57
• Not if you wire it up so it won't ever see two pins of the transformer of the other side. – Jeroen3 Aug 10 '19 at 20:02

It could be safe, but only if the 0-10V signal has a sufficiently high source impedance (which is probably not the case).

The danger is that depending on how the connector is wired, the 10V could be applied differentially across an Ethernet signal pair. If the current isn't limited by the source impedance, this could easily burn out the Ethernet transformer winding.

A commentor mentioned "Power over Ethernet", but that's a special case, in which the power is only applied common-mode to the signal pairs.

As a side note, Ethernet over twisted pair (XXX-base-T) is already "misappropriating" a connector originally designed for telephony. Fortunately, the two applications are electrically compatible.

• I added a pin-mapping description: actually one line is a 10V power-supply for the fans internal circuitry, so the corresponding source on my controller board will need to have a low impedance. – Sim Son Aug 10 '19 at 15:27
• OP could apply their 10 V as a common mode signal between two pairs, then the isolation transformers in standard networking equipment would block it from damaging low-voltage circuits, the same as when a PoE cable is connected to non-PoE equipment. – The Photon Aug 10 '19 at 15:30
• @ThePhoton I've added a schematic of the fan's circuitry and the RJ45's pin assignment. Could you explain if/how what you mentioned is possible in this specific case? I still wonder about those 10V DC supply... also, are the "isolation transformers in standard networking equipment" actually part of the RF45-jack. so, would you consider a Raspberry Pi a "standard networking equipment" which tolerates 10V common-mode signals? – Sim Son Aug 10 '19 at 16:00
• @SimSon, you'd need to change the pin assignments. – The Photon Aug 10 '19 at 16:41
• @ThePhoton seems like I have misunderstood the purpose of the +10V pin 3. My first approach would have been to apply 0-10V to the fan's pin 2 as I allready have a DAC on my pcb anyway. Instead applying e.g. a digital potentiometer between pins 2&3 would make my controller some kind of passive device and therefore be a save way to go, right? The question about common-mode signals would then be obsolete, right? – Sim Son Aug 10 '19 at 16:41