Receiver quality in 433 OOK plug

Jee Labs Talk » How can I fix ... ?

Receiver quality in 433 OOK plug

(21 posts) (6 voices)

Tags:

No tags yet.

panosnl

I just ordered the OOK plug and I am a bit disappointed with the reception.
I have hooked it up to 3.3V, so i have not tried 5V yet. I have attached a 17cm antenna.
My KAKU switch can receive easily through 2 concrete floors. The OOK plug doesn't receiver through 1 concrete floor. Even 2 walls (gipsblokken) results are unreliable. The outdoor Oregon sensor is also not noticed...
I enabled the activity LED and I see there is continuous pulse activity.
I have read a post in the Nodo forum about china receivers having noise levels above the TTL level causing the arduino to be constantly busy checking rubbish. http://code.google.com/p/arduino-nodo/issues/detail?id=82
Has there been a comparison of the OOK plug receiver with the aurel receiver from conrad?
I will try 5V next. Hopefully this is the solution.

Posted 11/02/2011 20:23:48
jeroen

Hi, I have both receivers so I did some tests with the low cost conrad receiver (130428 - 8A) and with the aurel ac-rx 5 volt standard ook receiver 100dbm. Low cost conrad is compared to the Aurel receiver indeed not that good. Conrad does not receive with any wall or floor between transmitter and receiver. And the aurel works fine with two concrete floors.

Posted 12/02/2011 10:54:43
johno

That is disappointing, I have an order pending for the OOK plug. I'm having second thoughts now.
jcw - Is this is likely to affect new deliveries?

Posted 13/02/2011 15:43:29
panosnl

I have been playing some more with the OOK plug.
I added a counter of detected pulses and I get about 3500 false pulse detections per sec. Processing these takes about 150ms per sec.
Also I experienced that the KAKU control triggers about 3 to 4 HEZ reports and 0 to 2 KAKU reports (KAKU is sometimes missed). When I disable HEZ in the code all KAKU is reported with 1 to 2 lines.
Furthermore My Oregon THGN132N is recognised as CRES instead of Oregon. When I disable Cresta in the code the sensor is not seen at all.
I have some questions: KAKU being recognised as HEZ: Is this receiver /noise related or something else?
Oregon THGN132N sensor not being recognised. Is this sensor not yet supported? Why is it recognised as Cresta?

Posted 13/02/2011 20:44:39
anpfeiffer

"Also I experienced that the KAKU control triggers about 3 to 4 HEZ reports and 0 to 2 KAKU reports (KAKU is sometimes missed). When I disable HEZ in the code all KAKU is reported with 1 to 2 lines." Indeed, the Intertechno(KAKU) decoding doesn't work well.

Posted 13/02/2011 23:10:55
jcw

The extra pulse triggers are quite common - these radio's have Automatic Gain Control (AGC), so basically they keep increasing their sensitivity until something is detected - meaning noise is bound to take over in quiet times. As you noted, the moment a signal is received, that sensitivity will adapt accordingly. The 15% processing overhead is not something to worry about, IMO.
The false and missed packet detections are caused by two different effects, I think: these decoder modules have not yet been tweaked to always accurately match pulse patterns. Note that many of them were constructed from looking at signal patterns with the OOK Scope, and then guessing what the different pulses mean and how packet framing should be mapped on top. And second, some devices tend to use the same protocol as others, so it is not always possible to uniquely determine the brand (CRESTA vs OS).
I don't have all the wireless remotes people use, so I'm hoping that with a bit of collaboration we can improve the detection process for more and more sensors. I do see the HEZ/KAKU mixup here all the time, so that's probably the first one to look into.

Posted 14/02/2011 18:16:11
panosnl

Receiver range was the most disapointing for me. This makes the OOK plug too unreliable to use. I'll have a try with Aurel receiver later to see if that is indeed an improvement as suggested.
Wrt the oregon sensor. I tried the code from http://jeelabs.net/projects/cafe/wiki/DecodingtheOregonScientificV2protocol .
Initially still only the cresta was reported. When I disabled the cresta code finally "OSV2 1A2D109B0020904539C6" was reported for my sensor thgn132N. It seems that the OOKrelay2 scetch does not yet support Oregon v2 (and V3?). Next step is to get T and Hum out of this.
I don't think the mixup is a protocol sharing since the reported cesta codes have different lengths and values over time which doesn't make sense. The OSV2 output looks more stable over time (only slow variation which can be attributed to T changes).
I there is anything I can contribute to fixing these issues I will be glad to help.

Posted 15/02/2011 18:08:40
johno

Hi,
Can you tell me which sketch you used to test the receiver? I can't see anything received by my OOK433 Having said that, I am also unsure which port it should be connected to.
JohnO.

