433.92 MHz data repeater - thoughts ?

manuka

Senior Member
The recent arrival of both Dorji 433 GFSK (Gaussian Freq. Shift Keying) data modules & 08M2 PICAXEs has set me pondering a decent matched application. Perhaps a 433.92 MHz ("434") data repeater may suit their joint horsepower ? The idea of course is that ground level gathered data, perhaps normally obstructed by buildings & terrain, could be received by an elevated repeater & then rebroadcast for the distant receiver. Radio hams & TV stations have been doing this for decades with voice & video, employing a split (typically 600 kHz for VHF hams) between the received & outgoing frequency so retransmission is instantaneous. In data terms however a store & forward approach on the same freq. would do fine, with short delays tolerable.

Years back I rustled up a simple DS18B20 temperature repeater, & this gave excellent results. However no serious attempt was made to ease the 10s of mA level current drain. If battery drain was to be minimised (especially if solar power was used) then some way of sensing incoming signals & awakening the snoozing setup would be essential. In spite of probing around the DRF7020D13 modules RF section, neither an RSSI or WOS (wake on signal) feature was found. Dorji make a simpler FM module which has these, although it may not be compatible. However their more powerful (& hence not legal) DRF7020D20 apparently does at least offer an RSSI tap. Note -the ease of use & GFSK (Gaussian Freq. Shift Keying) features of the Dorji are really appealing,so I'm tempted to continue with these.

Any thoughts or experiences welcomed. AFAIK such 434 MHz data repeaters are legal providing power levels are respected- quite a number of $$$($) commercial ones are available. Stan.
 

Attachments

Last edited:

manuka

Senior Member
Yes- I'd pondered that, & Dorji's DFR1212D10 has diverse "snooze" settings. However none of them seem true WOS & (being FSK) it may not respond to the narrow band GFSK. If a decently long header (of say 5 seconds?) is sent perhaps cyclic wake up could do however.

FWIW Forum regulars may recall a lengthy early 2011 thread of Grogsters on a similar vein. In his case he wanted a rural use "panic" pendant to have a long battery life but decent punch thru' vegetation & terrain. Although he's probably in the thick of spring lambing by now it'd be nice to hear how he got on.
 

hippy

Technical Support
Staff member
I think a store then re-broadcast system should work quite well with a PICAXE ( from as simple as an RFIN, PAUSE, RFOUT loop to filling then emptying the scratchpad ).

The main issue would be data format; well defined and it's quite easy, arbitrary data possibly very difficult other than replicating input as output on another frequency. Application specific repeaters would therefore be much easier than general purpose.
 

Goeytex

Senior Member
Dorji also has front end modules ( no Microprocessor) which allow you put down your own micro and write your own "protocol" and
bypass the limitations of the modules that include the micros. See the DRF4431F13-043.
 

Dippy

Moderator
Do they come properly calibrated out of the box?

Can you do a PIC-sswitched WOS circuit by measuring RSSI on a duty?
With a minor bit of thought you could combine threshold with RSSI WOS duty.

I think you would have to check for RSSI before retransmission unless the users were disciplined with packet sizes.
A combination of hippy's store-then-rebroadcast plus look-before-you-leap broadcast.
(You may have other PICAXErs around blatting data).

Of course, you can start covering more and more bases and end up using Network style software ... which isn't that hard for simple stuff.

Are there not better or more appropriate modules for a task like this?
(Hey Stan, you seem to have gone quiet on Hope and Shonki or Yukki or whatever that other unknown make was . Out of fashion or have they stopped sending freebies ;)
The trouble is if we all start following non-standard unknown RF devices then we have to keep rebuying when another lot comes along. There'll be another Chinese company next week. 2 pennies cheaper.)
 

srnet

Senior Member
Dorji also has front end modules ( no Microprocessor) which allow you put down your own micro and write your own "protocol" and
bypass the limitations of the modules that include the micros. See the DRF4431F13-043.
Where on the Dorji site is the referance manuals for these devices, I see the one you mention has an SPI interface so presumable there is a document explaning what commands you can send etc.

