HIGH RESOLUTION ECG AND DEPTH DATA LOGGER

A Novel Device to Study Breath Hold Diving Induced Variations of the PQ Interval

A. Sieber

1, 2

, R. Bedini

, X. Yong

, A. Navarri

, M. Dalle Luche

, A. L’Abbate

and P. Dario

Profactor Research and Solutions GmbH, Seibersdorf, Austria

Scuola Superiore Sant’ Anna, Pisa, Italy

CNR, Istituto di Fisiologia Clinica, Pisa, Italy

MEMS Center of Chongqing University, China

Keywords: ECG, breath-hold, apnoe, diving, PQ interval.

Abstract: Breath hold diving induces several physiological effects. The authors speculate that next to bradycardia,

vasoconstriction, splenic contraction and blood shift, the form of the ECG and especially the PQ interval are

also affected. Investigations of this effect requires a high resolution ECG monitor which is also capable of

recording depth. This paper describes our data logger prototype. It samples ECG data at 1200 Hz, is

equipped with three temperature and a pressure sensor that allows depth measurement up to 140 m and

stores all the data in an ASCII text file on a SD flash card in FAT 16 or FAT32 file format. The prototype is

then encapsulated in a Lexan tube with an outer diameter of 42 mm and an overall length of 18cm that

should withstand 20 bar pressure equal to 200 m depth.

1 INTRODUCTION

Freediving or breath hold diving history dates back

at least 4500 years ago to pearl divers of the south

pacific. In 1911, one of the first freediving

competitions was held when a Greek fisherman,

Yorgos Haggi Statti, sometimes called “the father of

freediving”, successfully reached more than 60m

depth with a total apnoe close to seven minutes.

Today breath hold diving enjoys a wide popularity,

both recreationally and competitively. Actual depth

record in the breath hold diving discipline No-Limits

is already beyond 200m (Nitsch, H., 11.07.2007,

214m No-Limit, Greece). A big stimulus for this

sport was and still is Luc Besson’s cult film “Le

Grand Bleu” or “The Big Blue”, which depicts the

life of the elite apnoe divers Jacques Mayol and

Enzo Maiorca and their life-long competition in

freediving. However research on breath hold diving

is also shown, where experiments in a mountain lake

in gorgeous scenery in Peru are carried out (Mayol,

J., 2000) to demonstrate bradycardia during breath

hold diving. Other adaptations made by the human

body while underwater and at high pressure include

(Gooden, B.A., 1994), (Andersson, J.P.A., Line,

M.H., Ruenow, E., Schagatay, E.K.A.,2002):

• Vasoconstriction: Blood vessels shrink. Blood

stream is directed away from limbs for the

benefit of heart, lungs and brain.

• Splenic contraction: Releasing red blood cells

carrying oxygen.

• Blood shift: Blood plasma fills up blood vessels

in the lungs and reduces residual volume.

Without this adaptation, the human lung would

shrink and wrap into its walls, causing

permanent damage at depths greater than 30

meters.

Recent experiments with a novel underwater

Doppler-Echography system also demonstrate

cardiovascular changes during breath hold dives at 3

and 10m (Marabotti, C., Scalzini, A., Chiesa, F.

Bedini, R., Reale, L., Passera, M. Belardinelli, A.

Pingitore, A., Cialoni, D., Data P.G,2005). We

speculate that these cardiovascular changes also

affect the timing of the ECG. To investigate possible

variances of the PQ or the PR interval dependent on

the depth, a novel data logger was developed that

combines a high resolution ECG (16 bit, 1200 Hz

sampling rate) and depth recorder in a small,

watertight housing.

269

Sieber A., Bedini R., Yong X., Navarri A., Dalle Luche M., L’Abbate A. and Dario P. (2008).

HIGH RESOLUTION ECG AND DEPTH DATA LOGGER - A Novel Device to Study Breath Hold Diving Induced Variations of the PQ Interval.

In Proceedings of the First International Conference on Biomedical Electronics and Devices, pages 269-275

DOI: 10.5220/0001052102690275

 SciTePress

2 METHODS

2.1 Introduction to ECG

The typical ECG signal (Figure 1) is characterized

by six peaks and valleys labelled with successive

letters of the alphabet P, Q, R, S, T and U

(Malmivuo, J., Plonsey, R., 1995).

Figure 1: Typical form of a ECG signal.

A typical electrocardiogram (ECG) shows peaks

of up to 5 mV. For the proper design of a ECG

amplifier DC components up to +-300 mV, resulting

from the electrode-skin contact, and a common-

mode component of up to 1,5 V, resulting from the

potential between the electrodes and ground, have

to be taken into account. The bandwidth of an ECG

monitoring system, depends on its application. It

ranges from 0,5 Hz to 50 Hz in intensive care units

up to 1kHz for late-potential measurements

(pacemaker detection). A standard clinical ECG

application has a bandwidth of 0.05 Hz to 100 Hz.

A typical PQ interval is about 160ms. A

correlation is given between heart rate and PQ

interval (Atterhög, J., Loogna, E., 1977). Recent

studies have shown that this is not always the case,

for example directly after a heavy workload (Busse,

M., Nißing, A., Tegtbur, U., Miltzow, S., Thomas,

M., Fikenzer, S., 2004). To study variances in the

PQ interval a high spatial time resolution is required

(Ward, S., Shouldice, R.B., Flanagan, M.,

Heneghan, C.,2004).

2.2 Principle Design of the Module

The prototype consists of two boards: the ECG

signal acquisition board including amplifiers and a

high resolution AD converter and the data logger

board, comprising mainly an 8 bit RISC

microcontroller, a display, a Secure Digital memory

card slot and a 10 bar pressure sensor to monitor the

depth.

Figure 2: schematics of the developed module.

2.3 ECG Acquisition Board

For the analogue front ECG amplification stage we

use the typical approach (Company-Bosch E.,

Hartmann, E., 2003) with an instrumentation

amplifier (IA) and a right leg common-mode

feedback op amp. For the IA we have chosen the

AD620 [Analog Devices], a low cost, high accuracy

instrumentation amplifier, with excellent DC

performance: CMRR>>100 dB to nearly 1kHz,

50µVmax offset voltages, low input bias current

(1nA max), and low input voltage noise

(0.28µVfrom 0.1Hz to 10Hz).

Figure 3: Schematic of the ECG amplifier.

The AD620 requires only a single external gain-

setting resistor R

(figure 3). Resistors R

and R

change the normal gain equation to:

Ω+

kkRkGain

22/)2/4.49(/4.491

To avoid output saturation, the usable gain is

limited by the output swing and the maximum input

voltage to the IA. Here, Gain is conservatively set to

9 by choosing RG=6,8 kΩ.

220k

22k

To AD7731

22k

C3 0.1u

1 2

10k

R9 1m21

R8 10k

IC4A

OP491

411

2.5V

IC2

AD620/AD

2 5

OUT

-REF

V+V-

RG1

RG2

100n

IC4D

OP491

411

11k

R10

220k

6.8k

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

270

The OP491 form Analog Devices is used in the

right-leg common-mode feedback circuit. It is a high

precision operation amplifier with a low power

consumption and high common-mode rejection

(70 dB minimum). This circuit applies an inverted

version of the common-mode interference to the

subjects right leg, with the aim of cancelling

interferences. The op amp has a voltage gain for the

common-mode voltage of 91 (R

=1MΩ/11kΩ)

with a low-pass cut off at about 160Hz for stability.

-3dB

=1/2π x 10kΩx0,1µF)). For simplicity, a

passive low-pass filter and a passive high-pass filter

are adopted to accomplish both gain and frequency

selectivity. The filter will allow all signals to pass

through unaffected as long as their frequency is

between the low-pass corner frequency at 160Hz and

the high-pass corner frequency at 0.1Hz

(f-

3dB

=1/2π x 1MΩx1µF)).