Posted 19/02/2011 21:49:04
joop
To test the receiver you can use the OOKrelay scetch. Place the receiver in port 4.
these are the ports defined in this scetch :
```
      
    #define PIN868 14 // AIO1  = Analog 0 = pin 23ATmega328(PC0) = port 1 jeenode 
    #define PIN433 17 // AIO4  = Analog 3 = pin 26ATmega328(PC3) = port 4 jeenode 
```
I did't test it I hope this info is right.
Posted 20/02/2011 08:17:29
johno

Hi Joop,
Thanks for taking the time but I'm still struggling. The repository http://jeelabs.net/projects/cafe/repository/entry/sketches/ookRelay.pde has the following lines: 28#define PIN868 14 // the 868 MHz receiver is connected to AIO1 29#define PIN433 4 // the 433 MHz receiver is connected to DIO1
I tried the above as is and also with PIN433 changed from 4 to 17, as per your post, connected to P4, still no joy. The PCB indicates that the data pins connect to the analogue port: http://jeelabs.net/attachments/834/jlpcb-125.png so the repository looks imperfect in line 29: "// the 433 MHz receiver is connected to DIO1" I wonder, does this thing actually run on 3.3V, documentation says 5V.
Perhaps mine is DOA.
JohnO.

Posted 20/02/2011 10:49:29
jcw

Did you add two solder jumpers? The docs at 433Plug> were wrong ("nothing to configure"), I've just corrected it.
I've tested that receiver and transmitter operate properly at 3.3V.

Posted 20/02/2011 11:16:00
johno

Hi jcw,
I decided to test before soldering it all and so I have the receiver board connected using three cores of an extension cable. Using + not PWR, the two data pins on the radio are already linked on the radio PCB. I have a 16.4cm antenna connected.
Is there a really basic sketch that just reads the analogue pin?
John.

Posted 20/02/2011 11:24:54
jcw

> Is there a really basic sketch that just reads the analogue pin?
One way to test is to read the AIO pin and set another pin on another port, constantly in loop(). Then use a LED to see what's going on.

Posted 20/02/2011 11:47:21
johno

That was good jcw, being a novice at code I have managed to get some analogue samples coming out and it feels to me like an on/off scheme happening on the pin. 0 258 258 258 258 0 258 258 0 0 0 0 B 0 0 0 258 258 85 0 21C 168 258 0 0 0 0 0 0 0 258 0 258 258 10 0 0 0 258 232 258 0 258 258 257 258 0 0 258 132 258 257 0 258 0 0 258 258 258 0 0 AA 0 -- Armed with this small success I'll try something closer to what's needed.
JohnO.

Posted 20/02/2011 12:58:07
johno

I'm happy now, my problems were all my own. I didn't have a KAKU device and you see nothing unless there is a protocol match. I have switched sketches to http://jeelabs.net/projects/cafe/repository/entry/RF12/examples/ookRelay2/ookRelay2.pde and after copying decoders.h into my /libraries/RF12 directory and connecting to JeeNode P4, all is now well.
Thanks everyone for supporting me up this steep learning curve.
JohnO.

Posted 20/02/2011 17:05:19
johno

Happiness is oh so transitory, having tested the range I am no longer so happy. About one metre through air is somewhat limiting the usefulness. I am testing with a Marmitek EasyTouch 35 handheld remote control decodes as XRF. The OOK 433 unit has a 16.4cm antenna fitted!
I wonder if tuning the coil thing can make things worse?
JohnO.

Posted 20/02/2011 17:23:32
johno

Range improves to around five metres through air when using RCT-08 remote control which decodes as HEZ. With sensible force I haven't been able to move the 'tuning' screw.
JohnO.

Posted 20/02/2011 17:44:30
johno

Hi,
I am running my OOK433 receiver with a 16.4cm antenna and would like to improve the range. Antenna theory seems a mystical from what I see on google and so I hope we have an antenna alchemist and a ground plane guru in the forum.
JohnO.

Posted 24/02/2011 08:01:58
panosnl

After my initial disappointment on the OOK plug's receiver range I want to give an update:
I replaced my 17.2 cm (lambda/4) soft wire with a 1 mm solid copper wire of 34.6cm (lambda/2). Now the KAKU transmitter is detected through 2 conrete floors...
The transmitter I have tested only using NODO to transmit KAKU command. Range is sufficient. Is there a jeenode sketch that includes KAKU commants using the 433 OOK plug?

Posted 24/02/2011 19:41:06
panosnl

Update 2:
I changed the Voltage from 5V back to 3.3V and I could not receive through 2 concrete floors anymore.
After changing back to 5V good reception is restored.
Question still remains: Is there a Jeenode sketch available for transmitting KAKU using an OOK plug?

Posted 28/02/2011 20:47:38
johno

Hi, I'm still on 3.3V now using a 34.6cm antenna, disappointing performance.

Posted 01/03/2011 16:34:25