Control.exe will absolutely not operate until the USB driver for SCAMP has been installed.  

What Microsoft operating system are you using?  

We have tested your SCAMP with Windows 2000 and also Windows 98SE.  If you are using any other operating system, try a different computer that has Windows 2000 on it.   You don't have to install the SCAMP driver,  just look to see if the "Found new hardware" screen appears.  You can select Cancel on the Found New Hardware Wizard screen.   If Windows 2000 recognizes SCAMP you'll have to give up using whatever operating system you have that doesn't recognize SCAMP.

Instructions for installing the driver are contained within the operator's manual at the Software Installation section.  These instructions depend upon the operating system recognizing SCAMP at the USB port.   When you connect the SCAMP for the first time to your computer's USB port, Windows 2000 should recognize it as a un-installed device and display "Found new hardware" as shown in the manual.  If this does not happen, try a different computer that has Windows 2000 on it.  You don't have to install the SCAMP driver, just look to see if the "Found new hardware" screen appears.  You can select Cancel on the Found New Hardware Wizard screen.   If a different computer will recognize SCAMP then you have some problem with the original computer.

If no computer recognizes SCAMP, then replace SCAMP's batteries as described in Maintenance \ Replacing the Batteries section of the manual.  Might as well get good at this since you'll do it often when operating SCAMP.  Try computers again.

If you continue to fail to see the Found New Hardware screens when you connect SCAMP via USB, then go to Device Manager within your Windows 2000.  Connect SCAMP via USB port and observe the Universal Serial Bus controllers part of the device tree.  It will take something less than 10 seconds for Windows 2000 to recognize a USB device and post this info to Device Manager.


Q. We have been trying to get data in the downwards profiling mode, testing the Scamp functioning in the 5 gallon bucket full of water. We have defined all the conditions in the Mission Tab as indicated in the manual, we have awakened the Scamp with the Start magnet (but release motor operates once instead of twice as mentioned in the manual ?) . But we cannot get any data file. What could be wrong?

The magnet causes SCAMP to wake up. It triggers its release motor once immediately to indicate it has awakened.

During the next 2 or 3 seconds SCAMP checks internal resources (batteries and memory) to be sure that it can perform a profile. If there are sufficient resources, then it triggers its release motor twice in quick succession. It then begins to evaluate the profile start conditions you defined.

When SCAMP determines that a profile start condition is true, it begins acquiring profile data and triggers its release motor continuously for 2 seconds (to drop any weight if in upwards mode).

After starting a profile, SCAMP acquires profile data and evaluates stop conditions to determine when to stop acquiring the profile.

Profile data acquisition is halted if a stop condition becomes true or if you connect to SCAMP via USB.

Since your release motor only runs one time, SCAMP is determining that there are insufficient resources for obtaining a profile.

Q. Another question is about the Scamp Control Dialog, in the System Tab. We get nothing in the battery status (no progress bar display on the right, voltage = 0), despite the fact that they are new and that main other features are operating (connection, channel testing, rotation of the release motor, ...)

The batteries should display a reasonable value, something like 7 to 9 Volts. Since they do not, then  I suspect that there is a battery problem. This would explain you problem in question 1 above.

Q. In the Test tab, when we test the Fast T0 channel for instance, it seems that the display does not give the right temperature values. The water we are using is about 18°C and the display shows us 13°C. What does this test mean exactly? Other channels are also giving dubious results.

If the batteries aren't working, then sensor values will not be correct.

Your objective is to get the battery indication on the System Tab to give reasonable values. Sensors and missions won't work until reasonable values are displayed. 

SCAMP has three battery packs:  two 9 V transistor batteries, and a 6 X 1.5 V AA battery group. The CPU operates from the 6 X 1.5 AA battery group. The sensors and A/D system operate from the two 9 V transistor batteries. Since SCAMP will communicate via USB I suspect that the 6 X 1.5 AA batteries are not dead. However if either 9 V battery is dead then the A/D will not work and SCAMP will not have a way to determine the voltage for any battery including the 6 X 1.5 AA. This is how it will communicate even though the battery levels it displays are not reasonable.

I suspect that you have problems with one or both of the 9 V transistor batteries or with the fuses connecting them to SCAMP.

Open SCAMP, then remove all batteries. Measure their voltage. If either battery gives an unreasonable value, then replace it. If you get reasonable results, then the 9 V battery fuses may be blown. This can happen if you momentarily touch the 9 V battery backwards onto the connector or from some other unknown cause.