The required power for the module is supplied with

a battery. The +5V supply required for the AD620

and OP491 is handled through the voltage power

regulator TPS76950 [Texas Instruments] which

offers the benefits of low dropout voltage, ultra-low

power operation and miniaturized packaging (5-pin

SOT-23 package). The +2,5 V reference for the

AD620 is accomplished with one of the four

operational amplifiers integrated in the OP491

(figure 4).

Figure 4: Power supply of the ECG board.

A AD7731 AD converter from Analog Devices

follows the analog section. It is a low noise, high

throughput 24 bit Sigma-Delta ADC with buffered

differential inputs and programmable low pass

digital filtering allowing adjustment of filter cut-off,

output rate and settling time. The device has a

proprietary programmable gain front end that allows

it to accept a range of input signal ranges, including

low level signals. Figure 5 shows the peripheral

setup of AD7731.

JAD is the interface aimed for the

communication between AD7731 and Atmega32

which is located on the data logger board.

Figure 5: Schematics of AD7731.

2.4 Data Logger Module

The core component of the data logger board is an

ATMEL Atmega32 microprocessor with the

following specifications:

• 32 kbyte Flash Program Memory

• 2 kbyte SRAM

• 1 kbyte EEPROM

• programmable 8 channel 10 bit ADC

• 16 MIPS @ 16MHz

• TQFP44 housing

The board can be programmed via J1. A Secure

Digital memory card connector is connected to the

SPI interface of the µ-processor (PB4-PB7). As the

SD card is powered with 3,3V, three voltage

dividers (R16 – R21) are deployed to decrease the

5V digital output signals to 3,3V. As the Atmega32

interprets digital signals above 2,7V as high, the

digital output can be directly connected to the input

of the SPI interface (MISO, PB6).

Two low drop low power consumption voltage

regulators are used to provide 3,3 V for the SD card

and 5V for the rest of the components. Additionally

the 3,3 V regulator can be switched in standby mode

via dropping the enable pin to ground (via PB2). The

EA-DP204 4x20 characters display can be

controlled in 4 bit parallel, 8 bit parallel or SPI

mode. To simplify the circuit by minimizing the

necessary connections we have chosen SPI mode.

Therefore the displays clock SCLK and serial input

SDI are connected to PB0 and PB1. The ADC’s

ready pin is interfaced to the external interrupt pin

INT1 of the µ-processor (PB3).

To dedicate the µ-processor’s inbuilt SPI bus

solely to the SD card, the routines for interfacing the

IC1

TPS76950

INPUT

GND

OUTPUT

R11

10k

R12

10k

0.1u

IC4B

OP491

411

2.5V

VCC

DIN(J4-5)

RDY(J4-4)

C13

22p

DVDD(J3)

DOUT

C10

0.1u

DIN

ECG Signal

XTA L

4.9152M

C11

IC6A

AD7731

AD7731BN

AGND

REFIN_+_

RDY*

DVDD

CS*

AIN 3/D1

AIN 6

MCLKOU T

AIN 5

DGND

POL

SY NC*

AIN 2

AIN 1

AIN 4/D0

MC LKI N

SCLK

RESET*

DIN

REFIN_-_

AVDD

DOUT

STANDBY*

2.5V

C12

22p

SCLK(J4-7)

R18

RDY

JAD

GND

SCLK

DOUT(J4-6)

DVDD

HIGH RESOLUTION ECG AND DEPTH DATA LOGGER - A Novel Device to Study Breath Hold Diving Induced

Variations of the PQ Interval

271

AD converter and the Display are software

implemented.

The device mainly addresses ECG recordings in

hyperbaric environments like in diving. Therefore a

digital pressure sensor needs to be integrated. For

this reason the MS5521B 14 bar absolute pressure

sensor was chosen [Intersema, Switzerland]. Even if

it is not mentioned in the datasheet, the sensor is

suitable for pressure measurement up to 33bar.

Moreover a digital temperature sensor is integrated

in the part.

This sensor requires a 3.3V supply, therefore

again voltage dividers (R5-R10) are used to reduce

the pins output voltage. The output of the pressure

sensor can be directly connected to the

microprocessor. For correct operation of the sensor a

clock signal is needed (MCLK). This clock signal is

generated with the internal Timer 2 from the Atmel

ATMEGA32.

