MAX-M8W-0-00 drops satellites

0 votes

We are using the MAX-M8W-0-00 with a Taoglas active GPS receiver. The unit gets about 9 active satelites and after a while they suddenly collapse to zero. The numbers collapse to zero in about 10 seconds. On one set of data logs, the unit ran for 1, 2 and 5 minutes before GPS stopped. We have several machines that do this. Testing is done in a field.

We do not think that someone is following us with a jammer.

Is the SW code a clue?

What can we do to isolate this bug?

Best regards,  Tim Orr

 

Hello, 

(Text put in the proper place this time)

We have more observations. We fed the MAX-M8W directly with a generated single satellite GPS signal through a coax cable and a UFL connector. This returned a 36dB S/N ratio. Our board gets hot. As the temperature went from +30C to +50C, the S/N ratio went down by 5dB. Would a temperature rise of this size cause a 5dB degradation? A small blast with freezer stopped it from functioning. I am not certain that a freezer test is valid as it would not happen naturally and freezer can be -50C. 

I could not see temperature mentioned in the data sheet other than the maximum limits. Does anyone have experience of temperature dependencies.

Best regards, Tim Orr

 

Hello, 

What is the sensitivity of the MAX-M8W to noise in the power supply? I could not find any information about this. The 'Migration' document seems to not recommend power supply filtering (with resistors or inductors). The supply is +3.3V which is obtained from a switcher with all of its inherent switching noise. Maybe I should have used local LC filtering for the module. 

The PCB generates about 3 Watts of heat and rapidly warms up. The air temperature inside the unit can reach +40C above ambient temperature.

Best regards,  Tim Orr

 

Hi, another question. When using the MAX-M8W to receive satellite data, what is the range of S/N values that can be expected. That is, the maximum S/N to the minimum S/N for satellite acquisition. I do not have a sense of what is considered good and bad.

Best regards, Tim Orr

 

Hello, I have some progress. 

When we use a large GPS/GLONASS antenna on a 3m cable we get lots of satellites with good S/N ratios even though it was raining. This proves that the MAX M8 works well, even with a switched mode supply. We use a Taoglas AP10F GPS antenna. This is a 10 x 10mm patch antenna with a SAW and LNA. When it is close to our electronics, the local RF interference pulls down the S/N ratio by about 12dBs. The antenna can be moved around to observe the best location for maximum S/N performance. We will have to provide some selective screening and add some RF noise reduction to get the best performance.

Yesterday the sun was shining and there were no clouds. Using the AP10F GPS antenna, we got best GPS S/N ratios of 40 and 42dBs and a couple of GLONASS readings.

Best regards, Tim Orr

 

Jan 29 2018 Our old bug has returned. The MAX-8 SN ratio falls and the number of satellites in view declines over a few seconds. After many unsatisfactory tests, I resort to the freezer spray and a heat gun which produced similar results on the MAX8 body. I remember another RF company saying that their super stable TCXO cannot cope with rapid temperature changes. For the MAX8, abrupt temperature changes make the declared SN ratios fall. This seems worse when the starting SN is poor. I blew on the module and we plotted the local temperature and SN ratio collapse and they correlated. I guess, that the local master oscillator drifts off and the performance falls away.

Our machine is a very small drone. The plastic body has optical openings in the top, which were not filled in because they have not yet been made. Under high thrust conditions, the air enters the body and blows on the module which causes this effect. These holes have been sealed with tape and the problem seems to have gone away! 

Can the MAX8 send out its own temperature data?

Best regards,  Tim Orr

 

 

by orrcam asked Jan 12
by orrcam edited Jan 30
51 views
Check the DC bias on the RF_IN pin in working vs stopped working conditions.
Reset receiver to default via UBX-CFG-CFG, make sure that it isn't in some power saving mode.
Check the stability, ripple and noise level of your source of VCC to MAX.
(Edit) use a scope to look for millisecond or microsecond dropouts on VCC.
Watch UBX-MON-HW to assess changes in  AGC and Jamming indicator.
Hello,

We have more observations. We fed the MAX-M8W directly with a generated single satellite GPS signal through a coax cable and a UFL connector. This returned a 36dB S/N ratio. Our board gets hot. As the temperature went from +30C to +50C, the S/N ratio went down by 5dB. Would a temperature rise of this size cause a 5dB degradation? A small blast with freezer stopped it from functioning. I am not certain that a freezer test is valid as it would not happen naturally and freezer can be -50C.

I could not see temperature mentioned in the data sheet other than the maximum limits. Does anyone have experience of temperature dependencies.

Best regards, Tim Orr
>> Our board gets hot.
In an abnormal fashion?

Check the performance of your supply in these conditions, ie mV level noise in the supply voltage, temperature likely to have an impact to external components.

The part has a TCXO, this should have reasonably consistent operation over temperature (performance see UBX-NAV-CLOCK). The operation limits here are likely -40C to 85C at the device.

See bottom of Table 9
https://www.u-blox.com/sites/default/files/MAX-M8-FW3_DataSheet_%28UBX-15031506%29.pdf

For a direct feed from a simulator via coax 36 dBHz seems rather depressing.
0 votes
50 mVpp is what u-blox recommends for power supply ripple.

Under unobstructed view of the sky C/N0 values with good antenna are around 45-50 dB/Hz
by grumpy answered Jan 19
Hello 'grumpy',

Many thanks for your reply and also to others that have replied. Our 3.3V supply has a noise level of about 15mVpp at  580kHz from a switcher. I will try a Microchip Ripple Blocker to reduce the supply noise. A low noise supply has to better than the output from a switcher.

Best regards,  Tim Orr