See the Maintenance|Replacing the Fuses in the SCAMP manual. Prior to doing anything, remove the 9 V battery harness (red,black, white wires and battery snaps). You will probably need to take off one electronics cover (Maintenance|Electronics Covers). Note how the connector is plugged on before you remove it. You MUST plug the connector back on in exactly the same way. It is not polarized and can be plugged on in two rotations.

Measure the resistance of each fuse. You can see the fuse on the red and white wires. Red goes to a 9 V battery (+) and White goes to the other 9 V battery (-) so you can guess which terminal in the battery snap to touch the Ohm meter to. You should measure about 6 Ohms resistance in each fuse. If you measure greater than 10 Ohms, the fuse is blown. If either fuse is blown then find the replacement 9 V battery harness we shipped with SCAMP. Measure its fuses. This is new and should give the correct 6 Ohm reading. If not then you've done the measurement wrong and must re-check things until you get the proper reading.

If you determine that the 9 V battery harness that you removed from SCAMP has a blown fuse and that the replacement harness is OK, install the replacement harness. Be sure you plug the connector on properly. Re-install the batteries, 9 V first, then 6 X AA. Be sure you don't touch the 9 V batteries the wrong way to the connectors.

You don't have to put the electronics cover on or close SCAMP.

Connect SCAMP via USB and run the control dialog. Batteries should indicate reasonable voltages. If so, put cover back on, close SCAMP, continue with testing.

If the fuses in the battery harness you remove are ok, or if after you replace them SCAMP still doesn't give a good battery reading then there are other problems.

Please tell me what resistance you measure and whether you get SCAMP going as I've described here.


Q. Why won't the MSDOS Host program, or the new Windows 2000 Matlab software load profiles produced by the MSDOS Host program?

You sent two profiles, both recorded in August 2003.  Each profile consists of three files: *.txt, *.raw,  and *.cfg.  Both cfg files seem to be corrupted.

The problem is not with the new Matlab software only; the profiles do not load properly into the old Host program either.   *.CFG files are very mechanical things and should have all information in exact positions within the file.  I looked at historical *.cfg files from and also at *.cfg  files from other SCAMP units.  All historical files from all units that I reviewed have 6127 bytes.  However the two *.cfg files that you sent have  6148 bytes.

As for the two files from Aug 2003 having 6148 bytes - either they were written wrong at the time they were collected or they have been corrupted during storage or handling since they were written.   Since files prior and after 2003 seem correct, then I suspect storage or transport, especially if they have been transported through or stored on a Unix or Linux system.

Customer reply:  Thanks for the reply. I figured out that all .cfg files are 6127 bytes in size on our MSDOS platform. It was the same even in Linux (which we prefer to use) and when I transferred the files from Linux to my Windows laptop to use your software, it jumped up to 6148. This is really funny to me, never realized this could happen. Anyway, now I changed all my files with the correct file sizes and it seems to work pretty well.

Q. I have question about the .dll called in Matlab m files. How can a .dll file be edited (to see what it is exactly doing)?

*.dll files cannot be edited by you.  You must have the source *.c file, which you edit and then compile to create the *.dll.  I do not distribute the source files with SCAMP, but will provide individual files in response to a specific need.  I've attached the source file for s_segmen.c so you can see what it does.   Now that you have this file, you could edit it and then compile it.  Type 'help mex' into the Matlab command space.  I think this brings back some information about mex files.  In addition to Matlab you'll need a C compiler.  I use Microsoft Visual C++ 6.0.

You'll also find that the *.dll are complied specifically for use on a Windows operating system.  They cannot be used with Linux or Unix or any other OS.  However others have compiled them for Linux and maybe even Unix. 

SCAMP's main objective is to determine the turbulent rate of energy dissipation. This is done by matching observed temperature gradient spectra to theoretical Batchelor spectra for a passively transported scalar. Where matches can be found the dissipation is computed from kB.

This method works well where the water is non-thermally uniform, but depends upon knowledge of the spatial distribution of temperature since the Batchelor spectral form is a function of wave number. Unfortunately, SCAMP can only measure temperature at fixed time intervals. The time-spaced interval is transformed to the distance-spaced interval required for Batchelor fitting by using SCAMP's velocity through the water. Since SCAMP is a vertically profiling device, this velocity can be determined from the rate of change of hydrostatic pressure.

