LM2901AV: Comparator Based Arinc 429 Reciever / Transmitter

Hi Jacob,

Can you provide your circuit and and expected supply voltages, input signal ranges and speeds? Taking a quick look at the spec, it looks a bit messy (tri-level -5 / 0 / +5V signaling).

12mA is getting close to the Lm290x current limit. We generally recommend 8mA or less. Most comparators are not designed for power. The TLV902x or LM393LV family is pretty beefy, but it is only 5V max.

When you say "drive", is this sourcing current or sinking current ? The LM2901 is an open collector and can only sink current to GND.

You *may* be able to do it with the LM311 of LM319 - which as a "floating" output transistor that can be operated either common collector or common emitter - thus pulling "up" or pulling "down". But you have to ensure that neither are on at the same time.

Sorry it's my first post and I tried to make it look tidy and inadvertently buried the link to my circuit in the blob of text. 

Receiver circuit

Transmitter Circuit

I would like to use a 3.3V supply for the receiver and +-5V for the transmitter. The fastest speed would be around 12.5kHz data rate, and I'm shooting for no slower than 1.5uS rise/fall time.

"When you say "drive", is this sourcing current or sinking current ?"

I overlooked this, I believe I actually need to source current, 

I'll also attach a picture of the circuit if that is easier to look at.

TX

RX

Hi Jacob,

Looking at the two circuits you've posted, it seems like sourcing capability is only required for the TX circuitry; signaling for RX only seems to only depend on the sinking capability of the comparator, which seem to be in the microamps range from the simulation you've sent.

As Paul said, you could evaluate the TLV9022 or LM393LV for the RX circuit as that family of comparators is meant for low voltage operation.

In terms of the TX circuitry, it seems like sourcing and sinking capabilities are necessary. This means that the comparator needs to have a push-pull output. With the +-10V rail, you could look into the TLV1814 40V push-pull quad comparator. However as Paul said, most comparator output stages aren't meant to drive high current loads. Looking at the datasheet, it seems like the source and sink capabilities of that device still result in the output being ~1V from rail at 12.5mA, which falls out of the +-0.5 spec. As a quick aside, where does 12.5mA drive specification come from? I'm not very familiar with the ARINC 429 specification.

it seems like sourcing capability is only required for the TX circuitry

I agree, my concern was just with the transmitter, I'm not sure why I included the receiver as well in retrospect.

As a quick aside, where does 12.5mA drive specification come from? I'm not very familiar with the ARINC 429 specification.

Referencing page 7, section 2.2.4.2 of this specification doc (https://toddheffley.com/wordpress/wp-content/uploads/2014/07/Arinc429WordList-copy.pdf )
 "The total receiver input resistance including the effects of RI, RH and RG in parallel should be 8,000 ohms minimum (400 ohms minimum for twenty receiver loads). " and "Differential Input Capacitance CI = 50pF maximum"

From what I understand 400 Ohm for a maximum case of 20 devices, so 5V/400ohms = 12.5mA. Let me know if I've steered wrong somewhere. Again this would be worst case but technically within the 429 specification. 

I will look into your listed comparators to see if it might get me closer to my goal. Thanks for the suggestion. 

Hi Jacob,

It's always fine to include more information as it helps us paint a picture of your system as a whole.

Thanks for sharing the specification; it seems like the minimum output resistance to ground would be 400Ω per the specification.

Could you confirm if this is how you expect the system behave?

RX: For a differential ARINC 429 that swings +-5V, the circuit outputs a differential TTL signal that swings 3.3V to GND.

TX: For a differential TTL signal that swings 3.3V to GND, the circuit outputs a differential ARINC 429 that swings +-5V.

Is this circuit expected to propagate "NULL" inputs? As Paul stated, tri-level would be tricky as comparator outputs are expected to rail.

Your description of the system is correct for RX and TX. And yes I would like to propagate the 'NULL' state on both lines. I was under the impression I could gate the power but I see the power-on time is 200uS which would be too slow. I will need to reevaluate this as I don't think tri-state buffers that can operate at +-5V exist. Let me know what you might suggest.

Thanks for the help. 

Hi Jacob,

There'll be some difficulty in propagating the "NULL" level with comparators. As you've stated, the power-on time is too long and the comparators would take too long before recovering to propagating non-NULL voltage levels. In addition to this, push-pull comparators with power-on-reset (POR) will have outputs that are pulled to the device negative supply whenever the comparator supply voltage is below minimum supply. This means that the output may not necessarily be Hi-Z whenever the power to the comparator is turned on and off.

As for what I'd suggest, I'm not entirely sure if this tri-level signaling can be done with a comparator. I'm not an expert on tri-state buffers, so it may be helpful to make an additional post to get input from the logic team on realizing this circuit.

Thanks,

Ho