# Forwarding Timeout, can "s" be floating point?

**URL:** https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812
**Category:** Digi TransPort WR (Series)
**Tags:** serial, wr44, forwardingtimeout
**Created:** [December 19, 2016, 7:07pm UTC](https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812 "2016-12-19T19:07:22Z")
**Posts on this page:** 8
**Page:** 1

<div class="post-metadata">

### Author: ![jserink](https://avatars.discourse-cdn.com/v4/letter/j/bc79bd/32.png) [@jserink](https://forums.digi.com/u/jserink)
#### Post date: [December 19, 2016, 7:07pm UTC](https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812/1 "2016-12-19T19:07:22Z")

</div>

Hello:

I have data coming in at 115200 baud with 1 bit per baud so effectively 115.2Kbps.  
1 stop bit and 8 data bits is 9 bits, lets call it 10 bits for ease of the Arithmetic, takes 87 microseconds. I would like to set the Forwarding timeout to 100microseconds or set s=0.1.

Is this supported?

Cheers,  
John

---

<div class="post-metadata">

### Author: ![NicholasWilson](https://avatars.discourse-cdn.com/v4/letter/n/c77e96/32.png) [@NicholasWilson](https://forums.digi.com/u/NicholasWilson)
#### Post date: [December 20, 2016, 2:06am UTC](https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812/2 "2016-12-20T02:06:51Z")

</div>

I think you are looking for ATS15=100

The S register definitions for a serial port are on page 132 of the manual.

---

<div class="post-metadata">

### Author: ![jserink](https://avatars.discourse-cdn.com/v4/letter/j/bc79bd/32.png) [@jserink](https://forums.digi.com/u/jserink)
#### Post date: [December 20, 2016, 2:37am UTC](https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812/3 "2016-12-20T02:37:27Z")

</div>

Hello:

Yes, it loos like it would be S15 but, as this is a Digi Transport manual, finding something in one section usually contradicts another. So, on page 132 this is what is says for S15:  
Reg Description Units Default Range  
S15 Data-forwarding-timer ms 2 0-255

On Page 271 we have the following:  
Forwarding Timeout s x 10 milliseconds  
The length of time that the router will wait for more data after receiving at least one octet of data through the serial port and transmitting it onwards. This timer is reset each time more data is received. The router will forward data onwards when either the forwarding timer expires or the input buffer becomes full. This parameter applies to ADAPT, TCPDIAL, TCPPERM and PANS.

Now, Page 132 says the units are in ms but page 271 says the units are in increments of “10 milliseconds”.

The range of values on page 272 and 132 appears to match, 0-255. That is sorted.

What units are values possible values of 0-255?

Cheers,  
John

---

<div class="post-metadata">

### Author: ![NicholasWilson](https://avatars.discourse-cdn.com/v4/letter/n/c77e96/32.png) [@NicholasWilson](https://forums.digi.com/u/NicholasWilson)
#### Post date: [December 20, 2016, 3:51am UTC](https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812/4 "2016-12-20T03:51:30Z")

</div>

It is a value in tens of milliseconds and the S register description is wrong.

If you set it to 255 it will take 2.55 seconds for a character to echo (which is how I confirmed this).

---

<div class="post-metadata">

### Author: ![jserink](https://avatars.discourse-cdn.com/v4/letter/j/bc79bd/32.png) [@jserink](https://forums.digi.com/u/jserink)
#### Post date: [December 20, 2016, 3:53am UTC](https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812/5 "2016-12-20T03:53:11Z")

</div>

What if I set it to zero?

---

<div class="post-metadata">

### Author: ![NicholasWilson](https://avatars.discourse-cdn.com/v4/letter/n/c77e96/32.png) [@NicholasWilson](https://forums.digi.com/u/NicholasWilson)
#### Post date: [December 20, 2016, 3:55am UTC](https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812/6 "2016-12-20T03:55:21Z")

</div>

Then it is “instantaneous”. I would suggest you test thoroughly if you put it in this mode. I suspect the 20ms default is there for a reason.

---

<div class="post-metadata">

### Author: ![jserink](https://avatars.discourse-cdn.com/v4/letter/j/bc79bd/32.png) [@jserink](https://forums.digi.com/u/jserink)
#### Post date: [January 17, 2017, 11:43pm UTC](https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812/7 "2017-01-17T23:43:49Z")

</div>

Hi All:

I just completed some testing with this. I used a GPRS receiver that had both an Ethernet port and Serial port. I connected the serial port of the GPS to the Digi and the the Ethernet port to the office switch. I then connected the office switch to the GPS Ethernet port. I configured the unit to output NMEA data, GPZDA, on both the Ethernet port, TCP 4002, and the Serial 1 port at 1 second epochs. I wrote a perl script to compare the arrival of both streams, one via the Asyn0 and 1 via directly from the GPS while varing the forwarding timeout value. I took one hundred samples and averaged them. Here are the results:

```auto

Forwarding	Average
Timeout Delay(ms)
90 905.84
80 806.64
70 705.98
60 607.13
50 506.98
40 413.43
35 356.79
34 346.78
33 336.66
32 327.79
31 316.64
30 306.46
29 297.20
28 291.12
27 277.03
26 267.33
25 257.60
20 208.22
15 157.43
10 107.73
9 97.64
8 88.47	
7 77.63
6 67.97
5 57.32
4 47.67
3 37.79
2 27.44
1 3.98

```

Hope this will be of some help.

Cheers,  
John

---

<div class="post-metadata">

### Author: ![jserink](https://avatars.discourse-cdn.com/v4/letter/j/bc79bd/32.png) [@jserink](https://forums.digi.com/u/jserink)
#### Post date: [January 17, 2017, 11:44pm UTC](https://forums.digi.com/t/forwarding-timeout-can-s-be-floating-point/17812/8 "2017-01-17T23:44:40Z")

</div>

Appologies, the tabs did not come out in my last message but I think you ca see the numbers…used code tags, seems to have worked.

Cheers,  
john
