"Zapata" project status

As of Saturday, November 20, 2004

The Quad PCI Radio Interface card (Rev. A) and its driver pciradio.c and VHDL files (pciradio.vhd, ucf, and rbt) released today, along with app_rpt version 0.19. See the app_rpt page for details.

[Sorry, we didn't update this page much for about 3 years.]

As of Tuesday, December 11, 2001

E-1/A-law support completed and verified for the Tormenta 2 PCI and Tormenta ISA cards.

As of Monday, October 22, 2001

Tormenta 2 PCI card (Rev. A) design tested and verified (along with associated driver) after much pain and torment.

[Sorry, there were a few things left out of this status report in the middle here, but it was getting rather frustrating and depressing]

As of Wednesday, August 1, 2001

Tormenta 2 PCI card board layout and CAD completed today. Will get boards back from the fabrication house by mid-August for testing.

As of Friday, July 13, 2001

Version 1.10 of the LINUX driver/library released today. Contains bug fixes.

As of Thursday, July 12, 2001

Tormenta 2 PCI card design engineering completed today. Will meet with circuit board CAD dude sometime during the week of July 23 to generate artwork for the cards, and may have them back and tested by the first or second week in August (if we are lucky and didn't screw things up too much).

As of Monday, June 25, 2001

Version 1.9 of the LINUX driver/library released today. The DTMF receiver is now fully functional and performs quite excellently under all tested conditions (hooray for Steve!!). Its sensitivity, accuracy, speed, and talk-off immunity (we use the Bellcore tapes for comparison, and get no more then 19 hits where, according to their spec, 470 hits is acceptible) compares with or beats any of the commercially-available products of its type (hardware and/or software based). Also includes a couple of very minor bug fixes and general improvements in the driver and library.

As of Monday, March 13, 2001

Version 1.7 of the LINUX driver/library released today. Major improvements have been made to the quality of the DTMF receiver. Its now to the point where it performs more then acceptably under virtually all conditions.

As of Thursday, March 8, 2001

Version 1.6 of the LINUX driver/library released today. This release is fully tested and functional (with some limitations of the quality of DTMF reception, but its at least better then it was). There have been some quite significant additions to the archetecture/functionality, including user-scalable buffer pool management, standard tones, such as ringing, busy, DTMF, etc., generated within the kernel, clear-channel (data) features, including multiple-channel interfaces (multiples of 64K), HDLC transmission and reception, and integration into Linux networking (it creates Linux network devices, such as hdlc0, if desired), non-blocking I/O, improved driver configuration script/interface, improved efficiency in DTMF decoding (and improved quality soon), and ability to use IDE UDMA hard drive, instead of expensive SCSI. Check it out.

As of Monday, January 1, 2001

Version 1.5 of both the BSD driver/library (click here to download) and the (still experimental at this point) LINUX driver/library released and is available on-line. This driver/library includes much improved tone (both DTMF and MF) detection. In both modes, it is able to detect tones in the receive level range of -20.0 dbm to +1.0 dbm, and the DTMF has quite improved noise and talk-off immunity.

As of Saturday, December 30, 2000

Version 1.4 of the (still experimental at this point) LINUX driver released and is available on-line.

As of Saturday, December 23, 2000

The BSD driver and library, version 1.4 has been released, and is available on-line. This includes a couple of minor bug fixes in the library, and adds support for FXO signalling, and Tone-muting while on a conference.

As of Tuesday, December 12, 2000

The software has been (experimentally. at this point) ported to Linux. Version 1.3 of the LINUX version of the Tormenta driver, the Zapata library, and the Emiliano MIPL browser has been released, and available on-line. Technically, this is an experimental release, and we will see if there are any problems/issues with using this technology with Linux that are so far unforeseen before considering it generally viable for use.

This software includes added support for channels using both loopstart and groundstart FXO signaling (to operate with an FXS channel, such as one connected directly to an analog telephone set).

The BSD version of this version (1.3) of the driver/library will be available in the next couple of days.

As of Wednesday, November 29, 2000

Version 1.2 of the Tormenta driver, the Zapata library, and the Emiliano MIPL browser released, and available on-line.

The Emiliano browser still is functional, however it is neither optimized nor pretty (nor documented, for that matter). This will hopefully change in the near future (we will honestly get to it some time soon!). The browser now supports the new additions to MIPL version 0.6.

The driver and library are in even better shape, including additions to documentation and some clean-up (making things more consistent).

As of Wednesday, November 22, 2000

Small design flaw (easily correctable) noted in the card design. Click here for more details.

As of Friday, November 17, 2000

Production version (Rev. A) of the Tormenta ISA PC card received from the PC fabrication house (finally!!), tested, and found to be working quite well. Boards (both stuffed and un-stuffed) are now available for purchase in small quantity.

Also updated the above link (Production Rev. A card) to include bill of materials (sorry for omitting it earlier).

As of Wednesday, November 8, 2000

Production version (Rev. A) of the Tormenta ISA PC card designed, and submitted to PC fabrication house for production. Boards should be ready in about 1 week. For those wishing to fabricate their own boards, the gerber photoplot files are available on-line.