DVDD

GND

DIN

DOUT

RDY

SCK

-> AD7731

DS18B20

U12

GNDGND

VDD

DS18B20

U13

GNDGND

VDD

TQFP44A

ATMEGA32

PB5

PB6

PB7

RESET!

VCC

GND

XTA L2

XTA L1

PD0

PD1

PD2

PD3

PD4

PD5

PD6

PD7

VCC

GND

PC0

PC1

PC2

PC3

PC4

PC5

PC6

PC7

AVCC

GND

Aref

PA7

PA6

PA5

PA4

PA3

PA2

PA1

PA0

VCC

GND

PB0

PB1

PB2

PB3

PB4

EA-DIP 204-4

4x20 Characters

EADI P204

GND

VDD (+5V)

VEE _ Kontrast (0.. 5)

RS (CS*)

R/W (SDI)

EN (SC K)

D0(SDO)

RES!

LED A

LED K

+5V

U10

TPS76950

EN!

GND

OUT 5V

47k

SP12 CON

J21

CON8

VBATT

+5V

Reed

SW KEY -SPST

J22

CON6

+5V

U11

LP2985-33

VIN

GND

VOUT

By pass

10M

+5V

3V3

MOS I

MI SO

SCK

R22

220k

R23

220k

MOSISCK SS

SD Card

MIS O

R16

3k3

PB5 (6)

PB6(7)

PB7 (8) PB4 (5)

R17

1k8

R18

1k8

R19

3k3

R20

1k8

R21

3k3

3V3

+5V

3V3

Reed2

DS18B20

MS5541B

SCLK

GND

VDD

MCLK

DIN

DOUT

1k8

3k3

J24

CON8

1k8

3k3

1k8

R10

3k3

Figure 6: Schematics of the datalogger board.

Body skin temperature and water temperature

are parameters that influence the physiological

diving response. Therefore two temperature sensors

are integrated in the design. These sensors can be for

example placed under the diving suit to measure the

skin temperature.

As temperature sensors two “one wire” DS18B20

[Maxim] were selected. The “one-wire” feature

allows to use just two lines for power supply,

communication and ground. In principle the two

sensors could be connected to the same port pin, but

to increase communication speed by parallel read

out, each of the sensors was connected to one

separate pin.

2.5 Software

The firmware of the module detailed in figure 7I is

developed in C under the Atmel AVR Studio 4

[Atmel] and the GNU C compiler WinAVR

(http://winavr.sourceforge.net/).

The implementation of a DOS compatible FAT

16 of FAT 32 filesystem on the SD card requires in

total 1,6 kbyte of ROM to mirror the boot sector, the

file allocation table (FAT) and to provide a buffer

for data storage.

Data storage on SD card in FAT 16 of FAT 32

file system is performed in blocks of 512 bytes each.

To enable a high sampling rate of 1200 Hz, it is

necessary to treat data storage and read out of the

ADC separately, as the data storage of a 512 byte

block may last up to 7µs.

Every time an AD conversion is complete, an

interrupt is generated via the external interrupt pin

INT1.

In the interrupt routine the result is stored in a

FIFO buffer. As soon as there are 10 entries in the

FIFO buffer, the data is converted in an ASCII string

and stored together with the actual depth on SD

card.

The “one-wire” protocol [Maxim] is software

implemented. Every two seconds the two

temperature sensors are read out.

The MS5541B is interfaced via SPI bus. Every

two seconds values for pressure and temperature are

read out and stored on SD card. Together with 6

calibration parameters stored on the sensor and

usually read out at the beginning of the program,

pressure and temperature can be calculated. To

reduce overall processing time depth and

temperature are not calculated on the microprocessor

but later under LabView.

A reed contact is used to switch the system on

and off via a magnet (in underwater applications

magnet switches are preferred as they require no

mechanic connection to a switch that needs to be

sealed). Once the magnet is removed, the system

switches in standby mode. Placing the magnet over

the reed contact creates an interrupt on INT0,

waking up the module.

