TecellaAmp API Installation & Release Notes

Installation Notes
======================================

- Make sure to install the amplifier drivers first and that they show up properly in the Window's Device Manager.

- Include TecellaAmp.h in your project.

- If you're using Visual Studio, include TecellaAmp.lib or TecellaAmp_64.lib in your project.

- Link your application against TecellaAmp.dll or TecellaAmp_64.dll.

- For 32-bit applications, make sure to include the following files in your running directory:
	- TecellaAmp.dll
	- okFrontPanel.dll
	- libusb0.dll
	- The proper bit file for your amplifier.

- For 64-bit applications, make sure to include the following files in your running directory:
	- TecellaAmp_64.dll
	- okFrontPanel_64.dll
	- libusb0_x64.dll
	- The proper bit file for your amplifier.

- If you receive the error code 0xc0150002 when running your application, try installing the latest .NET runtime from Microsoft.

- If you receive the error "unable to enumerate amplifiers" or "unable to load okFrontPanel.dll", try installing .NET 3.5 via Microsoft Update or Windows Update.

Release Notes
=====================================
Version numbers are formatted as X.Z
Changes in X are reserved for major overhauls of the API.
Changes in Z indicate new features have been added that do not break compatibility with previous versions.

v1.28
=================
- Uncomment dir.length logic in hardware_opal_kelly.cpp.
- Skip v1.27, so that LSB of v1.28 matches corresponding v106.08.

v1.26
=================
- For Apollo/Mach, set the current clamp (CC) gains as follows:

    10x gain with  0pF CC capacitance compensation
    10x gain with 55pF CC capacitance compensation

- Add support for gain over-ride files apollo_0.cfg (for VC) and
  apollo_1.cfg (for CC) for Apollo/Mach.

v1.25
=================
- For Apollo/Mach, set default Cfast offset to 0.
  This undoes the setting of 15pF from v1.23.

v1.24
=================
- "cfast_offset" now supported for apollo_0.comp, shasta_0.comp,
  and pico2_0.comp.  Sample file format is:

    apollo_0.comp (first line is comment line)
    cfast_offset 25

- Force re-programming of the firmware when selecting "Start" tab in
  TecellaLab.

v1.23
=================
- For Apollo/Mach, set default Cfast offset to 15.  This means that
  setting Cfast to zero will actually set the Cfast hardware to 15pF.

- For Apollo/Mach, add "cfast_offset" parameter to "apollo_0.comp" file
  to over-ride the above default setting.

    - Create a file called "apollo_0.comp".
    - The first line is comment line.  Suggest typing "apollo_0.comp".
    - As an example, on the second line, enter "cfast_offset 30"
      to change default Cfast offset from 15 to 30.

- Create Shasta Duo configuration to support both voltage clamp and
  current clamp.

v1.22
=================
- For Apollo/Mach, force Iclamp OFF when switching configurations.
  This fixes side-effect of v1.20, V1.21 where switching from CC to VC
  with Iclamp ON caused ringing on VC High Gain.

v1.21
=================
- Add delay before and after loading firmware.

v1.20
=================
- Do not reset the FPGA when changing user configuration.
- Maintain source select when changing user configuration.
- Revert back to block size of 1024 for Apollo.

v1.19
=================
- For Mach Duo2, use Low and High gains in Current Clamp mode to
  select the level of capacitance compensation.  22pF for Low setting,
  and 62pF for High setting.  Voltage follower gain is 4x for both
  settings.

v1.18
=================
- Do not re-load the FPGA when changing user configuration.

v1.17
=================
- Support for Mach for 1024 segments, without the channel 145-192 bug.

v1.16
=================
- Auto Artifact removal now works when combining Delta segments with
  Ramp segments.

v1.14
=================
- Auto Artifact removal now works on downward ramps, too.

v1.12
=================
- Auto artifact removal now works on each step of ramp segments.

v1.11
=================
- For Apollo and Mach, current clamp range is increased to +/-5nA.

- Revert maximum number of stimulus segments to 64.

v1.10
=================
- For Apollo and Mach, maximum number of stimulus segments is increased
  from 64 to 1024.

- For Apollo and Mach, fix the bug where an extra segment was inserted
  between stimulus patterns when the stimulus pattern was repeated
  more than once.

v1.06
=================
- Max_stimulus_segment is increased to 256 for Apollo, Shasta, Jet, and
  Triton+, but firmware changes are necessary.

v1.05
=================
- Disables Digital Leak Compensation

- Max_stim_seg is 128 for Apollo and Triton, but requires firmware change

v1.04
=================
- Reverse polarity of iadj to match that of Vcmd

v1.03
=================
- Set iadj register to 0 when switching to Current Clamp mode

v1.02
=================
- Use model cell for offset calibration of all current clamp modes

v1.01
=================
- Fix issues with Mach Duo when switching modes from voltage clamp
  to current clamp, and vice versa

v1.00
=================
- Support Mach Duo (voltage clamp and current clamp) that uses the
  Bolt daughter board

