THVD8000: THVD8000/CAN

Part Number: THVD8000
Other Parts Discussed in Thread: TCAN330, THVD8010, THS6222

Tool/software:

Hi team,

I have a question about THVD8000, could it work with CAN communication? 

BR

Jingguo

  • It can transmit and receive any half-duplex signal that is not faster than 1 Mbps. So it could indeed replace a plain CAN transceiver like the TCAN330.

  • Hi Jingguo,

    No - the THVD8000 cannot be used for CAN communication - it doesn't support the CAN PHY or higher level CAN standard; CAN transceivers - while differential signaling and termination is similar to RS-485/THVD8000 there are large mismatches in how the devices work at a system and part level that could cause system failure. It can be used in RS-485 like systems and maybe able to be used in place of CAN - but it cannot work with CAN devices. i.e. you may be able to replace CAN completely in the system with THVD8000 but the THVD8000 cannot support CAN standard. 

    Best,

    Parker Dodson

  • Hi Dodson, So I think the device is a good choice to replace RS485 communication, and how long the longest cable does it support? and how many nodes does it support?

  • Hi Jingguo,

    So in terms of how long the cable is and how many nodes it can support is variable due to how you need to design with this part. 

    For cable length:

    It depends on system setup and cabling. 

    In a proper system (cabling is 120 ohm twisted pair) and the start/end node are terminated with 120 ohms and the stub lengths for unterminated nodes are <=300mm (the stub length depends on device but THVD8000 and THVD8010 both are going to suggest unterminated stubs are less than or equal to 300mm for best performance) 

    It should approximately follow the following curve:

    It should be noted that the smallest modulation frequency is 125KHz and the fastest is 5MHz - so only that portion of the graph is relevant.

    Basically in proper setup at 125KHz you would be at 1100ft /335m (conservatively) and at 5MHz you could get 40ft/12m (conservatively) and up to 250ft/76m (allowing additional 5% jitter to output signal). 

    However improper cabling, termination scheme, or unterminated node stub lengths can cause worse performance than listed above. Daisy chaining the network is the best idea to eliminate the stub length issue as that is usually the hardest to hit since you don't have a lot of budget due to the fast rise/fall times on the THVD8000 and THVD8010. 

    For the number of nodes you can support:

    This device is 1/8th unit load - which typically means 256 nodes are allowed. However due to the coupling network  needed for PLC with the THVD8000/THVD8010 the inductors of the coupling network will decrease the total amount of nodes allowed in the system (or require a unrealistic inductance value) 

    8461.THVD80x0_Design_Calculator.xlsx

    I have added a design calculator that allows customer to quickly size the coupling network components based on their system. Please ignore unterminated system values - unless your system is very short (<= 300mm bus length)  we highly suggest to terminate bus. 

    As a note the equations used here are not the same as what is shown in the design guide and datasheet because the design guide and datasheet make the assumption that the node count on the bus will never be high enough to influence the coupling network - that is not strictly true so the equations in the excel sheet include a term for the total amount of nodes on the bus - what you will find is that on  low node count systems the equations in the excel sheet and the equations in datasheet will give you very similar answers but they will diverge for heavier loaded systems - so use the excel sheet calculator for best sizing advice across node counts. 

    Best,

    Parker Dodson

  • Hi Dodson,

    Thanks for your detail reply.

    Customer want use it with 220V AC bus, can the device support it?

    BR 

    Jingguo 

  • Hi Jingguo,

    So the device can be used in applications with higher voltages (like 220VAC) - but you can't do the standard implementation as described in the datasheet as there will need to be additional protection circuitry as well as an additional line driver (we have always used THS6222 for these types of applications)

    We do have a detailed write-up on how to work with higher voltages, you can read more about that here: https://www.ti.com/lit/pdf/slla590 

    We also do have a reference design based off of the application note above that you can find here: https://www.ti.com/tool/PLC010935BP (please note that this reference design refers to DC voltages - but the first note that I attached can be applied to both AC and DC systems because essentially you isolate the communication node with a transformer and high voltage capacitor so the actual power signal on the system doesn't really matter as long as the protection circuit protects the THVD8000). 

    Please let me know if you have any follow-up questions on the high voltage implementations that I linked to above - I understand this a non-trivial solution so I am more than happy to answer any further questions you have here!

    Best,

    Parker Dodson

  • Hi Dodson, Customer is interested in this device because they want to reduce two cables for cost down. but if need additional protection circuitry, what's more, the cost of implementation capacitor will also increase. The cost advantages will be unconspicuous. So in this application, Can you recommend the implementation capacitor spec to me? So that I can evaluate the system cost with customer.

  • Hi Jingguo,

    So with a 220VAC signal the HV capacitor should be rated to 2x the voltage for best results - so 440V rated or higher - technically you can use a capacitor at the same rating as your signal but any slight overstress would break the capacitor and you could also have to deal with a higher amount of derating - which for DC doesn't really matter in this case - but AC it does as the transformer and HV capacitor form a bandpass to help filter out the AC power signal (Capacitor is high-Z to AC signal and the secondary coil on the transformer is seen as low impedance to a standard 50 to 60 Hz power signal). Our test system did use a 1.5kV capacitor - but you don't need that for 220VAC signals. 

    In terms of cost down on the system - this solution may bring a cost benefit to your customer depending on how much cabling they are using currently and how many nodes they have in the system. In general for HV implementation the cost will go up and the minimum BOM per node is shown below:

    1 THVD8000

    1 THS6222 (will require a different supply than THVD8000) 

    1 HV Capacitor 

    1 Transformer

    1 Set of Protection diodes to clamp output of THS6222 path (To Line Driver VCC and GND) 

    1. Set of Protection diodes for THVD8000 input receive path (To THVD8000 VCC and GND)

    1 RC Snubber circuit by transformer on communication side. 

    and realistically you probably also should include the filtering elements shown in the design guide to have the best chance of a successful communication node. 

    Basically the question for your customer is this:

    Does cutting your cabling cost in half cost less than required circuitry for HV PLC communication using the THVD8000 + Line Driver (THS6222) - in general the longer the bus the more chance you will have for having a system that is more cost effective. 

    However there are systems where it is cheaper to keep separate wires for Power and Data it really depends on what they are already using for cabling, how much they are spending on it, and how many nodes would they need to support the high voltage implementation. Generally speaking though - cabling is still generally the most expensive part and realistically most of the components per node are not that expensive compared to proper cabling - but its not uncommon for end users to use cheaper cable as well - so it really will depend on customer specific system - I believe the reference design includes a BOM and that would be a good place for customer to start looking to compare that projected cost to what they would be saving on cabling (~1/2 the cost if they only need one set; assuming the cabling is the same type currently). The reference design keeps the communication true differential so it has two high voltage capacitors - but the original note and testing converted the differential signal to pseudo-differential which reduces the total amount of HV capacitors that are needed.

    Please let me know if you have any other questions!

    Best,

    Parker Dodson