Project

General

Profile

RFM12b minimum startup voltage?

Added by martin about 7 years ago

I am experimenting with energy harvesting breakout board (LTC3108) and have stumbled into the delicate matters of starting up a Attiny84+RFM12b bundle from gradually rising voltage source and quite limited “juice” that will be available (470uF tantalum capacitor). My plan so far is to set the Attiny84 BOD fuses to slightly lower voltage than the minimum startup of a RFM12b, so that it would be doing some preparatory work before the power hog comes in (the RFM12b); BOD will keep the MCU in uA consumption area until sufficient voltage is available. JCW has implemented a mosfet switch to keep the RFM12b calm until needed in the latest JN micro2, but I lack that and have to find alternatives..if possible.
So the idea is to wake up slightly before the RFM12b, power down all unused peripherials, and just sleep/monitor the VCC via bandgap method. When The VCC rises to about 2.2V (or maybe slightly higher to allow for the voltage drop not to BOD the MCU and RFM12b), I would attempt to instruct the RFM12b to power down its oscillator and resume with the sketch.

So is anyone aware at what voltage the RFM12b starts? The datasheet says 2.2V minimum operating voltage and JCW reports it actually works down to 1.8V, but what I am interested in is when does it really start.


Replies (10)

RE: RFM12b minimum startup voltage? - Added by JohnO about 7 years ago

Hi Martin, jcw spent a lot of hours trying to crack the problem you have taken on. It was the reason he designed the Micro V2 & V3.

Here is one of his blogs: http://jeelabs.org/2012/12/29/rfm12b-startup-power-consumption/

RE: RFM12b minimum startup voltage? - Added by martin about 7 years ago

aha, thanks. Exactly what I needed, although it looks bad :) I was expecting the RFM12b is kept in a reset until certain voltage threshold.
I will just have to see how it behaves, the 0.65mA draw may prevent the LTC3108 VOUT capacitor from reaching some useful voltage levels, I guess I have to experiment.

RFM12b minimum startup voltage? - Added by martynj about 7 years ago

martin, certainly a fascinating path to explore, don’t let me put you off. Unfortunately, it ends in a blind avenue.
The difficulty is that the RFM12B gets into its “power hungry” state early in the ramp up before it is ready to talk on the SPI bus.  This means if the ramp stalls due to the consumption rate exceeding what is available to continue the ramp up,  you never get to where it will accept commands to go to sleep!

The history of finding this out is there if you pick some of these Weblog threads One solution is implemented on the current JeeNode (v3) - have a look at the schematic for that and you will see a high-side PFET controlling the RFM12B Vdd exactly for this reason.  

Like most circuit design challenges, there are usually several ways to implement the same functionality, at different costs (component count, quiescent current etc).  For this case, much of the head scratching is done for you - look at the capabilities of the PGOOD pin on the LTC3108….

 

RE: RFM12b minimum startup voltage? - Added by martin about 7 years ago

I agree that the MOSFET switch is great idea, but I don’t have it on my boards. I couldn’t think of good use of the PGOOD in my situation, for a JeeNode micro v3 that would be just excellent wake up signal, but I am dealing with another situation where I don’t have control over the RFM12b’s power state.Looking at the datasheet I notice RFM12B has a NRES, that if brought LOW will put the RFM12b in a reset state, hopefully also in the uA area; since the Attiny84 will become self-aware at around 1.8V or probably less, I could use a jumper wire between the NRES on the RFM12b and a pin on the ATTiny84 to keep the RFM12b supressed until needed. Have you tested something like this?

RE: RFM12b minimum startup voltage? - Added by martynj about 7 years ago

martin,  sure - worth a try. Note that it is not really an issue of waiting until you need the RFM12B - better to grab control as soon as you can and put it to sleep - if it gets up the startup ramp successfully, it is still in ‘power hungry’ mode since the POR values stuffed into the control registers default to CLK “ON”, expensive until you can disable that. It seems an odd power up default, but the designer was thinking of the case where the attached MPU is “cheap and cheerful” and needs an external clock to get it started, unfortunately the reverse of what you have there.

RE: RFM12b minimum startup voltage? - Added by martin about 7 years ago

Tried hooking nRES to GND via 10K and another wire to a digital pin to bring it HIGH when needed; no notable effect really, so that idea goes down the sink. Apparently the CLK keeps ticking even in reset mode.

RE: RFM12b minimum startup voltage? - Added by martynj about 7 years ago

martin, didn’t Captain Hook have the same problem? ;-)

RE: RFM12b minimum startup voltage? - Added by martin about 7 years ago

martynj indeed he did :) He didn’t have a MOSFET to deal with it, just like me :)

RE: RFM12b minimum startup voltage? - Added by martynj about 7 years ago

martin, seriously - no space to cut an etch and squeeze in a high-side PFET?  At least that method is proven.

RE: RFM12b minimum startup voltage? - Added by martin about 7 years ago

martynj sure, I will make another board revision.
btw the LTC3108 is a monster, I made a short movie last night, take a look: https://vimeo.com/61918285

    (1-10/10)