- Some support for Jet Sonic

v0.98
=================
- Can define number of slots and channels per slot in apollo_0.comp

- Example (there should be no leading whitespace in actual file):
    apollo_0.comp   (1st line is comment line)
    slots 2         (Number of slots/banks is 2)
    chans 16        (Number of channels per slot/bank is 16)

v0.97
=================
- Fix Auto Artifact update

- Add current clamp mode register support for Mach/Sonic Bolt

v0.96
=================
- Remove error check for Invalid Ramp End Value

v0.95
=================
- Add vcmd.range file support for Pico (see v0.90 notes)

- Leak and Rseries interaction is only checked if dummy file 
  "check_leak_rseries" exists

- Build 64-bit version with br64.sh

- Create new DevID for Mach

v0.93
=================
- Add support for 384-channel Mach and Rio

v0.92
=================
- Remove 2 extra pulses at end of AutoComp

- Streamline debug logging

v0.91
=================
- Disable all Vcmd ownership logic

v0.90
=================
- Support Vcmd Range configuration file. Filename is vcmd.range
    1st line is comment line
    2nd line is minimum value in volts
    3rd line is maximum value in volts
    4th line is LSB value in volts
      - For 10-bit DAC, (max - min) / 1023 (or 1024)
      - For 12-bit DAC, (max - min) / 4095 (or 4096)
 
- Example format (no leading spaces in actual file):
    Comment
    min -1
    max 1
    lsb 4.884e-4
  Or
    Comment
    min -1.0
    max 1.0
    lsb 4.884e-4

v0.89
=================
- Support generic Triton with config file and comp override file.
  Rf of 10M/100M/500M/1G.  Uses "Salinas C" as device ID.

v0.88
=================
- Fix gain telegraph for Pico

v0.87
=================
- Change timing resolution to 10us for selected Triton models

v0.86
=================
- Added support for gain over-ride files for most models
    apollo_0.cfg
    jet_0.cfg
    shasta_0.cfg

- Format is the same as for Pico and TritonDuo over-ride files
  shown below.

- Added support for compensation value over-ride files for
  most files
    apollo_0.comp
    jet_0.comp
    pico_0.comp
    shasta_0.comp
    triton_0.comp
    triton_duo_0.comp

  Format is:

    triton_0.comp (comment line)
    cfast 200
    cslow0 100
    cslow1 100
    cslow2 100
    cslow3 100
    jp_min -254
    jp_max 256
    leak 2
    rseries 10000

v0.85
=================
- Added support for Triton gain over-ride file (triton_0.cfg)
- Format is the same as for Pico and TritonDuo over-ride files
  shown below.

v0.84
=================
- Boost settings are now linked to Rf Gain settings
- Boost is enabled by default on Pico2 for 3.3Gig and 10Gig settings
- Default settings can be over-ridden via the boost configuration file
- Format for Pico2 boost over-ride file (pico2_0.boost)
    - First line is comment line
    - Second line is number of Rf gains (default 5)
    - Subsequent lines (one per gain) are SPACE delimited with 3 fields
         Enable (0 or 1), Frequency (0 to 255), Gain (0 to 255)
         Example:
            pico2_0.boost (or any other comment)
            5
            0 128 128
            0 128 128
            0 128 128
            1 22 52
            1 70 178
- Format for Pico2 or Triton Duo gain over-ride file (pico2_0.cfg or triton_duo_0.cfg)
    - First line is comment line
    - Second line is number of Rf gains (5 for Pico, 4 for Triton Duo)
    - Subsequent lines (one per gain) are TAB delimited with 2 fields
         Display string, Rf value in exponent form
         Example:
            pico2_0.cfg (or any other comment)
            5
            10Meg	10e6
            100Meg	100e6
            1Gig	1e9
            3.3Gig	3.3e9
            10Gig	10e9

v0.83
=================
- Fix Pico AutoCal problem on Windows XP
- For Pico, offset the displayed Cfast value, so approximately 0pF is
  displayed after compensating source=head with nothing attached
- Speed up AutoCal

v0.82
=================
- Fix sources for Pico2

v0.81
=================
- Config file support added to Pico2
- Bug fixes for Pico2

v0.80
=================
- TritonDuo: add I=0 and Iclamp offset calibration
- TritonDuo: full Vclamp/Iclamp support completed

v0.79
=================
- TritonDuo: config file support
- TritonDuo: combine I=0 and Iclamp into single mode.
  Use IclampOn bit to switch between the two modes.

v0.78
=================
- Add Iclamp On/Off support for custom amplifiers

v0.77
=================
- Add dual-register source select support to enable disconnecting
  the head, and allow Imodel to be decoupled from Iclamp circuit.

v0.76
=================
- Add support for multiple Iclamp gain values for Tritons that
  support Iclamp.  Available gain values are 10x, 30x, 120x,
  and 250x.