Where do you buy these Dorji modules ?
 

srnet

Senior Member
Thanks, looks like the same RF chip as in the Hope RFM42 module.

Do the Dorji modules suffer from the same register corruption and reset issues (due to supply glitches) as the Hope module ?
 

Goeytex

Senior Member
I have only tested the DRF7020 series and have had no such issues. But that is a different
Chip ADF70xx and a different board.

Register corruption due to supply glitches suggests a marginal supply circuit, inadequate supply,
poor bypassing, or poor board layout . How was it determined that the problem
was actually "supply glitches" (whatever that really means) .

While I have not tested the Hope modules, Silicon Labs and Hope seem to have partnered up to provide
an FCC compliant module. It seems odd to me that Silicon Labs would promote a module that has
an inherent problem such as you describe.
 
Last edited:

premelec

Senior Member
I second Dippy's interest in past referred module comparison and note that Ciseco and Anaren.com modules seem to use TI chips that are quite effective and easier to configure than Ziggys - and where DO you buy Dorji stuff? Thanks...
 

Goeytex

Senior Member
I second Dippy's interest in past referred module comparison and note that Ciseco and Anaren.com modules seem to use TI chips that are quite effective and easier to configure than Ziggys - and where DO you buy Dorji stuff? Thanks...
To buy from Dorji, Contact Mark Yao at www.dorji.com.

It may be helpful to mention Picaxe and that I referred you. This is not an endorsement of Dorji RF Modules or their suitability
for any particular task. I am still testing and writing code for the DRF7020xx. They look promising and IMO should not be lumped in
with the fly-by-nite cheapies.

However I have found Dorji ( Mark Yao in particular) to be the most responsive, honest and above board of any overseas
company I have done business with and (better than some US companies).

William Roth (aka Goeytex)
 
Last edited:

srnet

Senior Member
Register corruption due to supply glitches suggests a marginal supply circuit, inadequate supply,
poor bypassing, or poor board layout . How was it determined that the problem
was actually "supply glitches" (whatever that really means) .
Its been better when its been fed with an LDO regulator for sure, but it still happens. It happens more often the higher the power.

I previously had a non-regulator power circuit, and the 'glitches' due to the variance of Vf from a diode in the supply line definetly caused the problem, no diode or heaps of resevoir capacitor either eliminated or reduced the problem respectivly.

The POR circuit seems oversensitive, the POR flag gets set and the registers get reset.

Mankau mentioned the same problem I believe in another thread ......
 

Goeytex

Senior Member
Not good. Sounds like it could be a layout or ground plane issue. I wonder if it has separate analog / digital
ground plane areas. I doubt it has anything to do with something inherent the Silicon Labs RF IC.

Soon, I'll be getting some Dorji modules that use this IC and will report back with my findings.

Supply "glitches" are many times unavoidable. When I scope the 7020D20 I see the Supply
voltage glitch/ drop when a packet is transmitted (almost 500mv dip) ...but never a reset or
loss/corruption of register data. The same is true with The Nordic based modules but the VDD
dip (measured at the IC) during TX is less pronounced as the TX power is significantly lower .
 

srnet

Senior Member
Sounds like it could be a layout or ground plane issue. I wonder if it has separate analog / digital
ground plane areas
Yeah, I realise that, but I have taken a fair bit of care with that.

The RFM module has 3 seperate grounds (tied together on its own PCB), I have tried single point grounding each of them in turn.

I will probably try a stripboard layout with a specific single point ground. See if it improves.

