HEKA Elektronik EPC 10 USB, Hardware Manual, Page: 51 / 81

7.2 Capacitance Compensation 45
overshoot for time constants τ
u
greater than about 1 ms; the 100 µs setting is appropriate for τ
u
values on the
order of 10 ms or longer. In practice, you can estimate τ
u
from the ratio of the settings of C-slow and R-series .
For example, if C-slow is 10 pF and Rs is 10 M Ω, the time constant is 10 pF
∗
10 M Ω = 100 µs . The use of the
Rs-compensation circuitry can be summarized as follows: When you set the capacitance transient cancellation ( C-
slow , R-series , C-fast , τ -fast ) to minimize the size of the transients when voltage pulses are applied, you have also
properly set them for series resistance compensation. Then you enable R-series and turn up the %-comp control
to the desired value. Any maladjustment of the transient cancellation will be apparent and can be compensated.
The
EPC 10 USB
makes the procedure very easy: C-slow and R-series values can be obtained automatically
by clicking on Auto C-slow . Beyond that, the Pulse Generator of
Patchmaster
provides an option where Auto
C-slow is performed before each command pulse, achieving an accurate update of R-series . The procedure will be
described in more detail later (see chapter 10.3.3 on page 57).
7.2 Capacitance Compensation
The
EPC 10 USB
provides automatic procedures for both fast and slow capacitance subtraction. In both cases,
the ongoing pulse protocols are suspended and short trains of square-wave pulses are applied (number and am-
plitude of these pulses are specified by
Auto C-slow Settings
in the
EPC 10
menu). The resulting capacitive
transients are averaged, leak-subtracted, and then used to calculate the required corrections to the components of
the compensation network. A detailed description of the procedures to estimate capacitance is found in Sigworth
et. al (1995). The procedure for
C-fast compensation (pipette and stray capacitance) is concerned mainly with
the very initial portion of the transient (the first 10
−
30 µs ), whereas the procedure for C-slow compensation
(whole-cell capacitance) regards a somewhat later time window. The width of this window is based on an initial
guess of what the time constant of the slow transient might be. Therefore, it is a good idea to set typical values
for
C-fast, τ -fast, C-slow and R-series in advance (they may be saved in the protocol for whole-cell recording, for
example). This will provide the fitting routine with reasonable starting values. Alternatively, the Auto procedures
can be performed alternately to arrive at settings for which the remaining transient is minimal. Settings for C-fast
are left at their respective values whenever an Auto C-slow is performed and vice versa. This way, a combined fit
can be obtained by alternately executing C-slow and Auto C-fast . It is better, however, to perform Auto C-fast in
the cell-attached recording configuration and not to touch it any more when proceeding to whole cell.
During Auto C-fast and Auto C-slow , certain settings are changed temporarily and restored upon completion.
These are:
ˆ
Rs-compensation is switched ”Off”
ˆ
The Gain is set to 0.2, 10 or 50 mV/pA, depending on the Gain range in use
ˆ
External stimuli are disabled
ˆ
Filter 1 is set to 10 or 30 kHz, depending on the required bandwidth
These changes are not displayed on the screen, since they usually are only effective for fractions of a second. In
Cap. Track , however, the changed values are effective for longer times, since the Auto C-slow procedure is called
periodically. It should be kept in mind that under this condition the parameters on the screen may not represent
the hardware settings.
The relationship between the values of the compensation network ( C-slow and R-series ) and those of the pipette-
cell assembly (Cm and Rs) is straightforward, if the membrane conductance is negligible. In this case, perfect
compensation will leave no residual current and
C-slow will be equal to Cm. If, however, there is a finite membrane
conductance, then some ambiguity exists, because a three-component network is being approximated by a two-
component compensation network. Details of the compensation procedure will then determine the residual current
(some filtered version of the command waveform) and Cm is likely to be underestimated.
The Auto C-slow
procedure is designed to provide unbiased estimates of the actual membrane capacitance and series resistance
(Sigworth, Neher & Affolter, 1995).
http://www.heka.com