v0.75
=================
- For Picos and Tritons that support Iclamp (current clamp)
  mode(s).  Allow connecting/disconnecting of Iclamp circuit
  in Iclamp modes to achieve true I=0 in any of the Iclamp
  modes.  In the future, I=0 mode will be phased out, since
  any Iclamp mode can be used in I=0 mode.

v0.74
=================
- Add support for both Vclamp and Iclamp in new Tritons
- More reliable enabling of 100% Rs comp circuit in Picos

v0.73
=================
- Fix gains for I=0 and Iclamp modes
- Add support for custom TEVC Pico

v0.72
=================
- Add support for higher bandwidth Triton+

v0.71
=================
- Fix Pico's JP range to +/-255mV

v0.70
=================
- Remove delay from writes to OpalKelly

v0.119.50
=================
- Adds streaming stimulus and segmented stimulus support for Pico.
- Unicode characters removed.
- Rare race condition fixed to prevent hang on Start/Stop.

v0.119.49
=================
- Adds support for additional amplifiers.
- Include Pico firmware by default.

v0.119.46 - 48
=================
- Internal Releases.

v0.119.45
=================
- Fixes the bessel settings for Amadeus.
- Adds clipping detection per-channel.
- Improvements made to auto calibration.
- Documentation updated and reorganized.
- Initial support added for custom amplifier config files.
- Adds support for more amplifier configurations.

v0.119.44
=================
- Requires Pico firmware to be upgraded to v007.
- Utility DACs fixed for Pico.
- Fixes Pico initialization if the system has not had an Opal Kelly based amplifier installed previously.
- Several bug fixes for issues that caused Pico to hang.
- Changed software gain values to match Pico's hardware.

v0.119.43
=================
- Internal release.

v0.119.42
=================
- Adds Beta 64-bit support.
- Removes dependencies on mingwm10.dll and libgcc_s_dw2-1.dll.
- Make acquisition start/stop faster for Pico.
- Fixes ramp stimulus segments for Pico.
- Fixes bug where the tecella_initialize() would always fail if an OpalKelly device has never been connected to the computer before.


v0.119.41
=================
- Fixes some of the settings for Pico.
- Fixes some corner cases where the stimulus would not program correctly.


v0.119.40
=================
- Remaps the channel steering output pins for Amadeus.


v0.119.39
=================
- Add tecella_chan_set_iclamp_enable and tecella_chan_get_iclamp_enable.
- hwprops_ex_01 structure changed.
- Structure of hwprops remains the same, but some values have been renamed.
- Converted icmd values in hwprops to reserved.
- Vcmd values in hwprops converted to stimulus values.
- Stimulus values should be interpreted as icmd or vcmd values depending on the current mode.


v0.119.38
=================
- Fix a spurious DEVICE_CONTENTION error that would occur during continuous acquisition if a stimulus is not used.


v0.119.37
=================
- Fixed an Amadeus instability introduced in the previous release.


v0.119.36
=================
- Changes some settings to improve the noise performance of Amadeus.


v0.119.35
=================
- New Amadeus firmware disables register writes during acquisition.
- Remove un-needed workarounds for Amadeus.


v0.119.34
=================
- Enumeration for Pico handled properly now.
- Support for Pico serial numbers added.
- Added Iclamp support for TritonPlus and Pico.


v0.119.33
=================
- Attempts to fix gain switching on Amadeus.
- Fix some more of the register writes for Amadeus.


v0.119.32
=================
- Fixes a bug introduced in 0.119.31 that breaks tecella_initialize_from_file for Amadeus.


v0.119.31
=================
- Fix some of the register writes for Amadeus.


v0.119.30
=================
- In tecella_auto_comp: Allow undercompensation coefficient to affect digital leak.
- Remove the restriction that the first segment cannot be a ramp segment.
- Using a ramp as the first segment reduces the max segments by one.


v0.119.29
=================
- Add functions tecella_initialize_from_file() and tecella_finalize_to_file().
- Supports saving calibration and register state to a file.
- Enables quick initialization.
- Can now quickly move ownership of the amplifier to different processes.


v0.119.28
=================
- Make auto calibration depend on the HPF in addition to gain.
- Add delays back into auto calibration if a HPF is present.


v0.119.27
=================
- Removes glitches at the end of a stimulus for Amadeus.
- Fixes stimulus repeat for Amadeus.
- Fixes hpf and gains for Amadeus Rev A.
- Gets rid of the "settling" in first acquisition after initialization for Amadeus.


v0.119.26
=================
- Fix stimulus for Richmond and Jet.


v0.119.25
=================
- Improves auto offset and auto scale for Amadeus.

v0.119.24
=================
- Enabled auto offset and auto scale for Amadeus.
- Fixes missing last stimulus segment for Amadeus.


v0.119.23
=================
- Fixes stimulus offset adjust mode when leak isn't enabled.
- Improves offset adjust mode, by automatically scaling the offset when leak/jp change.


v0.119.22
=================
- Adds a new offset adjust mode that adjusts the stimulus(jp) and response simultaneously during tecella_auto_comp() when using analog leak.