Its a shame I have the problem its a nice device otherwise. The WUT timer in particular is a useful to shut the 20X2 down into sleep, accurrate sleep times into months, with a sleep current of under 20uA, including the 3v LDO regulator ......
 

manuka

Senior Member
Fair comments gents regarding other offerings! I still use the likes of HopeRF gear to good effect & have never had the unreliablity mentioned by others. Hope's HM-TR however were only a few mW & their receiver not as sensitive as the recent GFSK offerings from Appcon,Sure,Sunray,HAC, Dorji etc etc. Yes- all manner of wireless data devices are stacked up here in my workroom, both awaiting attention & under evalaution. One visitor commented that I'd gear covering "DC to light", before noting some gamma ray detection gear I'm also pondering !

Keep in mind this game of ours is a VERY rapidly moving field indeed. The life of many electronic products is often just a few years before superior approaches arise. Huh -"as long as a few years"- try months in some cases. Having an educational background (where every $ counts & resources are often laughingly meagre) I tend to the cost effective & "can do". It's a Kiwi thing too.

As frequently mentioned before ( both by myself & others), NZ & Australia are now more closely linked to China than many Westerners may realise. It's something UK visitors, perhaps expecting some kind of South Seas Blighty increasingly often comment on... I've been importing gear from UK/EU/US US since -gasp- 1960, but these days most now comes from the likes of Shenzhen. However the UK sourced Ciseco "XBee killer" is next on my to do list,even though it's pinouts are a tad daunting for breadboard work. Stan.
 
Last edited:

manuka

Senior Member
Et voila -a simple proof of concept Dorji 434 + 08M(2) based simplex 'store & forward' data repeater ! I've had decades of ham radio 144 MHz VOICE repeater experience, so was familiar with probable benefits, & this quest was just intended as a check on DORJI based DATA repeater potential.

Raising the repeater above obstacles worked absolute wonders for improving coverage in difficult environments (trees/buildings/terrain etc), with a system performance boost that was most impressive ! This very basic approach can be massively enhanced & tweaked of course. At the very least the repeater (which will probably be hauled up a tree etc, & powered by battery or solar PV) can have some current reducing snoozing 'down time' as it's presently a rather high 20-30mA. Insights & variations welcomed. Stan.
 

Attachments

Dippy

Moderator
As a user-option, can you easily get the repeater to work on a duty or wake-up to reduce the current?
I mean; is it easy?
Obv the user will have to know this to increase the transmission to allow for snooze/duty.
 

john2051

New Member
Hi, An idea that is probably too complicated for what you want. I had a similar amateur radio problem in that I wanted to access the local packet radio
network on 433MHz. I did this in the end using two packet radio controllers (tncs), and two low power 433MHz modules up to the main transceiver.
This works, and has error correction & resend until on the same frquency.

Good luck, regards john
 

Goeytex

Senior Member
Seems to me that the primary TX could send a data byte to tell the Repeater to go to sleep for a fixed time period and then wake up and wait for data for another fixed period. The Primary TX then starts a timer to match the Repeater sleep period ..... Then sends data when Repeater is awake.
 

Dippy

Moderator
Good point.
Does that mean that the Transceiver at t'other end will have to be informed too?
i.e. to set it's timer so that any transmission that it amy do later on isn't wasted.
I'd be very worried about 'sleep' RC time errors.

Really, for general usage, I was just asking Stan if the Doji modules could do 'standard' duty/WOR style behaviour, and if a packet is received it could inform sleeping PICAXE.
I didn't intend starting a long (albeit interesting) discussion on power-saving techniques.

Certain modules e.g. CC1101 can sleep/WOR either with RC for minimum current , or, for a compromise of power/timing-accuracy, they can 'sleep' all the chip-modules but leave the crystal on. If you are using a pre-made module you may not have access to that feature even if available on the RF chip as it depends on firmware authors.
 

Goeytex

Senior Member
There are two variations of the Dorji modules, one uses an Analog Devices RF IC ( DRF Series) and the other uses the Texas Instruments IC same as the Hope Modules. The TI based modules may be more flexible. They also have a TI based front end only module (no micro) that opens up a world of possibilities for someone able and willing to throw down a micro and do the firmware code (similar to what I did with the INHAOS modules.)
 

