- Email: [email protected]
Modification of a consumer digital audio tape (DAT) for analog data recording
of Neuroscience Methods, 45 (1992) 155-158 0 1992 Elsevier Science Publishers B.V. All rights
155
Journal
reserved
0165-0270/92/$05.00
NSM 01429
Modification of a consumer digital audio tape (DAT) for analog data recording Christian Department
Liischer,
(Revised
Key words: Electrophysiology;
Single-channel
Peter Frei and Denis de Limoges
of Physiology,
Data recording
version
Unir,ersity
(Received received
acquisition;
of Berne, Berne (Switzerlund)
6 May 1992) and accepted 7 August
Patch
clamp;
1992)
Data
storage;
A-D
conversion;
A modification of a commercially available digital-audio tape (DAT) for DC recordings such as those used in patch clamp and other electrophysiological experiments is introduced. The modified DAT can record data from 2 channels plus a separate trigger signal for up to 2 h. The bandwidth ranges from direct current to 21 kHz; the signal-to-noise ratio is above 90 dB and the A-D resolution is 16 bit. A 120-min cassette yields a storage capacity of 1.4 Gbytes. The trigger is ‘hidden’ on I channel as a burst of l-ms duration, thus minimizing interference with the recorded data. The modified DAT offers a low-cost yet high quality solution to most analog data storage needs.
Introduction
In electrophysiological experiments continuous data recording is a crucial step in data flow in the laboratory, since most researchers do analyze data off-line. Even with the advent of integrated systems for data acquisition by microcomputer, tape storage of raw data has not been replaced, especially in the case of long-term high temporal resolution. Disadvantages with high capacity RAM disks are limited recording time and slow transfer to Winchester-type hard disks. Analog FM tapes are somewhat cumbersome, have a limited signal-to-noise ratio and are very costly. Previous solutions to this problem have used a combination of a digital audio processor and a video cassette recorder (Bezanilla, 1985). Newer digital audio technology offers an even simpler and less expensive solution: DATs offer a maxi-
Correspondence: Christian Liischer, Department of Physiology, University of Berne, Biihlplatz 5, CH-3012 Berne, Switzerland. Tel.: (41-31) 65 87 49; FAX: (41-31) 65 46 11.
ma1 storage capacity of 1.4 Gbytes (2 h at 2 * 48 kHz with 16-bit resolution) at low cost with the possibility to set identification marks (subcodes). Absolute time display and identification subcodes permit easy retrieval of data. The small cassettes are easy to handle and do not need much space. The use of a remote control can make the work comfortable. The consumer DATs do, however, have some drawbacks such as AC coupling, variable gain and impossibility to record a trigger without using an entire channel. These disadvantages make direct application of these devices impossible. It is the purpose of this article to give a detailed instruction for the modification of a commercially available DAT with respect to continuous recording of analog data. Description
Firstly, the modifications required for the Sony DTC-55ES will be described. All part numbers refer to this model. Diverging modifications are
156
then described in the appendix for the JVC Model XD-Z507. This communication may also serve as a guide for the modifications of other units available commercially.
0 -5 -10
368 level at 23.2kHz
-15
DC coupling
In the DATs that we have examined, the AC coupling which is present in consumer audio devices is made by analog as well as digital mean. In order to remove the analog AC coupling we replaced the 4 capacitors at the signal-input (C120, C121, C220, C221) by 100 0 resistors. The same was done at the output of the amplifier (replace Cl10 and C210 by 100 R resistors). Digital DC coupling is achieved by cutting the low-cut filter LCF (pin 12) of the digital filter (IC306) from line 10 and connecting it to ground. After these modifications we measured perfect DC stability (less than 1 mV shift/h) and 3 dB limit at 23 kHz (Fig. 1). Distortion to a step of voltage is less than 0.005% (1 kHz) according to the manufacturers specifications. Input and output amplification
For obvious reasons consumer audio tapes do have easy adjustable input and output amplification. In the original configuration the DTC-55ES has a maximum input sensitivity of f 660 mV and a maximum output of rt3 V. For our purpose we found it more convenient to have fixed gains adapted to the current or voltage clamp amplifier used in the laboratory. We adjusted the input to + 1 V and the output to +5 V, but as mentioned these values depend on the devices to which the DAT will. be connected (many microcomputer interfaces work at a +5 V range). In order to permit a maximum of flexibility, we have disconnected the original voltage divider at the input and constructed a similar circuit on a separate print using trim potentiometers. At the output stage we have added an additional amplifier (shown in Fig. 2) that has to be connected to ‘line out’. Two trim potentiometers allow one to eliminate any offset and to set the gain. Trigger circuit
For many experiments it is inevitable to have a trigger impulse which usually has transistor-tran-
frequency (kHz]
!w
b-----I
Original
Fig. 1. Sample recording and frequency responses of the modified DAT (JVC XD-Z507). The example shows an action potential (AP) of a rat embryonic dorsal root ganglion cell in culture. The AP is elicited by intra-cellular injection of a 0.5 nA current through a 60 MR electrode. The sweep was initially filtered at 10 kHz and recorded on the DAT. The signal was then played back and digitized using a microcomputer interface (MacLab ADI Instruments, Castle Hill, Australia) at a sampling frequency of 20 kHz (upper sweep labelled ‘taped’). The lower sweep represents the same transient in its original state, bypassing the DAT. The figures were printed on a conventional laser printer. No detectable difference is noticeable. The insert shows the bandwidth of the modified DAT. The sinusoidal test signal had an amplitude of 1 V U,,. The 3 dB limit is at 23.2 kHz. Notice the very steep drop at frequencies above 21 kHz due to digital filtering.
Fig. 2. Adjustable output amplifier. The circuit of an adjustable amplifier is shown. Two such units (1 for each channel) were connected to the ‘line out’. The 2 potentiometers allow adjustment of gain and offset. Maximum output is +6 V. Position b indicates the connection with the trigger circuit.
sistor logic (TTL) level. In order to avoid using an entire channel for this signal, we have designed an electronic circuit to superimpose the trigger with low interference to channel B (Fig. 3). In brief, any incoming trigger signal is converted
channel B
into a 1-ms burst which activates a 20 kHz/2.5 VpP oscillator. The oscillator signal lasting 1 ms is written to channel B (Fig. 3a). It has to be noted that during this 1 ms all signals on channel B will therefore be lost. We have, however, made the
Input ,Ok
__---, channel
B voltage devider
A
wggerinput
l-l..
..n
B
Fig. 3. Trigger input and output. A: the circuit which initially transforms an incoming trigger into a defined pulse of I-ms duration by using part B of the monostable multivibrator (CD4538). This signal then activates the oscillator (NEziSS) which write5 consequently a gated burst to channel B. Notice that during I ms all incoming signals to channel B will be lost. B: the circuit that transforms the sinusoidal signal back into a TTL pulse using part A of the dual monostable multivibrator (CD4538). Position 11 indicates connection with output amplifier.
158
experience that virtually all recording protocols permit a short gap on 1 channel, especially when the microcomputer interface permits prestimulus trigger mode. In play-back mode this oscillating signal is detected by a filter, compared to an adjustable trigger level and converted into a TTL signal (Fig. 3b). Time variability of the trigger has been measured to be below 50 ps (trigger signal generated at the second oscillation of the burst signal instead of the first). Theoretically it is possible that a signal of channel B will erratically incite the trigger. Such a signal would have to have harmonics > 17 kHz with an amplitude of 90% of saturation lasting at least 0.1 ms. In our tests, no biological signal ever was confused for a trigger.
Appendix Modification of the model JVC XL)-Z507
This model is almost identical to the previous regarding its technical performances. The modification described above applies with little adjustment; it is even somewhat simpler since there is no digital AC coupling. The analog AC coupling can be removed by shorting 4 capacitors of the input circuit (ClOl, C201, C108, C208) and by replacing the 2 capacitors Cl53 and C253 of the ‘line out’ by 100 fi resistors. The implementation of fixed gain and trigger is identical to the one described for the Sony DAT.
Acknowledgements Discussion We have tested and successfully used our implementation in triggered as well as continuous recordings. Compared to the FM tape used before, we were pleased with the ease of use. We have also tested the frequency and transient responses as well as the DC stability (Fig. I>. We think that this modification offers a useful low-cost solution for many researchers in various fields who need to record analog data continuously. Although we have not implemented our modification an a portable DAT, this could be a very convenient solution for field work.
We wish to thank Sony Switzerland as well as Spitzer Electronics (JVC Switzerland) for their collaboration and for providing the DATs. We also thank Drs. H.-R. Liischer, E. Niggli and J. Streit for helpful comments on the manuscript. Supported by Swiss National Science Foundation (Grant No. 31-27553.89).
References Bezanilla, F. (198.5) A high capacity data recording device based on a digital audio processor and a video cassette recorder. Biophys. J., 47: 437-441. The Sony DTC-55ES. In: Service Manual. Sony Corporation Audio Group. JVC Service Manual. Digital Audio Tape Deck. XD-Z505. Victor Company of Japan. Maebashi-city, Japan.
155
Journal
reserved
0165-0270/92/$05.00
NSM 01429
Modification of a consumer digital audio tape (DAT) for analog data recording Christian Department
Liischer,
(Revised
Key words: Electrophysiology;
Single-channel
Peter Frei and Denis de Limoges
of Physiology,
Data recording
version
Unir,ersity
(Received received
acquisition;
of Berne, Berne (Switzerlund)
6 May 1992) and accepted 7 August
Patch
clamp;
1992)
Data
storage;
A-D
conversion;
A modification of a commercially available digital-audio tape (DAT) for DC recordings such as those used in patch clamp and other electrophysiological experiments is introduced. The modified DAT can record data from 2 channels plus a separate trigger signal for up to 2 h. The bandwidth ranges from direct current to 21 kHz; the signal-to-noise ratio is above 90 dB and the A-D resolution is 16 bit. A 120-min cassette yields a storage capacity of 1.4 Gbytes. The trigger is ‘hidden’ on I channel as a burst of l-ms duration, thus minimizing interference with the recorded data. The modified DAT offers a low-cost yet high quality solution to most analog data storage needs.
Introduction
In electrophysiological experiments continuous data recording is a crucial step in data flow in the laboratory, since most researchers do analyze data off-line. Even with the advent of integrated systems for data acquisition by microcomputer, tape storage of raw data has not been replaced, especially in the case of long-term high temporal resolution. Disadvantages with high capacity RAM disks are limited recording time and slow transfer to Winchester-type hard disks. Analog FM tapes are somewhat cumbersome, have a limited signal-to-noise ratio and are very costly. Previous solutions to this problem have used a combination of a digital audio processor and a video cassette recorder (Bezanilla, 1985). Newer digital audio technology offers an even simpler and less expensive solution: DATs offer a maxi-
Correspondence: Christian Liischer, Department of Physiology, University of Berne, Biihlplatz 5, CH-3012 Berne, Switzerland. Tel.: (41-31) 65 87 49; FAX: (41-31) 65 46 11.
ma1 storage capacity of 1.4 Gbytes (2 h at 2 * 48 kHz with 16-bit resolution) at low cost with the possibility to set identification marks (subcodes). Absolute time display and identification subcodes permit easy retrieval of data. The small cassettes are easy to handle and do not need much space. The use of a remote control can make the work comfortable. The consumer DATs do, however, have some drawbacks such as AC coupling, variable gain and impossibility to record a trigger without using an entire channel. These disadvantages make direct application of these devices impossible. It is the purpose of this article to give a detailed instruction for the modification of a commercially available DAT with respect to continuous recording of analog data. Description
Firstly, the modifications required for the Sony DTC-55ES will be described. All part numbers refer to this model. Diverging modifications are
156
then described in the appendix for the JVC Model XD-Z507. This communication may also serve as a guide for the modifications of other units available commercially.
0 -5 -10
368 level at 23.2kHz
-15
DC coupling
In the DATs that we have examined, the AC coupling which is present in consumer audio devices is made by analog as well as digital mean. In order to remove the analog AC coupling we replaced the 4 capacitors at the signal-input (C120, C121, C220, C221) by 100 0 resistors. The same was done at the output of the amplifier (replace Cl10 and C210 by 100 R resistors). Digital DC coupling is achieved by cutting the low-cut filter LCF (pin 12) of the digital filter (IC306) from line 10 and connecting it to ground. After these modifications we measured perfect DC stability (less than 1 mV shift/h) and 3 dB limit at 23 kHz (Fig. 1). Distortion to a step of voltage is less than 0.005% (1 kHz) according to the manufacturers specifications. Input and output amplification
For obvious reasons consumer audio tapes do have easy adjustable input and output amplification. In the original configuration the DTC-55ES has a maximum input sensitivity of f 660 mV and a maximum output of rt3 V. For our purpose we found it more convenient to have fixed gains adapted to the current or voltage clamp amplifier used in the laboratory. We adjusted the input to + 1 V and the output to +5 V, but as mentioned these values depend on the devices to which the DAT will. be connected (many microcomputer interfaces work at a +5 V range). In order to permit a maximum of flexibility, we have disconnected the original voltage divider at the input and constructed a similar circuit on a separate print using trim potentiometers. At the output stage we have added an additional amplifier (shown in Fig. 2) that has to be connected to ‘line out’. Two trim potentiometers allow one to eliminate any offset and to set the gain. Trigger circuit
For many experiments it is inevitable to have a trigger impulse which usually has transistor-tran-
frequency (kHz]
!w
b-----I
Original
Fig. 1. Sample recording and frequency responses of the modified DAT (JVC XD-Z507). The example shows an action potential (AP) of a rat embryonic dorsal root ganglion cell in culture. The AP is elicited by intra-cellular injection of a 0.5 nA current through a 60 MR electrode. The sweep was initially filtered at 10 kHz and recorded on the DAT. The signal was then played back and digitized using a microcomputer interface (MacLab ADI Instruments, Castle Hill, Australia) at a sampling frequency of 20 kHz (upper sweep labelled ‘taped’). The lower sweep represents the same transient in its original state, bypassing the DAT. The figures were printed on a conventional laser printer. No detectable difference is noticeable. The insert shows the bandwidth of the modified DAT. The sinusoidal test signal had an amplitude of 1 V U,,. The 3 dB limit is at 23.2 kHz. Notice the very steep drop at frequencies above 21 kHz due to digital filtering.