v0.119.21
=================
- Fixed stimulus start synchronization issues on Jet and Richmond.
- Improve performance of acquisition and filtering.
- Automatically enables digital leak if used in tecella_auto_comp.
- Fixed crash with asynchronous vcmd when tecella_debug is enabled.


v0.119.20
=================
- Triton+, Shasta, and Apollo now suppurt extended stimulus functionaliy.
- Segment duration lsb for Shasta/Apollo changed.
- Changed the bessel value<->frequency conversion to better match the hardware.


v0.119.19
=================
- Changed Amadeus to only have a subset of gains so the gains can be set normally using tecella_chan_set_gain.


v0.119.18
=================
- Changes to TECELLA_HW_PROPS_EX_01 struct.
- Prevent gain divisors from being 0 on Amadeus
- Improved formatting of the register labels.
- Fixes Jet turbo mode.
- Added tecella_get_slew_label.


v0.119.17
=================
- Fixes a floating point exception that would occur during initialization.
- Extended stimulus interface added to control slew rate and synchronized digital io.
- Initial support for Amadeus added.
- Added functions to control a high pass filter.
- Added functions to steer stimuli to different destinations if supported by the amplifier.
- Added a TECELLA_HW_PROPS_EX_01 structure to maintain backwards compatibility for TECELLA_HW_PROPS.  TECELLA_HW_PROPS_EX_01 is subject to change, so be careful.


v0.119.15 - v0.119.16
=================
- Improvements made to autoScale.


v0.119.14
=================
- Update certain amplifier configurations.


v0.119.13
=================
- Improved auto compensation a bit. 
- Added voltage monitor support for Apollo and Shasta.
- Fixed auto leak check that makes sure rleak > rseries.
- Added support for another TritonPlus configuration.


v0.119.12
=================
- Improve auto analog leak compensation using an adaptive method.
- Fix autocomp when both analog and digital leak are used.


v0.119.11
=================
- Output tlc files for auto compensations in debug mode, to make debug easier.


v0.119.10
=================
- Fix scale for certain configurations


v0.119.9
=================
- Fixes TECELLA_INTERNAL_ERROR that would occur if the user acquires with a sample_period_multiplier other than 1 and then runs one of the tecella_auto* functions.
- Compiled with auto vectorization enabled to improve performance of filtering functionality.
- Added offset adjust modes to adjust the stimulus offset using JP or the response offset after tecella_auto_comp() with use_leak=true.
- Added an auto-zero function, which adjusts the response offset at the beginning of every acquisition.
- Added extended stimulus functionality for hardware that supports it.
		- Subject to change, but includes slew control and digital vcmds synchronized to the analog vcmd.
- New Triton+ configurations added.
-


v0.119.8
=================
- All auto compensation / calibration functions would silently produce incorrect results if the holding voltage wasn't set to the same initial voltage used by the auto compensation / calibration function.  The voltage difference would cause the auto function to see a spike at the beginning of acquisition, throwing off calculations.  This release properly manages the holding voltage so the spike no longer appears.  The holding voltage is restored before returning.


v0.119.7
=================
- Fixes autoscale bugs.
- Add new / fix some old amplifier configurations.


v0.119.6
=================
- Fix a suprious DEVICE_CONTENTION error that would occur after setting a stimulus without any channels assigned to it.
- Fix a segfault that might occur during autocomp in a very rare corner case.


v0.119.4-5
=================
- Fix some spurious TECELLA_INTERNAL_ERRORS.


v0.119.3
=================
- Fix the remaining spurious device contention errors (especially regarding multiple stimuli.)


v0.119.2
=================
- For vcmd ranges of +-2V on TritonPlus, use a step size of 1mV instead of .98mV.


v0.119.1
=================
- Re-enabled utility trigger out support on TritonPlus.


v0.119.0
=================
- First crack at merging the v0.106.x branch into the main release.
- Fix some spurious device contention errors, however others remain.
- Adds support for a new amplifier configuration.
- Fixes spurious Opal Kelly library not found error.


v0.118.0
=================
General new features:
- Single official Opal Kelly driver used now that works on all Windows platforms (XP/Vista/7+32/64 bit).
	*The user must uninstall the old drivers manually, then install the new drivers.
	*You no longer need okFrontPanel32.dll and okFrontPanel64.dll - just okFrontPanel.dll.
- New dll requiriment, which supports throwing exceptions accross dll boundaries: libgcc_s_dw2-1.dll.
- Added digital leak, which replaces leak calibration.
- Added system monitor functions.
- Autocomp can do digital leak and artifact removal now.
- Added filtering API - useful if you wish to run a low pass filter before using a lower sampling rate.

Changed/new structs:
- TECELLA_HW_PROPS: nslots field added.
- TECELLA_SYSTEM_MONITOR_RESULT: new
- TECELLA_ERRNUM: added TECELLA_ERR_DEVICE_CONTENTION error.