In a moored situation some other method of determining velocity must be provided. Yes, you can moor SCAMP but it will not, of its own measurements, be able to determine dissipation. You must provide some other way to measure water velocity past SCAMP's temperature sensor.

Assuming that you provide some independent velocity measurement there are yet more subtle problems. SCAMP's temperature sensor cannot respond to temperature variations that occur too rapidly. This limits SCAMP's use to roughly 10 cm/sec. This velocity is established when profiling by adjusting SCAMP's buoyancy. In a moored situation you'll have to take what you get. If you expect 5 - 12 cm/sec SCAMP may work.


Q. I would like to use SCAMP deeper than 100m. Is this possible?

SCAMP is designed for 100 meters maximum depth. We have no experience with its use at deeper points. SCAMP's pressure transducer is a 10 Bar device (145 PSI). At 200 meters it will be 2X range and may be permanently damaged. This can be swapped for a higher pressure sensor, but PME can only calibrate to 70 meters so the higher range would have to be extrapolated unless the customer does the depth cal.

Also , I can't promise that the system will be free of leaks or will resist implosion. My gut feeling (about 80% confidence) is that it will survive. I wouldn't use a Fast C sensor, but the other sensors (Acc CT and Fast T)  should be OK at 200 meters. F-meter shouldn't have any problems either.

So I think SCAMP will work but you should plan to install a pressure sensor with the appropriate range.

I strongly suggest that if you intend to use SCAMP deeper than 100m, make the necessary modifications, and then thoroughly test it at 200 meters. Better to have it implode in a test chamber then at the end of a retrieval line.>

SCAMPs built with serial numbers ranging from SN0001 to SN0030  hold data internally, but have space for only 100 meters of 8 channel data. We are re-designing the SCAMP to contain up to 8000 meters of 8 channel data in multiple files. These SCAMPs can be retro-fitted with this new feature.

Q. This experiment I am doing in Australia will be in a fairly benign, calm, sheltered piece of water. Would reducing the free-fall rate (sy to 5 cm/s) be a useful thing for me to try? Could I get better vertical resolution in the dissipation?

Sorry, I don't have a yes or no answer for you.  Yes, you'll increase the wave number that you can resolve with FP07.  Maybe not as much as you think since there are boundary layers around the thermistor that grow thicker as velocity decreases.  There is the "frozen turbulence" Taylor hypothesis which offers less guidance as velocity decreases.  SCAMP's travel rate for slower speeds may become less steady... No quantitative information however. I don't know of anyone who has experience or comparisons at 5 cm/sec so no help there.  Suggest you simply try both, then look at the data.  I wouldn't just take 5 cm/sec data only however.  A careful investigation of fall rate with comparisons to other instruments would make a nice paper!

Q. Just a quick question about the fluorometer range.  I noticed that on a couple of casts the fluorometer pegged at the top of the range (not too surprising, there's a lot of chlorophyll here). Is there something in the calibration constants that I can alter to reduce the sensitivity and increase the range if necessary. I'm not sure what to expect in the estuary I'll be working in, so want to be able to cope with anything.

Yes, there are two ways to change the fluorometer scale.  Some background:

Fluorometer signal begins as light pulses received by a photodiode within SCAMP's end cap.  Photo diode converts these to small current pulses.  These are converted to voltage pulses by the first stage amp on the fluorometer circuit, then AC amplified by a second stage.  Third stage is an adjustable gain (potentiometer on the 5500 board - only one inside SCAMP!) DC amplifier.  After this the signal is rectified, filtered, and passed to SCAMP's A/D card by two more stages.  A/D card has digitally controllable offset and gain circuits that add another point where overall sensitivity can be changed.  Ok, so there are two points where the signal can be adjusted: potentiometer and A/D card.

Changing the A/D card is easy.  Just go to SETUP\CHANNELS\FLUOROMETER and increase the max/min range shown there.  This simple change could solve the problem, but maybe not.  It turns out that the actual voltages of the pulses within circuits on the 5500 board must not exceed processing ranges on the board.  This can only be determined by looking at the pulses on an oscilloscope under the condition of max C.  If your high C range causes the circuit voltage to exceed limits, nothing permanently bad happens but you'll see a slope in the otherwise straight calibration curve in the high C regions.  Your max/min change will only expand the range before the A/D clips.