Fig. 2. Adjustable output amplifier. The circuit of an adjustable amplifier is shown. Two such units (1 for each channel) were connected to the ‘line out’. The 2 potentiometers allow adjustment of gain and offset. Maximum output is +6 V. Position b indicates the connection with the trigger circuit.
sistor logic (TTL) level. In order to avoid using an entire channel for this signal, we have designed an electronic circuit to superimpose the trigger with low interference to channel B (Fig. 3). In brief, any incoming trigger signal is converted
channel B
into a 1-ms burst which activates a 20 kHz/2.5 VpP oscillator. The oscillator signal lasting 1 ms is written to channel B (Fig. 3a). It has to be noted that during this 1 ms all signals on channel B will therefore be lost. We have, however, made the
Input ,Ok
__---, channel
B voltage devider
A
wggerinput
l-l..
..n
B
Fig. 3. Trigger input and output. A: the circuit which initially transforms an incoming trigger into a defined pulse of I-ms duration by using part B of the monostable multivibrator (CD4538). This signal then activates the oscillator (NEziSS) which write5 consequently a gated burst to channel B. Notice that during I ms all incoming signals to channel B will be lost. B: the circuit that transforms the sinusoidal signal back into a TTL pulse using part A of the dual monostable multivibrator (CD4538). Position 11 indicates connection with output amplifier.
158
experience that virtually all recording protocols permit a short gap on 1 channel, especially when the microcomputer interface permits prestimulus trigger mode. In play-back mode this oscillating signal is detected by a filter, compared to an adjustable trigger level and converted into a TTL signal (Fig. 3b). Time variability of the trigger has been measured to be below 50 ps (trigger signal generated at the second oscillation of the burst signal instead of the first). Theoretically it is possible that a signal of channel B will erratically incite the trigger. Such a signal would have to have harmonics > 17 kHz with an amplitude of 90% of saturation lasting at least 0.1 ms. In our tests, no biological signal ever was confused for a trigger.
Appendix Modification of the model JVC XL)-Z507
This model is almost identical to the previous regarding its technical performances. The modification described above applies with little adjustment; it is even somewhat simpler since there is no digital AC coupling. The analog AC coupling can be removed by shorting 4 capacitors of the input circuit (ClOl, C201, C108, C208) and by replacing the 2 capacitors Cl53 and C253 of the ‘line out’ by 100 fi resistors. The implementation of fixed gain and trigger is identical to the one described for the Sony DAT.
Acknowledgements Discussion We have tested and successfully used our implementation in triggered as well as continuous recordings. Compared to the FM tape used before, we were pleased with the ease of use. We have also tested the frequency and transient responses as well as the DC stability (Fig. I>. We think that this modification offers a useful low-cost solution for many researchers in various fields who need to record analog data continuously. Although we have not implemented our modification an a portable DAT, this could be a very convenient solution for field work.
We wish to thank Sony Switzerland as well as Spitzer Electronics (JVC Switzerland) for their collaboration and for providing the DATs. We also thank Drs. H.-R. Liischer, E. Niggli and J. Streit for helpful comments on the manuscript. Supported by Swiss National Science Foundation (Grant No. 31-27553.89).
References Bezanilla, F. (198.5) A high capacity data recording device based on a digital audio processor and a video cassette recorder. Biophys. J., 47: 437-441. The Sony DTC-55ES. In: Service Manual. Sony Corporation Audio Group. JVC Service Manual. Digital Audio Tape Deck. XD-Z505. Victor Company of Japan. Maebashi-city, Japan.