As of Tuesday, November 7, 2000

Version 1.1 of the Tormenta driver, the Zapata library, and the Emiliano MIPL browser released, and available on-line.

The Emiliano browser is functional, however it is neither optimized nor pretty (nor documented, for that matter). This will hopefully change in the near future. At least the audio sample rate conversion routines sound good now (not icky, like before).

The driver and library are in good shape, well documented, and seem to be functioning quite well.

As of Monday, November 6, 2000

Received, constructed, and tested PC boards of the initial version of the Tormenta PC card. They work quite well, with only a minimal amount of rework (wire jumpers, etc.). Documented the changes that need to be made to final production version.

Driver/Library Version 1.0 released today. This supports the actual production version of the Tormenta PC card. Will get together with our CAD dude in the next couple of days to incorporate the changes into the production version of the board design, after which we will do a board production run which should take aprox. 1 week.

Also, apparently the problem with ESF framing was unique to the wirewrap card (probably a board noise problem). The PC card version does not exhibit this problem. ESF is totally working and supported now.

As of Sunday, November 5, 2000

Driver/Library Version 0.8 (pre-release) released today. Added Caller*ID support.

As of Saturday, November 4, 2000

Driver/Library Version 0.6 (pre-release) released today. Added MF decode and generation to library. Added 'dial' (and extended tone generation) functionality to library. Tested and fixed outbound functionality (line starting, etc). Added transfer capability to the emliano MIPL browser (via the "xfer:" URI). Also found that there is a weird flaky problem with ESF line coding. Either there is a bug in the Dallas chip, or there is a noise problem on the wire-wrap prototype card. Either way, its not a big issue. It is EXTREMELY RARE for a channelized T-1 service not to be regular AMI/D4(SF). I actually have never seen one that wasn't, so if ESF just won't work, it really won't be missed. We'll see after we get the more "real" circuit boards on Monday.

As of Wednesday, November 1, 2000

Initial version of Tormenta PC card designed. This first attempt is mainly to determine feasibility of a 2 layer design. Should be getting boards back on Monday, November 6. If this design works out, we will change the layout slightly and get rid of the redundant clock-generation circuitry, and do an initial production board run, which I would imagine should take about 3-4 weeks.

As of Monday, October 30, 2000

Driver/Library Version 0.4 (pre-release) released today. The Zapata library now fully supports multiple channels/contexts per-process and is fully cleaned-up. The only thing remaining on this is to add 'dial' functionality (for outbound calls).

In addition, Current library source, and Current library header file are now available on-line.

The emiliano MIPL browser still needs better-sounding rate conversion (6000-8000 samples/sec and vicea-versa) routines (still have the same, old, icky-sounding ones), but works beautifully at 8000 samples/sec.

As of Sunday, October 29, 2000

Current driver source, Current header file, and Switching/Conferencing Archetecture now available on-line.

As of Saturday, October 28, 2000

Driver/Library Version 0.3 (pre-release) released today. The 'tor' driver now fully supports switching and conferencing, including the conference-link functionality (a conference listening to another conference), and pseudo-I/O channels (I/O channels not associated with a physical T-1 channel). Added extremely basic voice conferencing application (for demo/testing) to the 'emiliano' MIPL browser.

As of Wednesday, October 25, 2000

Driver/Library Version 0.2 (pre-release) released today. The 'tor' driver now fully supports alarm handling, flexible configuration specification (things such as framing, line coding, sync source). D4 (SF) and ESF framing, and AMI and B8ZS line coding are now selectable, as well as being able to specify primary and secondary network syncronization sources. Also handles loopback, both network-initiated and user-initiated.

As of Monday, October 23, 2000

I just released version 0.1 (pre-release) today. It seems to work well with the one wire-wrapped prototype card I have built. I will be meeting with my wonderful board CAD dude next week to have the PC boards layed out and should have a few of them in about 2 or 3 weeks.

This release includes:

The 'tor' driver version 0.1. It answers phone calls, handles audio in both directions, decodes DTMF, and does TDD in both directions. It does
not, as of yet, handle alarms, or loopback, but these are comparatively simple additions.

The 'Zapata' API library. It is reasonably fully implemented, but it needs, however, some major clean-up. Also it needs an improved, (okay
acceptibly working) sampling rate conversion routine to and from 6000 samples/sec to 8000 samples/sec. The one I wrote sounds icky.

The 'emiliano' MIPL browser. It seems to work. It uses the 'Zapata' API library. Its a good utility and example.


With the exception of items protected by the GNU General Public License (which are clearly indicated as such), the technologies, software, hardware, designs, drawings, schematics, board layouts and/or artwork, concepts, methodologies (including the use of all of these, and that which is derived from the use of all of these), all other intellectual properties contained herein, and all intellectual property rights have been and shall continue to be expressly for the benefit of all mankind, and are perpetually placed in the public domain, and may be used, copied, and/or modified by anyone, in any manner, for any legal purpose, without restriction.