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

272

Figure 7: Software flow diagram.

2.6 ECG Data Processing

A software to preview and analyze the collected data

software was developed under Labview 7.1

[National Instruments]. To suppress the 50 Hz

noise, a optimized Notch filter is implemented (2

order IIR filter, 45-55 Hz). (Josh, Y.V., Dutta Roy,

S.C.,1997), (Chivapreecha,, S., Dejhan, K.,

Yimman, S., 2005) (see also figure 9A and 9B).

3 RESULTS

A first prototype with the following specifications

was build up:

Power consumption

34 mA

Module size 80x30x25mm³

Battery supply 5,5 … 10

Sampling rate ECG 1200 Hz

Resolution ECG 16 bit

Resolution depth sensor 10 bit / 10cm

Sampling rate depth 120 Hz

Bytes / sec 6,4 kbyte/s

File format FAT16/FAT32

Depth sensor 14 bar (33 bar)

Housing, lexan: 250 m rated

The power consumption depends also on the SD card (Kingston

256 MB: 34 mA; Lbd, 32MB: 130 mA). In Standby Mode the

power consumption drops to 0,3 mA.

Moreover the board is equipped with a second

ECG channel with a second AD converter. To

enable sampling of low frequency signals a AD8730

AD converter from Analog Devices was additionally

integrated in the first prototype. It offers 10

programmable channels with a resolution of 16 bit

(application: measurement of skin conductance,

breathing sensor, temperature, etc.).

Figure 8A and 8B: First prototype in the double o-ring

sealed housing.

Figure 9A: ECG raw data.

Figure 9B: ECG raw data after filtering with the 50Hz

Notch filter.

HIGH RESOLUTION ECG AND DEPTH DATA LOGGER - A Novel Device to Study Breath Hold Diving Induced

Variations of the PQ Interval

273

Figure 8A shows the first prototype encapsulated

in the lexan housing. The electronics consist of two

boards, one (yellow-brown) with the µ-processor,

SD card slot and the display and a second (green)

board with the ECG circuit (Figure 8B).

Figure 9A shows a sample ECG recoding. The

Notch filter allows an effective suppression of the

50Hz noise signal (figure 9B). This noise

suppression is especially necessary when recording

the ECG signal on the surface when the diver is

preparing for the apnoe immersion. Once under

water 50Hz noise signals are usually not seen due to

the electrical conductivity of salt water.

Previous experiments have shown that during the

first meters of the immersion electrodes with a

sponge give bad signals. We speculate that this is

due to the small amounts of air next to the sponge

that is getting compressed by increasing deph

causing electrode movement thus causing also

artefacts in the ECG signal. To avoid this problem

Kendall Arbo H34SG [Tyco Healthcare] electrodes

were selected, as there electrode gel is placed over

the electrode without a sponge.

As described earlier (Bedini, R., Reale, L., 2003)

recording of an ECG underwater requires a proper

sealing of the electrodes. Therefore we use two

components impression material (Elite H-D+,

Zhermack Hydrphilic Vinyl Polysiloxane)] (Figure

10A). This works well for short dives like breath

hold dives. If the electrodes are exposed for longer

times like several hours to sea water, we connect the

electrode cables directly to the electrode and seal the

connection with Epoxy (5 min Epoxy, 155105-1,

R&G GmbH, Waldenbuch, Germany).

3.1 Pool Tests

Several tests on apnoe divers were carried out in a

10.5 m deep research pool (Divesystem, Massa

Marittima, Italy).

Figure 10A and 10B: 10A: sealing of the electrodes, 10B:

test dive in the research pool.

For visualisation and analysis of the recorded data a

software was developed under National Instruments

Lab View 7.1 which is also offering several data

processing utilities.

Figure 11: Lab View data visualisation software, in the

upper graph the ECG is visualized, down you can see the

depth profile.

3.2 Long Term ECG Recordings

An adapted version of the system with a to 250 Hz

decreased sampling rate was used to record the ECG

of Francesco Colletta during his world record dive in

Siracusa, Sicily, Italy on September 8th to 9th. The

total dive time was 32h.