Dippy

Moderator
"They also have a TI based front end only module (no micro)..."
Aha, that's interesting. Which TI module? I have done extensive PIC work with CC1101 and found it to be excellent.


Slight aside (or word of warning).
I should add to novice PICAXErs; writing sophisticated firmware to make an all-singing Transceiver device with RFchips like CC1101 is WELL beyond novice/keen. Though something basic may be possible for more experienced AXErs.
Not just code but RF quality PCB too.
Please don't think I'm being smug, I'm simply saving you time and money.
 

manuka

Senior Member
Seems to me that the primary TX could send a data byte to tell the Repeater to go to sleep for a fixed time period and then wake up and wait for data for another fixed period. The Primary TX then starts a timer to match the Repeater sleep period ..... Then sends data when Repeater is awake.
Good thinking Goeytex (& glad you got the correct RF IC)! This approach catches my fancy as well, & seems preferable to multiple sends of the data in the hope that the snoozing repeater may just happen to awaken for a reception 'window'. As much 433 MHz data isn't too time critical (crucial info would probably go via cellular now anyway) perhaps the likes of a minutes delay could be quite tolerable ? These Dorjis are transceivers after all, so could TX back to confirm 'down time' was beginning. Perhaps every miniute the repeater could awaken for ~5 seconds (in approx. synch with the TX alarm beeps) and listen to see for the 'radio check'. If this developed instead into actual data then the repeater could be fully awakened & organise for - well what else- repeating to the distant receiver.

Back of an envelope calculations say this may extend repeater battery life by maybe an order of magnitude. Average drain maybe then a few mA,so the likes of 3 xAA batteries should last months.
 
Last edited:

manuka

Senior Member
Here's a simple technique that just involves the repeater awakening (via a PICAXE sent “high” at the Dorji’s ‘EN’ pin3) every minute for a few seconds, and listening for the data being continually fired from the transmitter. Upon detection such data is organised for normal repeating to the distant RX, then the EN pin3 is made “low” for snoozing again. If nothing is heard the repeater’s “lost” LED is activated before the program loops. Only small wiring changes are needed in the original circuit – Dorji pins 3(EN) and 5(TXD) now run to PICAXE pins 3(data in) and 4(sleep/wake control). Simple trial code is shown below.
Code:
'Dorji 433 MHz simplex data repeater Oct 2011
'enhanced for low drain sleep (<1mA)
rptdorji:
high 4    'high to Dorji EN (pin3)to awaken
wait 1    'wake up delay
for b1=1 to 2
serin [2000,lost],3,t1200,("ABC"),b0
serout 2,t1200,("ABC",b0)
next b1
low 4    'Low to Dorji EN (pin3) for module hibernation
sleep 28 'sends PICAXE to low drain sleep (units ~2.3sec) 
goto rptdorji

lost:
pulsout 1,100 'uplink lost- LED alert
goto rptdorji
This crude &#8220;machine gun&#8221; approach requires multiple sends of the data in the hope that the snoozing repeater may just happen to awaken for a reception 'window', but was shown to be reliable and effective with current drains of <1mA noted during PICAXE & Dorji sleep. Depending on the &#8216;sleep&#8217; duration, drains now average just a few mA overall, & repeater battery life may extend by perhaps an order of magnitude-maybe to months. A smaller solar PV could now handle this.

Goeytex mentioned a more elegant scheme - whereby the TX &#8220;pings&#8221; the repeater regularly, telling it to sleep unless fresh data was available. Timing drifts may arise of course (as accurate time keeping is not available on a simple system), but a degree of synchronisation could develop if the repeater acknowledges TX commands before hibernating. The modules are transceivers after all, so a reply transmission can confirm 'down time' was beginning.

My Dorji musings are hosted at =>www.picaxe.orcon.net.nz/dorji434.htm Comments etc welcomed! Stan.
 

