Announcement

**beerygaz** · January 7, 2016, 10:27 AM

Jan, try the new version, I think there's an issue reading he holding registers, the new version will add more detail if that's the case and we can see which one is a problem.

**janjanssen** · January 8, 2016, 05:31 AM

Thx beerygaz, will test next week, camping with the family for a week. Cheers, Jan

**janjanssen** · January 14, 2016, 08:09 AM

Solved - used different addressing method

Originally posted by beerygaz View Post

Jan, try the new version, I think there's an issue reading he holding registers, the new version will add more detail if that's the case and we can see which one is a problem.

Hi beerygaz,

Success with the latest version ... but the problem reading Holding or Input registers was probably all along due to a misunderstanding in how the addressing works in the PI (which is different from how the yoyo HS2 modbus PI worked): to read Holding Register at address "40001", I previously entered "40001" as the address, but I realise now you already assume that Holding Registers start at "40000" and therefore to read Holding Register "40001", one needs to enter "1" as address. Once I used this addressing method, the PI works like a charm!

Other feedback:
- I'd like to poll faster than 1 second, so could your provide and option that polls as fast as possible, i.e. after all the devices have been polled, it starts a new round immediately; this is an important request from my perspective so I can read out potmeter values more or less in real time
- the flag "Zero Based Addressing" seems to work the other way around, i.e. I have to enable the flag to use "1" to read from address "1"?

Anyway, starts to look good.

Cheers,
Jan

**beerygaz** · January 14, 2016, 08:17 AM

Good feedback Jan. Thanks. I'll follow up on the zero-based addressing.

Polling faster may be possible but it will put a lot of load on HS as I update the HS devices as soon as I read a value. Let me see what I can do and what it does to CPU and HS3 responsiveness.

Sent from my iPhone using Tapatalk

**beerygaz** · January 14, 2016, 09:09 AM

Jan, anything below 1s and I start running into timeout issues on my modbus device connected at 19200 baud. The issue may be that I'm reading each register individually. There is a modbus command to read multiple registers at once which I'm sure will be a lot quicker, but then the issue is I need to do a lot of coding to sort the registers you want to us, bundle together the contiguous ones to read simultaneously and then red the others individually. See how that might be an issue?

We you getting sub second, or even sub 500ms read times with YoYo's plugin?

**beerygaz** · January 14, 2016, 12:17 PM

Jan, if you grab the latest release and set the poll time to 0 then it will actually set the poll time to 300ms. Let me know how that performs.

Sent from my iPhone using Tapatalk

**John245** · January 14, 2016, 03:40 PM

Originally posted by beerygaz View Post

Good feedback Jan. Thanks. I'll follow up on the zero-based addressing.

Polling faster may be possible but it will put a lot of load on HS as I update the HS devices as soon as I read a value. Let me see what I can do and what it does to CPU and HS3 responsiveness.

Sent from my iPhone using Tapatalk

Beerygaz,

FYI. In my case the zero-based addressing is working OK.

The device is working according to the manual and your PI shows the same behavior as the Modbus Poll program with regards to addressing.

---
John

**John245** · January 14, 2016, 04:27 PM

Jan,

I believe you are using an Ocean controls KTA-225. Do you have a linkt to the Modbus information of this device?

---

John

**janjanssen** · January 15, 2016, 09:57 AM

Ocean Controls KTA-226

Hi John,

I am using KTA-226, here is the link to the manual:
https://oceancontrols.com.au/datashe...226_Manual.pdf

Regards,
Jan

**janjanssen** · January 15, 2016, 10:38 AM

Originally posted by beerygaz View Post

Jan, if you grab the latest release and set the poll time to 0 then it will actually set the poll time to 300ms. Let me know how that performs.

Sent from my iPhone using Tapatalk

Hi beerygaz,

Thx, already much better, but you can still go faster. I am using 115200 as baudrate and when I look at the signal on my oscilloscope, I can see the polling happening each 300 ms (see attachment). To poll 8 input registers and 8 coils only takes about 30 ms, so you could go down to 50 ms or 100 ms poll time. It does seem to have a bigger impact on the CPU then the yoyo PI (which could go down to 50 ms I think): at 300 ms the CPU load is ~12 % for HS3 and ~8 % for HSPI_MODBUS, from memory even the fast polling with the yoyo PI did not use more than 2 or 3 % CPU time.

Anyway great progress again!

Cheers,
Jan

Attached Files

**beerygaz** · February 5, 2016, 11:45 PM

Jan. Have you tried the latest version that gives you full control over the interval down to the ms.

Sent from my iPhone using Tapatalk

**John245** · February 5, 2016, 11:54 PM

Hi beerygaz,

How can we test the latest version? I have still the expired trial issue

Sent from my iPhone using Tapatalk

**janjanssen** · February 11, 2016, 07:56 AM

Originally posted by beerygaz View Post

Jan, if you grab the latest release and set the poll time to 0 then it will actually set the poll time to 300ms. Let me know how that performs.

Sent from my iPhone using Tapatalk

Hi beerygaz,

Just came back from a business trip, will try when I can install the latest version from the Homeseer store.

Jan

**beerygaz** · February 12, 2016, 12:01 PM

The store now has the plugin available.

Sent from my iPhone using Tapatalk

Announcement

Can't read register

Can't read register

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment