Sim900 auto-Power down after network registration

exactly the same happens with both your modules?
did you check the antenna?
what's the result for the AT+CSQ command with and without antenna?

Well with the antenna plugged in the AT+CSQ results 19,0
With the antenna unplugged it is 5,0

Also with the antenna unplugged it can stay on after "call ready" with 4.2volt. With the antenna plugged it can stay on after the "call ready" with voltage more than 4.5V

i have the same problem too. check your power supply. you have to use 6.8W power supply at least. because sim900 needs 2A.

I use a 500 mA power supply and works fine. Of course I have some large caps there. 2A are only required during transmit pulses.
Try with a SIM card from another operator. I had a similar issue, some M2M cards didn't worked with some versions of firmware from sim900. Also, if you can put an image of the PCB would be helpful.

This is absolutely a power supply issue, I see this problem all the time with cell radio modules! The module is seeing small spikes on the power supply that dip low enough to cause it to reset - there's a circuit in the module to detect low power supply voltage.

Things you need to be aware of:
- You need very good grounding on your PCB so that there is very voltage loss and ground-bounce. A full ground plane is highly recommended.

- You need adequate decoupling at the module BAT pins - the caps need to be physically next to the pins. I suggest a low ESR (> 10 mOhm) Tantalum cap in parallel with a .01uF MLC.

- You need a power supply that has very good transient response. The module will draw upwards of 2 A whenever it transmits. The power supply will see a huge change in the load impedance and if the transient response is not good enough its output will dip. You can have a power supply with a 20A output rating and poor transient response and still have problems.

- You need a low impedance power supply line, and your regulator should be as close as possible to the module. Like the ground plane, a full power plane is your best bet. If that's not an option, use as wide a power trace as possible. For short runs, I'd suggest at least 80-100mils.


You may need a good storage scope to see the spikes on the power supply - they can be under 100nS and still cause problems.

GSM man, sorry but I do not agree with you, at least partially. Even if it looks that way, I think it is not. All the things about transient response and the layout issues are right, but since the modem registers itself and then after a time it disconnects, it cannot be the automatic powerdown of the modem from a deep spike of VCC (under its threshold). The registration phase of the modem takes several seconds and a lot of packets are transmitted at full power. If it is a case of power supply (either bad transient response or not enough current capability), the first spike will kill the modem and put it to powerdown.
The schematic contains several errors in my opinion. The communication between the SIM card and the modem is very fast and the presence of capacitors C5,C6,C7 with those values from schematic, will severely distort the signal, especially considering they are mounted after resistors. There is no place for a low pass filter over the high speed data lines. Tlogic, I don't know who told you to put those caps or where did you see a similar schematic, but it is wrong. The only things allowed instead of those caps are some ESD clamping components if the SIM contacts are available from outside the case. If not, remove the caps and the resistors. The only required cap is C8=100nF. Also, remove R12, it is not required on SIM900. Also, the power on circuit is not correct, the POWERKEY should not be pulled up to VBAT. Remove R13.

pisoiu said:

... but since the modem registers itself and then after a time it disconnects, it cannot be the automatic powerdown of the modem from a deep spike of VCC (under its threshold).

Click to expand...

The Original Post says the module resets (as indicated by the messages from it) when the SIM is inserted, but not when it's not inserted. The reset occurs immediately after the module registers with the network. Also, note that it doesn't reset with the antenna disconnected.

These are the classic signs of noise on the power pin. The fact that the voltage drops 200mV is also a clear indication that the power supply is inadequate. This is GSM101 - the first thing you need to know when designing with these modules is the power requirements are quite stringent. You need to understand what those requirements are, or you will surley have troubles in the field.

I'll keep my original opinion. I have sold so far few thousands products (3 or 4 separate designs) with sim300/sim900 so I understand it quite well. At the time of design, when something was wrong with the power supply, the modem was shut down instantly at the first spike under the limit voltage. There was no way to register and indicate that registration (+creg:2) with bad power supply. I also had situation when the modem staid on even with 1 volt deep spikes, as long as they are not under 3.4V or so.

To design the circuit of the sim card, i used as a guide the olimex sim300 development board ( https://www.rlocman.ru/i/Image/2010/01/21/3.gif ) (capacitors etc.). I put the pull up resistor to the power key because when i designed the board i used sim300 then sim300 production stopped and i bought sim900. I see that it is wrong because sim900 has internal pull up resistor but i don't think that this is the reason why sim900 powers down.

The power supply i use is this one 0-30 VDC STABILIZED POWER SUPPLY WITH CURRENT CONTROL 0.002-3 A which is very stable and i don't think that this is the reason.

As GSM Man said, the fact that the power supply is capable of keeping 3A continuously, does not mean it has a good transient response. Just remove all the caps from the SIM card (make the connections exactly as in the SIM900 datasheet) and try again. Look in the datasheet and make all the connections (SIM, power supply, power on circuit) exactly how it is described there.

tlogic said:

The power supply i use is this one 0-30 VDC STABILIZED POWER SUPPLY WITH CURRENT CONTROL 0.002-3 A which is very stable and i don't think that this is the reason.

Click to expand...

It's all about the Transient Response; for your app the Vpk Undershoot is the critical factor.