Changed function signatures / functionality:
- tecella_auto_comp: Performs digital leak and capacitance compensation. Removed leak calibration.
- tecella_stimulus_zap: zero length duration implies a "holding Zap".

Removed functions:
- tecella_auto_calibrate_leak_enable: use digital leak subtraction instead.

New functions:
- tecella_system_monitor_run: use this to retrieve system voltages, power usage, and such.
- tecella_system_monitor_get: "
- tecella_chan_set_digital_leak: applies digital leak
- tecella_chan_get_digital_leak: "
- tecella_chan_set_digital_leak_enable: "
- tecella_chan_get_digital_leak_enable: "
- tecella_auto_artifact_decay_start: allows artifact removal to trail off so there isn't a big jump.
- tecella_auto_artifact_get_size: allows user to get/set the artifact being removed.
- tecella_auto_artifact_get: "
- tecella_auto_artifact_set: "
- tecella_sw_filter_enable: applies a digital filter to the incomming data.
- tecella_sw_filter_set: allows a user-specified filter kernel.
- tecella_sw_filter_set_box: a convenience function for setting a simple box filter.
- tecella_sw_filter_set_lanczos: a convenience function for setting a lanczos filter.
- tecella_acquire_read_f: same as tecella_acquire_read_d, but for floats.

Extra notes about this release:
1) We upgraded to the official Opal Kelly Driver that works for all Windows platforms.
Why: Previous versions relied on a beta driver for 64-bit versions.
How this affects you:
a) You have to remove the old driver and install the new driver.
b) To uninstall the old driver, you may have to manually go to the device manager and select "delete the driver software for this device" when uninstalling the device.
c) Your application only needs okFrontPanel.dll instead of okFrontPanel32.dll and okFrontPanel64.dll.

2) Added digital leak compensation, which replaces leak calibration.
Why: Leak calibration (zeroing the offset after auto analog leak) was a hack and digital leak is the more general solution. Digital leak compensation allows for a more continuous leak compensation when used in conjunction with analog leak compensation. Analog leak compensation is still needed, however, to improve the dynamic range of the amplifier.
How this affects you:
a) When running auto leak, you should always use analog leak in conjunction with digital leak.
b) After running auto leak, use tecella_chan_get_digital_leak to get the amount of leak being compensated digitally.

3) Added system monitor functions.
Why: These functions allow you to monitor the voltages of the amplifier which will help diagnose a problematic system.
How this affects you:
a) When you want to add diagnostic functionality, you can implement the calls to tecella_system_monitor_run/get.

4) Autocomp can do digital leak and artifact removal now.
Why: For digital leak, see (3). We added artifact removal to tecella_auto_comp because it is basically free to perform simultaneously with any other compensation.
Pitfalls: Compared to a stand-alone artifact removal via tecella_auto_artifact_update, it may not be as accurate if you simultaneously auto comp cfast/clsow. Results are still pretty good, though.
How this affects you:
a) The call signature of tecella_auto_comp has changed.
b) You should consider setting use_digital_capacitance to true whenever you can (for example during every auto leak), since it's free.

5) Added filtering API
Why: Allows you to specify a low pass filter, which is useful for reducing noise when using a lower sampling rate.
How this affects you:
a) If you use a sample period multiplier of N, you should call tecella_sw_filter_set_box() with width = N before starting acquisition.
b) Alternatively, you may specify your own filter.

6) TECELLA_ERR_DEVICE_CONTENTION error if an acquisition, autocomp, or vcmd change overlaps.
Why: Device contention bugs are difficult to debug at the application level, so the API tries to fail early if it detects device contention.
How this affects you:
a) The following functions may now return TECELLA_ERR_DEVICE_CONTENTION: tecella_acquire_start, tecella_acquire_start_stimulus, tecella_stimulus_set, tecella_chan_set_stimulus, tecella_auto_comp, tecella_auto_artifact_update, tecella_auto_offset, tecella_auto_calibrate, tecella_auto_scale.

7) New dll requiriment: libgcc_s_dw2-1.dll.
Why: Dynamically linking this library supports throwing exceptions across DLL boundaries. Static linking of this library does not.
How this affects you:
a) You must include libgcc_s_dw2-1.dll with your application.
b) You may now throw exceptions from your callback functions and they should properly unwind the stack even as they go through the TecellaAmp.dll.



v0.117.1
=================
- Keep track of individual stimuli for TECELLA_ERR_DEVICE_CONTENTION.

v0.117.0
=================
- **INTERFACE CHANGE**
- Associates tecella_auto_scale and tecella_auto_calibrate with a stimulus.
- More error checking.

v0.116.2
=================
- Refactor flow control / device ownership.
- Improve error reporting and add a new error: TECELLA_ERR_DEVICE_CONTENTION.
- Removes 1s delay when programing stimulus for some models of Apollo.

v0.116.1
=================
- Auto compensation with leak is clamped to within valid values.
- Fixes a bug where pairing up amplifiers would crash when running ZAP.