It is safe, however, to DECREASE the circuit gain from the present setting.  This is done by turning the pot (a 20 turn device) screw a few turns as required CCW.  Remember how many turns in case you want to turn it back up again later.  If you decrease the circuit gain then you'll maintain linearity and at the same time expand the effective A/D clipping range.

Q. What is the recipe for Copro?

Mix 1000 micro-grams of copro powder in 100 ml of 6N HCl.  Then add distilled water to make 2 liters.

I purchased my copro from Sigma Chemical Co 1-800-325-3010.  Their part number is C-7157 for 1 mg of the substance.  The proper name is Coproporphyrin III Tetramethyl Ester.  1 mg isn't much, so I simply put the entire bottle in 100 ml and shook until the powder dissolved, then removed the bottle.  I doubt that the resulting 2 liter concentration of 500 ug per liter is very accurate, but you can tell if the f-meter changes its sensitivity if you use the same solution.  Wrap aluminum foil around the storage bottle and store in a dark place.  They say it lasts for a while.

As I remember I mixed 500 ug/l concentrations.  I had to trust the supplier to weigh the sample correctly.  I think I received 1000 ug in a little bottle.  I measured two liters of water, then completely rinsed the bottle in them.  I stirred then poured the solution into two 1-liter bottles which I sealed and wrapped with aluminum foil.  I've had my solutions for several years now.


Q. Do you think SCAMP, or any of the individual sensors, will get upset about high concentrations of hydrogen sulfide?

uDO won't like this.  H2S is known to give Clarke type oxygen sensors problems.  I'm not really sure what happens to them however.  No experience on this score.

Q. I changed the Acc CT sensor, and also put on a new DO sensor. I tried a communicate with SCAMP, and the serial connection worked. I then tested all the channels, and they all gave by total rubbish. Both C (Acc and fast) gave something like 40 (absolutely constant over the 100 samples), and temperatures were both large and negative. I asked for the battery voltage, and got a circuit temperature of 154 degrees (it's warmer in Australia, but not that warm) and all three battery voltages were -33 V. I took out the batteries and replaced them; same results. I then disconnected the two new sensors that I installed, again same results. I've had a careful look around and I am sure that I have not dislodged anything. Any ideas?

My bet is that you've blown a fuse in the 9v battery connection cable.  It's really easy to do.  It sometimes happens if you it the transistor cans when installing the A/D side shield.  Here's how you can tell:

Method 1: Take off the A/d side cover.  This reveals the 5026 board.  Look in your manual in the back for pictures of the boards.  You'll find the 5026 shown there if you don't know which one it is.  This is the A/D board.  Get SCAMP to awaken by communicating with it.  Measure the voltage from each T0-5 transistor can to the chassis.  These should be almost exactly equal to the battery voltage, about 9.2 volts for fresh batteries.

Method 2:  Take off the A/d side cover.  Unplug the 9V battery connection wire.  Measure resistance from battery terminals to corresponding pin on 3 pin board connector.  Two of the battery connector terminals will go to the center pin (black wire)  A (-) connector will go to the white wire pin and will have the fuse resistance of about 7 ohms.  A (+) connector will go to the red wire pin, again with 7 ohms resistance.  Blown fuses do not entirely open, they have blown resistances in the 1000 ohm range.

It is also possible to blow the fuses by contacting any of the many power supply pins on the boards so maybe there.  My money is on a blow out due to the covers somehow.  I favor this since it seems you described the problem as occurring right after sensor replacement, with no normal operation interval.


Q. T1 was removed and I had switched off the channel and forgot to switch off the gradient channel for T1. What are the effects of this?

When no sensor is installed the + terminal of the op-amp that services Fast T1 is simply open - not connected to anything.  The circuit output is undefined in this case but I've not seen any problems.  It is not impossible however that Fast T1 oscillated badly and maybe swamped out both gradient channels.  All gradient gains are implemented by the same IC, I use a dual DAC, one for each gradient channel so channel-channel cross talk at higher frequencies is not impossible. 

Q. How do I test the gradient channels?

Disconnect a Fast T sensor and replace it with the circuit shown.  Set the oscillator for 10 hz.  Record a short profile.  You should see a sinusoidal variation of temperature on the Fast T channel, and its gradient on the Grad Fast T.  This should work for both T0 and T1 provided you connect to the proper sensor port.  One thing to watch out for:  It is possible to drive the gradient circuit into over-range internally and not see the clipping distortion in the recorded data.  (The over-range point occurs before the anti-alias filter and the action of the filter removes over-range clipping features from the data.)  Do the math on the Fast T channel and check that the peak values don't exceed the limits shown by HOST for the Grad Fast T channel at whatever gain you are using.  You may have to turn the oscillator amplitude way down...