Check out this YouTube video on PS Transient Response
YouTube - Agilent DC Power Supply Transient response what is it? how is it measured? and why is it important?. Pay particular attention to the comments around the 4 minute mark where we address the issue of Cell Radio Power Supplies.

hy

I got a similar problem with my SIM900. But my poer dow is random. So sometimes it starts up without any problems sometimes it doesn't (~1 out of 10 tries doesn't work)
I the SIM900 with 3x Tantalum electrolytic capacitor, SMD 470 uF 6 VDC (ESR 0.065 Ω).
Still I get a voltage drop of 840mV over a time of 1.2mS from 4V (LiPo Battery) to 3.2V that's too much I think but what can I change?
As I mentioned even with this hard drop it does work... but not always

Do I need a larger cap?
Thx
Andy

Usually it should work even with lower capacitors. If you have such huge spikes with accumulator, may be something wrong with the antenna or antenna connection. An antenna with high VSWR or a mismatch in the connection can cause such behavior. Also, make sure you follow indications of GSMMan from post #25. Make sure the trace which connects the antenna pad to the antenna is a controlled impedance trace at 50 ohm. If you use grounded coplanar waveguide type, a trace of approx. 1mm with 0.2mm gap to ground plane on each side, on a FR4/1.6mm, is close to 50 ohm. The ground plane under the trace should be continuous and if possible, guard the trace with GND via fence. If you have the possibility, check the antenna with a VNA.

thx for the answer.
I just checked my wiring and oops there was a little mistake... So now I got of 280mV drop in Burst mode (calling the Phone or send SMS, GPRS etc.)
I think I can live with that and do hope that this is the Burst...
Can it happen that the GSM module is going to use even more Amps or are those the maximum when making a call or a GPRS connection?
What if the Signal strength is low (far away Cellular Antenna)? (can't test it do live in a House right to an Cellular Antenna...

Thx
Andy

GekoCH said:

I got a similar problem with my SIM900. But my poer dow is random. So sometimes it starts up without any problems sometimes it doesn't (~1 out of 10 tries doesn't work)

Click to expand...

You're lucky your are seeing it fail that often. I've been called in by clients that have 1,000's of units in the field before they started seeing the problems.

GekoCH said:

I the SIM900 with 3x Tantalum electrolytic capacitor, SMD 470 uF 6 VDC (ESR 0.065 Ω).
Still I get a voltage drop of 840mV over a time of 1.2mS from 4V (LiPo Battery) to 3.2V that's too much I think but what can I change?
As I mentioned even with this hard drop it does work... but not always
Do I need a larger cap?

Click to expand...

Adding larger caps may cause it to reset less frequently, but it's not the cure. Remember this is a `system problem' and you have to address it as such. Look at my earlier post and make sure that your ground system is good, and that the power leads are adequate. I've seen designs with 6 inch leads on a LiOn battery that cause problems. Also, check out the video that SoCalDude referred to.

GekoCH said:

I just checked my wiring and oops there was a little mistake... So now I got of 280mV drop in Burst mode (calling the Phone or send SMS, GPRS etc.)
I think I can live with that and do hope that this is the Burst...

Click to expand...

As SoCalDude pointed out, the critical thing to look at is the Vpk undershoot voltage. That 280mV drop you are seeing may be accompanied by a much larger spike. You need a fast scope to see that (preferably a DSO) as even a spike under 50nS may cause the module to reset.

GekoCH said:

Can it happen that the GSM module is going to use even more Amps or are those the maximum when making a call or a GPRS connection?

Click to expand...

There are a lot of things that will effect the power draw; both the quiescent power and the transmit power. If you are connected to a cell on the 1800MHz or 1900MHz band your transmit power (and therefore your current draw) will be half of what the 850MHz or 900MHz band will be. You have to make sure you test it adequately on the high-power band. You need to do some field testing as well to look at the worst case conditions.

The

I confirm that it's the power supply problem. I had same problem before. I've solved this problem by using 5.2A Li-ion battery instead of regulated power supply (30A @5V). After replaced the power source, my SIM900 module runs steadily and there's no huge voltage drop anymore.
There's an important thing about this issue, you have to make sure that the power wire and the module MUST HAVE A FIRMLY CONNECTION.

Hello All,

I am developing an application using SIM900B GSM module. After going through various posts I noticed that the I am facing similar problems with my board.

1) The board uses a 3.7 V Li-ion battery 1020mAh capacity. The problem I am facing is the SIM900B GSM module restarts sometimes while registering with the network. While the case is this particular problem occurs randomly.

The board has a complete ground plane on it. Is there any way to fix this problem?

please watch the 3.7V with scope while registering with the network.

rohit.bakshi said:

The board has a complete ground plane on it. Is there any way to fix this problem?

Click to expand...

Gsm Man addressed all of the things you need to look at in his post #25. Even in a battery powered design, a full ground plane is just "part" of the solution. Other things to look at include:

- Impedance of your power supply traces (how wide are they)
- The length of your battery leads - the shorter the better
- How far from the module are your filter and decoupling capacitors - need to be as close as possible
- Are you using a low ESR tantalum capacitor on the BAT pin
- Does your battery have the capacity to drive the module
- The state of your battery - old batteries may exhibit poor transient response
- Have you looked at the module's power supply pins using a "good" scope to measure your overhead

Is the problem finally solved or not...?
I have the same issue.