Attachments

Last edited:

Dippy

Moderator
The (general purpose) method I used with CC1101 is to set the CC1101 WOR-related registers.

For example, tell it to sleep for 250mS and then wake-up and have a listen. If the RSSI level indicates a burst you stay on until packet is received.
If no RF is detected it goes back to sleep, this would take about 300microseconds. Even with a 250mS period you can see that the duty cycle is very low and thus so is the average current. (CC1101 = 15mA in Rx and 300nanoAmps in Sleep; 800:1 duty , do the maths).

At this stage your PIC / PICAXE is still asleep, so a handful of microAmps.
The RF chip then wakes the Micro out of sleep and transfers packet in a mS.
(I can't remember precise figures I used, but you should get the general idea).

Average consumption would vary on RF traffic, but a once (~10mS) per 10 seconds burst would probably result in an average consumption of maybe 150 microamps. A nice lithium would keep that going for a few years ; no solar panel, no poo/thief-panics, hideability.

This is the advantage being able to 'get' to a RF chip's guts ; lower power, quicker, slicker, more efficient.
I daresay a similar performance could be had by controlling the SI4431.
 

manuka

Senior Member
Thanks Dippy. My ponderings are more at a PICAXEable "ease of use" level of course. FWIW further items have just been released by Dorji, including a cheaper DRF4432D20I module with simpler antenna & activated Group ID features - you can see it hung on the side of their test bed below. (Refer details of the latter => http://dorji.com/docs/app/DAD03A.pdf ). No - I don't have one!

Extra- http://www.picaxe.orcon.net.nz/dorjigas.jpg shows a gas meter hosting one of their SiLabs based DRF1212D10 modules. I've a couple of these simpler transceivers,& hadn't given them much consideration as they've both lower TX power (10mW) & a less sensitive RX (~ -110dBm). However they DO have about a dozen adjustable wake up periods! Refer page 8 at => http://www.dorji.com/docs/data/DRF1212D10.pdf

Stan.
 

Attachments

Last edited:

ciseco

Senior Member
Stan,

You mentioned earlier in the thread, complexity of pinout on the XRF. Could it be the small 2mm headers thats the show stopper (the very reason why we have other versions coming out with different connectors) the xbee layout has become some sort of defaco, but I know not every likes the connector layout.

For most uses on the XRF just 4 pins are needed 1,2,3, & 10 (pwr,rx,tx,gnd).

Dippy have you not tried the newer TI SPI chips, they'll do 7Km, should from what I read be a drop in for the earlier version you have.

Miles
 

Goeytex

Senior Member
The DRF1212D10 is actually based upon the Semtech SX1212 RF IC.

To get the 7km range you likely have to exceed legal TX power limits in most areas.
 

manuka

Senior Member
The DRF1212D10 is actually based upon the Semtech SX1212 RF IC.
I've just inspected a DRF1212 pair here & confirm this!

* Although I've diverse adapters for 2mm, classic 2.54mm (1/10th inch) remains my (& most folks) prefered pin spacing...

* 7km LOS is quite feasible with sensitive GFSK modules using 10 or 25 mW level TX powers, especially if a (legal) RX antenna is used !
 

ciseco

Senior Member
@Goeytex

It's a TI product, the CC1120, not a semtech part

They say - "Featuring range significantly beyond 10-kilometers (139-dB link budget) and 65-dB adjacent channel rejection, typically 30 dB more than nearest competition, the sub-1 GHz CC1120 performance line family provides an unmatched solution for industrial, scientific and medical (ISM) frequency bands at 169, 433, 868, 915 and 950 MHz"

There's a video kicking around the TI site of them running at 7km on simple antennas, here's the main link, should be somewhere of there. From what I have read/watched this distance can be done in ETSI regions. This is believable, haku got 3.4km out of ours at +10db, these are capable of +16db and the receiver is a hell of a lot more sensitive. Shame they don't yet do a SOC version otherwise we'd be all over it for a 10Km XRF :)

http://www.ti.com/ww/en/analog/sub_1_GHz_performance_line_rf_ic/index.shtml
 

rgooge

Member
Stan,

Just fell upon this thread, I assume this is a hypothetical application, but for pure simplicity between two fixed points a passive back to back high gain yagi antenna system might work? One pointing at the source and the other at the destination linked with some good coax.