Another simple test might be to stratify a bucket of water, then profile SCAMP through it by hand, being careful not to crash the sensors.  You could review the profile data, verify that the gradient is bad, then connect a dummy sensor and repeat.  Maybe the gradient will clean up.  I don't know for sure that the gradient problem is related to the lack of a Fast T1 sensor, but it seems most likely at this time.

(feedback from:  Jonathan Sharples, University of Southampton)     We used SCAMP from a small vessel (freeboard about 2.5 metres), deploying over the windward side. With 200m of cable we could profile down to about 40 to 50 metres before the wind drift of the ship made the line too taut (and it was not very windy, about 5 - 10 knots). SCAMP did not like swell too much either, and in swell  1 - 2 metres tended to somersault in the upper few metres (novel use of the PAR sensor....). Good data was not recorded until about 10 metres depth due to a combination of swell somersaults and the ship's wake.

We used a modified spinnaker boom to deploy SCAMP, which kept it away from the ship's side.

Paying out the line from the hand-held roll was ok (but have a good strong lanyard connecting the cable roll to the ship!) Pulling SCAMP back to the surface against the ship's drift was a little scary initially, as the line gets rather taut. But, we did not get close to the breaking strain; you just have to overcome your fear. 

Recovery onto the boat was not so easy, but waiting for the boat to roll toward SCAMP and hauling inboard worked ok (we were not using the conducting cable: hauling SCAMP out of the water using that may not be a good idea).

Note also, profiling to 40 or 50 metres means you don't get many profiles to combine for statistically reliable dissipation estimates. We had to use the longer 200m non-conducting cable, so data downloads took a lot of time. Ultimately, we were able to estimate tidal averages of dissipation with confidence, but could not reliably resolve intra-tidal signals.

(feedback from Barry Ruddick, Dalhousie University)  -  I've actually been afraid to use SCAMP from a large ship.  The two main factors are:

1.  The 10 cm/s fall rate means that the ship drifts away from the instrument faster than it falls.  Profiles are very limited in depth range because of that.

2.  If it's windy, the drag plate could cause the instrument to swing and bang against the side of the ship, breaking sensors. I think it might be possible to use SCAMP from a relatively small ship, anchored in protected waters.  I would try using it from the forepeak, so the instrument can swing without hitting the ship, and to prevent the instrument from seeing the wake.

The certain way is to use SCAMP from a launch, so it can be set directly into the water.  John Dower and I have done this without problems.

(feedback from John Dower, University of British Columbia) -  Barry and Jonathon are right, deploying and recovering SCAMP from a ship is a nail-biting experience!  I found that deploying from a launch or (better yet) an inflatable boat was much easier on my blood pressure.  In any case, even if it were easy to deploy from a ship the amount of turbulent noise created by the ship makes it preferable to get away from the ship in a small boat.

Even from a small boat, however, the downwind drift of the boat can make it hard to get deep profiles.  My best results came when I measured off 100m of the kevlar line, attached a small float to it and then tossed (well, lowered, actually) the whole thing (i.e. SCAMP, the kevlar line, and the float) over the side of the inflatable and then simply followed the float until SCAMP had reached the desired depth.  Just make sure the float is sufficiently large!

The only other problem I encountered was condensation inside the housing when SCAMP was used in *really* cold waters.  In coastal Newfoundland, where I've used SCAMP since 1997, the summer air temperature can be a muggy 25C, the surface temperature can be 12C, while at 50m depth the temperature falls to about -1C. When I opened the housing after a few such casts (to change batteries) I found a fine layer of condensation all over the inside of the housing. After that, I always taped a few of those small desiccant packs inside....that seemed to solve the problem.

Short of that, the only thing that I find to be a bit annoying is the long time taken to download a deep cast from SCAMP while siting in a tiny rubber boat in the wind, rain, and hail!

Q. Can you just confirm that the engineering units were calibrated as micro Einstein's per meter squared per sec?

Q. I have a query about the circuit temperature. It seems to be reading too low. In the field today the air temp was about 16 C and the water 14 C, but the circuit T always said about 10 C.