Noisy ground on the super carrier board with 2p24

Istha Powron

I have been working on a wirless control for a robot and have it communicating well.

The robot end consisted of a BS2p24 installed on a BS2P24/40 demo board connected to a 433mHz receiver and transmitter module.

Once it was all working, I transferred the BS2p24 to a super carrier board and then built a separate board with the transmitter and receiver. The two are connected via a ribbon cable.

When I turned the system on, I found that the receiver was constatntly busy receiving a ridiculous amount of data. I won't go through the hours of fault finding, swearing and throwing things around the room.

The bottom line is that there appears to be high levels of noise on the ground line of the Super Carrier Board.

Just the board with BS2p24 powered by a battery shows around 200 mV of noise at a frequency of around 10mHz.

This is enough to set off the receiver. I have tried changing baud rates, and have tried installing combinations of caps to quiet things down but it is still problematic.

When I reinstalled everything on the BS2P24/40 demo board, it wokred perfectly again. While there was noise on this board, it was in the order of 10 mV. I guess the large copper ground plane on this board supresses a lot of the noise.

I have searched the forums, but have not come up with a solution. Is this a common problem with the Basic Stamps and development boards? If so, what is the standard fix?

Thanks,

Istha.

If so



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔





The robot is going to lose. Not by much. But when the final score is tallied, flesh and blood will beat the damn monster.
Adam Smith

Mike Green

This is not a common problem with either the Basic Stamps or the development boards including the Super Carrier Board. It would help if you could provide more information like photos of the Super Carrier Board that you're using and drawings of the circuit that's on the Super Carrier Board and the receiver/transmitter board. It may also be related to the ribbon cable connection to the receiver/transmitter board. If there's a poor ground connection in either connector or in the cable itself, that could contribute to the problem. Another question is whether there's adequate bypassing of the power leads on the receiver/transmitter board (always at least a 1uF-10uF electrolytic and a 0.1uF ceramic capacitor).

Istha Powron

Mike,

Thank's for your response.

I have two Super carrier boards.

Even a bare board with just the BS2P24 and NO connection to anything else (no ribbon cable) shows similar levels of noise. In fact now that I have caps between 5V and ground on the tranceiver board, the noise is less than on the bare board.
You have basically answerd my question. I shouldn't be seing this.

For interest's sake, I have attached two photos. The first is the trace of the levels between VSS and VDD on the header of a bare supercarrier board with·a BS2p24 installed. Nothing else. You can see the amount of noise. And NO, it makes no difference if I run the board from a 9V battery.

The second shows the levels between GND and 5V on the transceiver board taken from the smoothing caps (10nF & 100uF). You can still see that there is 100mV + of noise. It makes no difference if the HB25s are powered up or not.

At this stage all I can think of doing is providing seperate power to each board, common grounds and low pass filters on the comms lines. Though if I go too low, it will stop the serial stream.

I don't quite understand the bypassing caps you suggest. Should I place one between the supply & ground pins on·both the transmitter and receiver board?

I think I have the technology virus. This should be easy.

Regards,

Istha

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔





The robot is going to lose. Not by much. But when the final score is tallied, flesh and blood will beat the damn monster.
Adam Smith

Mike Green

The datasheet for the LM2940 regulator states that at least 22uF should be used as an output filter capacitor for stability of the regulator. I don't know why the schematic of the Super Carrier Board only shows a 1uF capacitor. I suggest that you put a 22uF or 33uF 6V-10V Tantalum capacitor in place of or in parallel with the 1uF one and see how your noise is doing. You might put a 0.1uF ceramic capacitor across the Vdd/Vss lines near the regulator and near the header to the "daughter board". I tend to err on the side of more 0.1uF capacitors ... always.

It would be interesting for you to remove the BS2p from the Super Carrier Board and measure the noise on the supply without it. You might also try to measure the noise right at the regulator terminals rather than at the AppMod connector.

Post Edited (Mike Green) : 7/29/2007 4:53:41 AM GMT

Istha Powron

Thanks again Mike.

I'll pick some up in the morning.

Istha

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔





The robot is going to lose. Not by much. But when the final score is tallied, flesh and blood will beat the damn monster.
Adam Smith

Phil Pilgrim (PhiPi)

I agree with Mike about the 22uF cap. According to the LM2940 datasheet, a cap of at least 22uF with an ESR between 0.1 and 1 ohm is required on the output for stability. The 1uF cap shown in the schematic isn't even close. Rather than soldering another cap in parallel, I'd remove C2 and replace it, just to keep lead inductance to a minimum.

If that doesn't solve the problem and if the Vdd and/or ground traces that connect to the AppMod header pass through any other circuitry, I'd cut them and run wires directly back to their respective connections on C2.

And sprinkle 0.1uF ceramic bypass caps liberally on your TX/RX daughterboard at each point where Vdd powers active circuitry. Also it's good practice to install a tantalum cap (10uF or more) at the point where Vdd enters any daughterboard — especially when fed by ribbon cable.

-Phil

OzStamp

Hi

The HB25 in the picture with the DC motors .. are they running when the noise is bad ?

Your workbench .. what is the surface ... it is wood with a v
arnish... static from body...
What is the floor your physically standing on ... nylon carpet ???

All these things could contribute to the noise your seeing.

Rgd Ronald Nollet Australia

Istha Powron

Ronald,

Makes no difference if the HB25s are on or off or if the six dc motors are running or not.

Static? Makes no difference if I am at home or in the static sensible electronics workshop at work.

I am now convinced the noise is comming from the oscillator causing resonance somewhere.

I have so many capacitors on the boards that the lights dim when I power it up and it keeps going for ages after I turn it off.

I really should just give up and buy some decent tranceivers.

Thanx,

Istha

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔





The robot is going to lose. Not by much. But when the final score is tallied, flesh and blood will beat the damn monster.
Adam Smith

Tom C

Istha Powron,

I am using the same board with the new 912MHz transreceiver, but have yet to try the board out with the transreceiver in operation.

I presently have a BS2e installed on the board and the transreceiver is mounted in the prototyping area.

I also have an ADC0834 in the prototyping area to digitize a couple of joysticks and that is working fine so far.

Can you try a slower Stamp like the BS2e?

The installation of a bigger cap on the regulator output that meets the manufacture's spec makes sense and I plan to sprinkle some 0.1uf cap around the prototyping area.

I will report back when I get the transreceiver working.

Regards,
TCIII

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
If you are going to send·a Robot·to save the world, you·better make sure it likes it the way it is!

Istha Powron

Success at last.

Someone will say this is obvious however;

Changed BAUD from 4800 Inverted to 9600 True, and it now works like a dream.

Thank you all for your interest and support.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔





The robot is going to lose. Not by much. But when the final score is tallied, flesh and blood will beat the damn monster.
Adam Smith