Vector - Analog Input Discontinuities

Our Vector recently came up from a deployment at 90m (9.5oC).  It was used to power (at battery voltage) and sample (fast - 8Hz) oxygen and nitrate sensors (analog 1 and 2.)  A 'new' (date of manufacture 9/'05) lithium battery (11.72V @ room temp) was installed.

After ~24hrs, the counts from each analog sensor became very erratic.  This corresponded to a gradual flooring in the battery voltage record to ~8V.  The battery voltage then cycled noisily over the next ~12-24hrs between 8 - 9.2V.  Now, after deployment the battery is reading 11.56V.  On previous (successful) deployments, the battery voltage in-situ averages 8.8V (+/-0.2V) with a new lithium pack.

In spite of this behavior (and the manual specification for 9V power), the acoustic functions performed perfectly.  There seems to have been no effect on these signals from the waining battery voltage.

What is the true voltage floor for the Vector?

Is power preferentially supplied to acoustics under low voltage conditions?

As background, the O2 sensor here uses few mAmps.  The nitrate sensor circuit only consumes mAmps as well.  It's basically just an optoisolator (LED) that is activated when the ADV external power comes on.  Neither of these sensors appears to have suffered any internal or external (cable) ground faults.

What other diagnostics do you recommend so that I don't have premature interruption of analog data?

Thanks,
-Chris

Dear Chris

This is the second time this month I have heard about Lithium battery packs that have excessive voltage drops. In our testing of the Tadiran packs, the voltage dropped no lower than 10V and this was reason why we chose to use them.  There could be many reasons why the battery voltage is too low, some of them related to the battery and some related to the instrument, but my short term recommendation is to use alkaline as an alternative.  Meanwhile, we run some tests here to see if we can find out what is going on.

If the battery voltage is lower than 8 V, the Vector will shut down (or not start a burst if the level is this low already at the time of the first sample).  What may have happened in your case is that the instrument starts but the voltage is reduced below 8 V during the acoustic transmit pulse. As a consequence, the power supply to the 5V analog net becomes too low and the analog sensor reading is disrupted.

I will come back with more information when we have checked some of the batteries we still have in stock.

Best regards,

Atle Lohrmann

Dear Atle,

Thank you for your response.  I neglected to mention this Vector was set up to sample continuously.

Considering your thoughts, I checked the burst and ensemble counter.  However briefly, the Vector does indeed shut down - at each 8V  incident, 'burst' ticks up and simultaneously, 'ensemble' recycles.  I would expect 'burst' to remain at 1 through the deployment, and 'ensemble' to cycle repeatedly through 65536.

I'm doubtful that the batteries caused my problem.  I took the pack in question apart and find its diode protected (3 parallel X (3 x 3.9V series)).  I think it would have to have had a bad cell on each circuit - unlikely; not the case anyway.

This application requires continuous multi-day sampling.  The standard alkaline battery pack does not have the amp-hours.  You may recall, I've used an external alkaline pack in the past.  That Vector/battery combo resulted in measurable counts of high-frequency noise intruding on the acoustic/sensor data.  This confounded the goal of the work.

My experience is (depending on temp/depth) that standard lithium packs powering a Vector with two analog sensors sit at +/- 9V.  Would you recommend installing a lithium pack with a slightly higher voltage?

I think now the source of this problem lies with the sensor(s.)  I'm exploring the possibility that there is a pressure-induced ground fault.  I'll let you know what I find.

Thanks for looking at the batteries on your end.  But, I think now  I'll have to look at my sensors more carefully.

Cheers,
-Chris

Hi

The initial characteristics of the battery packs have not changed; we are measuring about 11.5V when they are loaded. We are now running several packs overnight to check whether the discharge characteristics have changed.  

The other application we saw where battery voltage dropped too far was also a continious deployment (Aquadopp profiler).  It is possible that the age has affected the battery ability to recover between each acoustic transmission (ping).

- Atle Lohrmann

Hi again,

We have tested two different batteries to depletion and we have asked the battery manufacturer if it could be a problem with shelf life.  

At this point we believe the low voltage is due to a combination of:
- Working full blast (not letting the batteries any time to recover)
- Low temperature (<10 degree C)

Since you cannot do anything about the temperature, our recommendation to ease up a little on the sampling.  For example, you could have burst that last 15 minutes and then put in a 5-minute break.  This should allow the cell chemistry time to "recover" and avoid the excessive voltage drop.

The batteries you used are about two years old.  The battery manufacturer says the shelf life is 10 years so that age should not matter.  We are getting some freshly produced batteries in October and we will run them through the same intense tests to find out if their statement is correct even for this special situation.


Best regards,

Atle Lohrmann