v0.116.0
=================
- tecella_auto_artifact_update() takes in t_hold and t_step arguments and uses t_step to determine for how long artifact removal should work.
- Added tecella_auto_artifact_decay_start(), that allows a user to specify when artifact removal should start decaying to prevent a sudden step if the artifact removal length is not long enough.
- Fixed a few bugs with amplifier pairing.

v0.115.0
=================
- Added preliminary support for pairing amplifiers together to make them look like a single amplifier.
- Added support for amplifiers that can be reconfigured on the fly.  tecella_user_config_get() tecella_user_config_set().
- Removed support for changing between voltage/current clamp, since no amplifiers support it.
- Fix bug where data could be lost as it's read from the amplifier.
- Fix small probability of hang when stopping acquisition.
- TECELLA_HW_PROPS added two new feilds: user_config_count, user_config_name.
- Fix a few bugs where enabling/disabling channels wouldn't work right.
- Fix "r6025 pure virtual function call" bug that shows up intermittently under heavy load with multiple threads.

v0.114.0
=================
- tecella_auto_comp takes a new parameter, under_comp_coefficient, to give the user control over how much capacitance should be over/under compensated.
- Added a new function, tecella_auto_artifact_get_capacitance, that returns the equivalent capacitance of the artifact removal.

v0.113.1
=================
- Fix bug in autocomp algorithm.

v0.113.0
=================
- Accidentally changed serial_number field of hw_props in v0.112.0 to char from wchar_t.  Changed it back to wchar_t.
- Fixes Zap for old Tritons.

v0.112.0
=================
- Adds an enumeration interface that allows the user to work with multiple amplifiers simultaneously.
- The second argument to tecella_initialize() is now a device index instead of a model type.  Using the TECELLA_HW_MODEL_AUTO_DETECT value as the index should still have the same effect.

v0.111.11
=================
- tecella_auto_comp with leak works with rseries enabled again. (Fixed Rleak>Rseries restrictions used during autocomp.)

v0.111.10
=================
- Artifact removal is now only disabled during tecella_auto_comp() if a cfast/cslow is used.
- Fixes bug where auto leak calibration would be unstable (infinity).
- The auto calibrate leak no longer keeps the baseline at 0 for all values of leak:
	1)The baseline will be 0 for the leak value calculated by tecella_auto_comp.
	2)The baseline will be the orignal value if leak is set back to zero.
	3)All other values of leak will linearly interpolate the baseline between 1 and 2.

v0.111.9
=================
- Added fuctions tecella_auto_calibrate_get/set()
- tecella_auto_comp now auto calibrates for leak whenever leak changes for a given jp and vcmd.
- Added function tecella_auto_calibrate_leak_en to enable/disable previous functionality.
- Fixes bug in previous version that would result in incorrect leak calibration.

v0.111.8
=================
- A new restriction is enforced that requires Rleak>Rseries.
	This restriction is taken into account automatically during tecella_auto_comp.
	If the user breaks this restriction, an error will be returned.
- Fpga would not be reset if firware was deteceted during initialization.  Fpga is now reset just in case.

v0.111.7
=================
- Add new TritonPlus configuration.

v0.111.6
=================
- Programming the firmware would produce spurious stimulus pulses, so the firmware is only programmed on the first initialization after power-up instead of on every initialization.
- The source now defaults to NONE instead of HEAD after initialization, so that the user may specify a holding voltage instead before HEAD is selected.
- Removed the serial number dependency for some TritonPlus configs.

v0.111.5
=================
- Special case a single TritonPlus amplifier with a different leak range.

v0.111.4
=================
- Rseries is no longer disabled automatically during tecella_auto_comp().

v0.111.3
=================
- ZAP was broken for several releases.  Fixed now.

v0.111.2
=================
- Added new TritonPlus configuration.

v0.111.1
=================
- Improvement to auto leak for higher gains.

v0.111.0
=================
- Makes pulse length in tecella_auto_comp programmable.

v0.110.1
=================
- Adds auto offset after auto comp.

v0.110.0
=================
- Support for Shasta added.

v0.109.2
=================
- Fix the range checking on ramp segment types.

v0.109.1
=================
- Update to a few TritonPlus configurations.

v0.109.0
=================
- Add a tecella_auto_calibrate_get to retrieve the per-channel calibrations.

v0.107.1
=================
- When tecella_finalize is called, the vcmds are now put into a state that shouldn't oscillate during the next call to tecella_initialize.
- Add 1s delays after changing JP during tecella_auto_offset to allow the response to settle.


v0.107.0
=================
- !!Binary interface change!! - Add "iterations" argument to tecella_auto_comp().
- Re-enable rseries properly after tecella_auto_comp().
- Disable leak during tecella_auto_offset().
- Fix rounding error on register bounds checking.
- Change leak units to be Ohm^-1 to make the scale linear.  Previous units were incorrect.


v0.106.3
=================
- Added a tecella_acquire_start_stimuli function that can start multiple stimuli simultaneously.