Figure 12 A,B: Francesco Colletta after 32h underwater

and our especially for this dive adapted ECG measurement

device (at the end of each of the two black cables one

temperature sensor is encapsulated in epoxy resin).

4 CONCLUSIONS

Research of breath-hold diving induced effects on

the ECG requires a high resolution ECG and depth

storage device. A novel prototype was developed

that is able to store one or two channel ECG with 16

bit resolution at a sample frequency of 1200 Hz on

secure digital memory card. This high sampling rate

is the basis for a precise estimation of the PQ-

interval. The authors are convinced that data gained

from elite apnoe divers during the world

championship in October 2007 in Egypt will lead to

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

274

a better understanding of the physiological effects

of breath hold diving.

Other applications of this device are ECG, depth

and temperature recording during SCUBA dives.

5 FUTURE WORK

4 more prototypes will be fabricated to be then

deployed during the AIDA freediving world

championship in October 2007 in Sharm el Sheikh.

An enhanced second version of this prototype is

planned to address SCUBA diving applications. It

will be expanded with the following sensors:

breathing frequency, water temperature, skin

temperature and skin conductivity.

Software will be developed to allow an

automatic measurement of the PQ-interval. This will

be based on triggering on the QRS complex and then

calculating the time to the beginning of the P wave.

REFERENCES

Mayol, J., 2000, Homo Delphinus The Dolphin Within

Man, ISBN 1928649033.

Gooden, B.A., 1994, Mechanism of the human diving

response. Integr Physiol Behav Sci 29: 6–16, 1994

Andersson, J.P.A., Line, M.H., Ruenow, E., Schagatay,

E.K.A.,2002, Diving response and arterial oxygen

saturation during apnea and exercise in breath-hold

divers, J Appl Physiol 93: 882–886.

Marabotti, C., Scalzini, A., Chiesa, F. Bedini, R., Reale,

L., Passera, M. Belardinelli, A. Pingitore, A., Cialoni,

D., Data P.G,2005, Echocardiographic changes

during breath-hold diving, Proceedings of Blue 2005,

Pisa Dec 1-4, 2005, Ed. Star CNR Pisa.

Malmivuo, J., Plonsey, R., 1995, Bioelectromagnetism -

Principles and Applications of Bioelectric and

Biomagnetic Fields, Oxford University Press, New

York.

Atterhög, J., Loogna, E., 1977, P-R interval in relation to

heart rate during exercise and the influence of posture

and autonomic tone, Journal of Electrocardiology 10

(4): 331 – 336.

Busse, M., Nißing, A., Tegtbur, U., Miltzow, S., Thomas,

M., Fikenzer, S., 2004, EKG-Parameter und

Herzfrequenz bei Belastung II. PQ-Zeit und

Herzfrequenz bei Belastung, Klinische Sportmedizin,

KCS 2004, 5(2): 45-49.

Ward, S., Shouldice, R.B., Flanagan, M., Heneghan, C.,

2004, Electrocardiogram Sampling Frequency Errors

in PR Interval Spectral. Analysis, Proc. IEEE

PGBIOMED’04, Southampton, U.K.

Company-Bosch E., Hartmann, E., 2003, ECG Front-End

Design is Simplified with MicroConverter, Analog

Dialogue 37-11.

Josh, Y.V., Dutta Roy, S.C., 1997, Design of IIR Digital

Notch Filters, Circuits Systems Signal Processing,

Vol. 16, NO. 4, 1997, PP. 415-427.

Chivapreecha, S., Dejhan, K., Yimman, S., 2005, Design

of IIR Notch Filter for Removal of Baseline wander

and Power Line Interference from ECG Signal,

ICCAS2005.

Bedini, R., Reale, L., 2003, Rassegna per immagine sulla

Ativita di Biotelemetria nel Diving.

HIGH RESOLUTION ECG AND DEPTH DATA LOGGER - A Novel Device to Study Breath Hold Diving Induced

Variations of the PQ Interval

275