Frewave DGR-115 and Modbus ASCII?

Hi all,
Does anyone know if Frewave (in my case DGR-115) supports Modbus ASCII mode?
Thanks.

It's a radio modem. Most radio modems don't care about the data they carry, they just move what's on the serial input to the radio output on one end and reverse that on the other end.

That said, Modbus ASCII uses 7 bit words. If the serial setup involves defining a word size, you'll have to make sure that it supports 7 bit words.

danw said:

It's a radio modem. Most radio modems don't care about the data they carry, they just move what's on the serial input to the radio output on one end and reverse that on the other end.

That said, Modbus ASCII uses 7 bit words. If the serial setup involves defining a word size, you'll have to make sure that it supports 7 bit words.

Click to expand...

Thanks for the response Dan.
I am a master and I read/write data from another PLC where there is a pair of Freewaves between. The slave is connected to the radio modem directly over RS-232, on my side I have a TSXETG100 (Modbus-Ethernet Gateway) which is configured as Master, ASCII, Even, 19200 and pretty much it - I guess # of bits and stop bits are configured in the Radio Modem. I can read/write data and everything looks good on the slave side but on my side sometimes I can see garbage in the registers I read. I read 10 registers and and I know that only the first register is a time counter which increments every second and all the rest should be zeros, but from time to time I see non-zeroed values in some of them, sometimes even in register 1 - value is way off than the counter.
Not sure what could be the issue in this moment. Any ideas?

have you tried contacting freewave support? Their support has been excellent every time I have called them up.

diat150 said:

have you tried contacting freewave support? Their support has been excellent every time I have called them up.

Click to expand...

Hi diat150,
I sent a web query to Freewave about supporting Modbus ASCII and the support person was not sure if Freewave supports it or not. According to Dan the radio modem should support Modbus ASCII as it is a part of Modbus protocol itself (and this makes sense).

If the protocol is Modbus ASCII and you can read the first register as a valid number (can you check it?), and you're not getting errors, and you don't see the start character (colon ASCII 3Ah) and an end-of-message character (CR LF, ASCII 0Dh,0Ah), then the protocol must be functioning to decode the ASCII characters into an integer or whatever the count format is and not display raw control characters.

Noise usually causes an error where the message is not even interpreted. So seeing a false bit or bits due to noise seems unlikely. Modbus ASCII uses an LRC check, not a CRC check in each message. If the LRC detects an error, the operating system reports an error, it does not decode the message.

Something in my memory says that an LRC check is not as good at finding the same percentage of errors as CRC, but even so, noise doesn't sound likely.

How positive are you that the other registers are locked at a value of zero?

Dan,
I am 100% positive on the other registers been zeros.
One more: if I read 2 registers, sometimes I see garbage in the 3rd register which does not make sense to me. Same if I read only 1 register - sometimes I see garbage in register 2.

The topic of error checking in Modbus for radio modem communications was discussed in this thread:
http://www.control.com/thread/1340906590#1340906590

The FreeWave modems default to a 10 bit word with 8 data bits, 1 stop bit and no parity (there’s always a start bit). 7, Even, 1 is also a ten bit word. If the modem is set to its default settings I.E. a standard 10 bit word (8,N,1) and the equipment is using a non-standard 10 bit word (7,E,1) most of the time the data will be OK. However from time to time you will get corrupted data. It all depends on how the data is broken up to be sent across the radios. If a communications packet is broken into two hopping packets then chances are the packet will get corrupted when put back together because of the mismatched protocol. To set the FreeWave ports to 7,E,1 set the “Data Parity” setting (via the “Baud Rate” screen) from “0” (8,N,1) to “1” (7,E,1).
If that doesn’t make any difference the problem isn’t related to the radios (most likely). What can’t happen is RF noise causing corrupted data. If the radios are in a high noise environment, which is difficult to do with FreeWave radios, they will drop and resend packets. The packet itself can’t be corrupted. With the radios running over the air at an RF speed of 115.2Kbps there is a lot of room for the radio to sort out getting packets through in bad RF environments before 19.2Kbaud is affected.

Dan, thank you for the link and sharing your knowledge.

Firejo, thanks a lot for such detailed information. It maybe exactly what you are talking about, I will check on site.

> What can’t happen is RF noise causing corrupted data.

I tend to agree with Firejo, I don't think noise is likely to create false data bits. The protocol's LRC is likely to catch errors and refuse to interpret the message and flag an error.

The combination of false bits that can fool the LRC just seems to be low probability enough that you wouldn't see the errors as regularly as you do.

FreeWave (and a few others) use a very complex error checking routine and if things show up from across the RF link and they aren’t perfect they get tossed. I’ve been working with the FreeWave radios for around 18 years now and I’ve never seen noise cause corrupted data. It’s been explained to me (a couple of times now) but it’s far enough above my head that I lose it after a few days.