v0.106.2
=================
- Added 5 second delays to tecella_auto_offset after any change to JP to let the response settle.


v0.106.1
=================
- Removed vcmd check for 3 or more segments and starting/ending with a SET type segment.
- Ramps segments are not always reported as out of bounds now.


v0.106.0
=================
- Sometimes the number of samples read would be less than expected.  Should be fixed now.
- Fixed bug where graph would appear "rounded" on certain TritonPlus models.
- !!Binary interface change!!  Will crash if tecella_hw_props is not updated with 5 new fields marked with a "+" below.

 typedef struct tecella_hw_props {
 	TECELLA_HW_MODEL hw_model;
 	wchar_t     device_name[32];
 	wchar_t     serial_number[16];
 	int         hwvers;
 	int         nchans;
 	int         nsources;
 	int         ngains;
 	int         ncslows;
 	int         n_utility_dacs;	
 	int         nstimuli;
 	int         max_stimulus_segments;
 	bool        supports_async_stimulus;
 	bool        supports_oscope;
 	bool        supports_vcmd;
 	double      stimulus_value_min;
 	double      stimulus_value_max;
 	double      stimulus_value_lsb;
+	double      stimulus_ramp_step_size;
 	bool        supports_icmd;
 	double      icmd_value_min;
 	double      icmd_value_max;
 	double      icmd_value_lsb;
+	double      icmd_ramp_step_size;
 	double      stimulus_segment_duration_max;
 	double      stimulus_segment_duration_lsb;
+	int         stimulus_delta_count_max;
+	int         stimulus_repeat_count_max;
+	int         stimulus_ramp_steps_max;
 	bool        supports_zap;
 	double      zap_value_min;
 	double      zap_value_max;
 	double      zap_value_lsb;
 	bool        supports_bessel;
 	int         bessel_value_min;
 	int         bessel_value_max;
 	double      utility_dac_min;
 	double      utility_dac_max;
 	double      utility_dac_lsb;
 	double      sample_period_min;
 	double      sample_period_max;
 	double      sample_period_lsb;
 	int         bits_per_sample;
 } TECELLA_HW_PROPS;


v0.105.1
=================
- Added tecella_debug() that logs API function calls + arguments to a file.
- More robust bounds checking / error detecting / vcmd validation.
- Change bit file search back to the old way:
	- Looks in executable directory instead of working directory.
- Fix sample notifications that broke in v0.105.0.
- Make hardware initialization a little faster.
- When setting registers values, the values are now rounded to the nearest valid value instead of truncated a lower value.


v0.105.0
=================
- !!WARNING!!: Your program will crash if you call tecella_auto_offset() using the old argument list.
- Binary interface change to accomodate the following:
- tecella_auto_offset() now performs automatic JP calculation.
	ie: tecella_auto_offset() compensates for EXTERNAL offsets by offseting the stimulus.
- tecella_auto_calibrate() performs the same function as the old tecella_auto_offset()
	ie: tecella_auto_calibrate() compensates for INTERNAL offsets by offseting the measured current.

v0.104.11
=================
- Stimuli with more than ~2^32 samples would acquire incorrect amount of data or crash.  Limit extended to ~2^64 samples.
- Add tecella_utility_set_stimulus() to the example program.
- Fixes inversion of Richmond/Jet holding voltage.
- Fixes a memory leak/accumulation that occurs if a notification callback is not set.
- tecella_auto_comp() does not change leak anymore if asked not to.
- Changed bessel initialization for TritonPlus.
- Change rseries ranges from 10MOhm to 20Mohm for some hardware configurations.
- Fixed bug on TritonPlus where graph looks filtered after enabling rseries.
- Acquisition stops automatically if no data is present on WallE.
- Beta support for 64-bit Vista added.
	- okFrontPanel32.dll and okFrontPanel64.dll must be used instead of okFrontPanel.dll.
- Fixed regression that occured since v0.104.7, where base acquisition sample period was not actually taken into consideration.

v0.104.7 - v0.104.10
=================
- Support for WallE hardware added.
- Fixed unsigned / signed samples for WallE.

v0.104.6
=================
- Fix range for cfast on Jet and Richmond systems.
- More tweeks to auto artifact.

v0.104.5
=================
- Contains fixes from both v0.104.3 and v0.104.4.
- Triton Plus holding voltage bug fixed.

v0.104.4
=================
- Direct branch from v0.104.2.  (Does not contain anything from v0.104.3)
- Fixes stimulus min/max for Richmond.

v0.104.3
=================
- Autocomp will now work even if auto offset hasn't been run or if there is a junction potential present.

v0.104.6
=================
- Fix range for cfast on Jet and Richmond systems.
- More tweeks to auto artifact.

v0.104.4
=================
-Direct branch from v0.104.2.  (Does not contain anything from v0.104.3)
-Fixes stimulus min/max for Richmond.

v0.104.3
=================
-Autocomp will now work even if auto offset hasn't been run or if there is a junction potential present.