Of coarse (spelling intentional ;-) ) your thread starter may have been the use of these modules for such an application!

- Rob.
 

manuka

Senior Member
rgooge: Yes- point to point techniques abound, but the quest related to work arounds when such obstacles as buildings, heavy vegetation & terrain totally obstruct the LOS transmission path. Roving transmitters are better tracked by an elevated repeater as well.
 

manuka

Senior Member
There we go -a verified 1200 bps cross harbour (~8 km ) LOS link via a simple 433 MHz ~20 mW simplex digipeater. The uplink transmitter was deep indoors my sea level home,some 500m from the hillside store & forward repeater. Without the elevated repeater,point to point TX-RX ground links thru' vegetation, buildings maxed at ~400m, or ~1km LOS (line of sight).

All 3 parts -transmitter, repeater & receiver- used PICAXE-08M2 controlled Dorji DRF7020D modules connected to similar SMA fitted omnidirectional whip antenna. Coding was straight forward, & included minor remote control to command the repeater & enhance it's 3 xAA battery life. Some weeks-months use could result. I'll look to write this up as a finished project when tweaks & circuit boards are attended to.

Elevation can be everything at UHF of course, but this was a pleasing verification. Stan.
 

Attachments

Last edited:

Goeytex

Senior Member
Nice.

I added an 80mm (3") round aluminum disc as a counterpoise under my antenna and increased the range significantly.
The ground plane of the module is of limited area so this helps.
 

Dippy

Moderator
Nice one Stan.

Ciseco, thanks for info. No, I hadn't seen that. I'll have a looksee.


One general point - and I stress it doesn't refer to products or people on this thread.
There is sometimes a 'competition' for range ; mine's bigger than yours.
I fully realise that achievable range has a number of variables that apply , but my concern is where people select RF devices simply on huge (claimed) range rather than appropriateness.
Many of the smarter modules have adjustable output power, but many don't.
And I wouldn't be surprised if many of the unbranded modules don't even comply with regulations.

I'm not telling you to make a compliant device. That's a whole DIFFERENT subject for another day.
My concern is where/when people select a monster-range module simply because it has a big number or is cheap.

I know, from numpty experience, that a fat 433/434MHz Tx can upset some UK digital terrestrial channels.
All I would ask is that, where appropriate, consider your neighbours BEFORE your wallet.

So, a gentle plea when using 'popular' frequencies; please select the RF device appropriate to your conditions (including legal). And, if using a smart RF, keep the power down to an appropriate range. And please minimise the carrier duty.
Also, consider repeaters as mentioned by Stan. You are less likely to interfere with others when using a repeater rather than some oomphy transmitter/transceiver which can do the distance in one go.
Radio channels are not perfect and can spill over.
All I'm asking is to think about others as well as your project... even if it costs you an extra quid.
 

srnet

Senior Member
You can make a parabolic reflector for poking onto the stubby antennas, at least that way you are directing the power where you want it.

Made out of card and alluminium foil\tape, laminated for durability.

I use them for my WiFi setup at howm, they definetly help, cant see why they would not work at 432Mhz.

Templates can be found on the Internet, simple to make.
 

John West

Senior Member
Nice.

I added an 80mm (3") round aluminum disc as a counterpoise under my antenna and increased the range significantly.
The ground plane of the module is of limited area so this helps.
If you have the room for it, try a 1/4 wavelength-long stiff wire in place of the counterpoise, pointed in the direction of the other station. It's a simple but functional somewhat more directional antenna.
 
Top