v0.104.2
=================
- Fixes bug in last two releases where no samples would be acquired if acquisition were started with start_stimuli=false and continuous_stimuli=true.
- Fixes crash that occurs after 10-15 minutes.
- Add ramp+delta support to TritonPlus.


v0.104.1
=================
Note: This release has only been tested with Richmond Hardware.
Known bugs: Stimulus notifications at the very beginning of acquisition may be lost.
-Fixed autocomp bugs and make it more consistent.
-Fixed tecella_utility_set_stimulus() function.
-Holding voltage now output whenever a stimulus is not playing.
	-This allows deltas/repeats to occur without having to visit the holding voltage each iteration.
	-Note: The user must now set the holding voltage explicitly using tecella_stimulus_set_hold(), instead of relying on the first/last segment of the vcmd.
	-Note: Minimum segment count is still 3 due to hardware/firmware limitations.
-Stimulus notifications from internal API acquisitions are no longer reported.
-First call to tecella_stimulus_start() failed in previous version.  Now fixed.
-Fixed stimulus notifications for Triton and Triton Plus.
-Enhancements to Artifact removal.


v0.104.0
=================
Support for stimulus only channels added via tecella_utlility_set_stimulus(_).

Timestamp functionality implemented and is now a 64 bit unsigned integer (unsigned long long).

tecella_chan_set*() and tecella_stimulus_set() functions queue their commands and return immediately instead of blocking until the request is complete.

Added stimulus notification callback (TECELLA_STIMULUS_CB), which is guaranteed to be called before notification of samples in acquire callback.

Gain labels (from tecella_get_gain_label()) now reflect actual feedback resistor values of the connected hardware.

Auto offset and auto scale are not run during initialization any more and should be called manually after tecella_initialize().

Fixed bug where maximum reported stimulus segment was incorrect.

Fixed bug where tecella_auto* functions would not work if user acquired with a sample period other than 50us.

Fixed bug where sample counts for ramp/delta/repeat segments were incorrect.

Fixed bug where sample count returned by tecella_stimulus_sample_count did not match actual number of samples returned.


v0.103.0
=================
Ramp and delta segments fully supported.
Autocomp on a vcmd while acquiring on others works now.
Fixed various bugs that were notifying the API user of internal data.
TecellaTLC split off into a separate dll to fix 2GB limit.
	-Compiled by gcc instead of msvc, so please notify us of any incompatibilities.
Fix invalid memory access if acquiring when a vcmd has not been set.
Interface for timestamping data added (but not implemented).
Interface for signaling last sample added (but not implemented).
Numerous memory leaks patched.

v0.102.3
=================
When programming a vcmd, tecella_stimulus_set takes in two integers (step_count and repeat_count) instead of an int and a bool (iterations and loop).
	Be careful, as old code will still compile but will likely be invalid!
Added initial support for delta and ramp stimulus types.
	Known bugs, to be fixed:
	A stimulus with a delta segment is limited to 3 segments.
	A stimulus cannot have both a delta and a ramp segment.
Added initial support for artifact removal, which digitally compensates for any imperfections in the analog compensation (calculated by tecella_autocomp or otherwise.)
	Known bugs, to be fixed: Doesn't work all the time with async stimuli.
	Known limitations: Does not work for delta and ramp segments.
"Invalid argument" error code added.

v0.102.2
=================
Automatic gain calibration done during initialization.
Function added to specify the stimulus value on the fly.  (See tecella_stimulus_set_hold.)
Fixed sample period specification - it can now be any multiple of 50e-6 seconds.
Added utility dac interface.  (See tecella_utility_dac_set.)
Changed error codes.


v0.102.0
=================
Better async vcmd support:
	Individual stimuli can be stopped.
	More options to only acquire while a vcmd is playing.
	Frame size reduced from 1 second to ~10us for reduced latency.
Auto Compensation
	Works for Richmond and TritonPlus now.
	More options added to specify test pulse amplitude and which compensations to use.
	Note: support for autcompensation while acquiring is not supported yet.
Revamped error reporting:
	Initialization failures returned.
	SW buffer overflow.
Rounding errors fixed for tecella_get_stimulus().
Simple holding voltage can be specified with tecella_stimulus_set_hold() instead of having to programming a vcmd.
tecella_stimulus_zap() works now.
Polarity of vcmd/acquisition fixed for several products.
Capture bit removed from stimuls segment.

v0.101.5
=================
Support added for Jet.
Autodetect support added for initialization.

v0.101.4
=================
Bessel filter channel selection works properly now.

v0.101.3
=================
Preliminary multi/async vcmd support for hardware that supports it.

v0.101.2
=================
This is our first full release for which we expect the API to remain fairly consistent.

A new acquisition scheme has been implemented that doesn't block the calling thread.

Known issues, that will be fixed in the future:
	-Changing the channel settings can be a little slow.
	-Other than tecella_initialize(), error reporting isn't very robust.
	-Zap is not yet implemented.
	-Auto-detect of the model of the connected device doesn't work.
	
