Clifton Laboratories will be at Dayton! I'm helping Larry, N8LP, staff the Telepostinc booth and will bring a Z90 for those who might wish to see one in person. I will also have a couple Z100s and a Z1502 TDR interface for you to examine. Or, if you only wish to stop by and introduce yourself, please do so.

Because of sales tax reasons, no sales will be completed at Dayton. If you wish to purchase a Z100 (or a Z1502, for that matter) I will give you payment information which you may follow. Or, if you wish to pay for a Z100, I will accept cash or check, but your kit will be shipped when I return to Virginia.
 

My wife is off to Italy for a class on watercolor painting, so I can hang around my basement lair even more than normal for the next couple weeks.

I've made minor changes to the Z100 manual today and posted it to the web, available by clicking here or via the Z100 page or the Documents page. The current manual edition is version 1.2, dated 31 May 2007.

The change provides a useful trick to bending the short LED lead provided by Bill, VK4SQ—bend the lead around the shank of a 0.100" diameter drill shank. (Use a #38 drill [0.101" diameter], or a 3/32" drill. I built a Z100 yesterday using the modeler's bar clamp suggested by W2PY and VK4SQ's drill bit technique with excellent results.

I also have added an extra illustration to cutting the gray film using a template made from the clear lens's protective mask. This technique also works well.

I hope, over the next two weeks, to get a better handle on the TMS320VC33 DSP and University Development kit. There are three elements to learning a new micro-controller, and I'm having troubles with all three at the moment.

1. Learn how to use the programming tools. (Made more difficult by TI not updating most tools for Windows 2000 or XP, and doing the minimum necessary to make a few tools work for W2K/XP.)
2. Understanding the target's instruction set and internal organization, such as register assignments, internal hardware modules and the like. Here, at least, TI has provided adequate documentation, but not first class documentation.
3. Understanding the hardware needed to make it run, and the associated software interaction. Again an area where TI has not provided adequate documentation or sample programs.

The difficulty is that I'm still at the point where I can't always determine whether an error is due to 1, 2 or 3, thereby making it difficult to correct.

I'm gradually making some sense out of it, but TI has made the process far more difficult than it needs to be.

I suppose it's a generational thing, but I find it difficult to use on-line documentation for anything other than searching for small information segments. I can't easily read hundreds of pages of PDF-formatted text on the computer screen and get anything out of it. Hence, I spent a good part of yesterday printing selected TMS320VC33 documentation, totaling over 1500 pages. I ran the printouts by the local Kinko's printing shop this morning for binding and this afternoon I've been going through the material in a form that I can read and annotate, add Post-it notes to and generally find more conducive to learning.

 Along this line, I also ordered a hard copy of the most recent version of Steve Smith's book, The Scientist and Engineer's Guide to Digital Signal Processing. Dr. Smith makes a free PDF copy of the 1997 edition available at http://www.dspguide.com/ but I wanted a hard copy of the most recent (2002) version. His book, by the way, is exceptionally clearly written, and I highly recommend it if you have even a passing interest in DSP. Read the on-line copy—you'll like it. He has managed to make a topic normally encrusted with arcane mathematics understandable to anyone who has passed high school algebra. The computer routines are presented in BASIC and can easily be translated into your favorite programming language. I've used his book, and Numerical Recipes in Pascal, extensively when working on a Fast Fourier Transform for an audio spectrum analyzer program running on an 18F4620 PIC (written in Swordfish, of course.) Numerical Recipes is still in print and runs about $75 from Amazon, with used copies being available in the $15 range.

In going through the TMS320VC33 University Kit software today, printed documentation in hand, I discovered that it does have a rudimentary IDE for assembler programming. The IDE is part of the program DSK3DW.EXE. I would not call it a full-featured IDE, or even an easy to use IDE, but it does permit one to write assembler code in a window and then assemble, load into the DSP and execute the code, whilst viewing memory and register values, all from the IDE. The usual features, such as single stepping, breakpoints, etc. are supported.

Errors do not place the cursor at the error location, and the general interface is more than a bit clunky, but it seems usable.

At this point, I'm concentrating on the mechanics of how one deals with the assembler and debugger as well as how the TMS320VC33 is structured (memory, registers and the like) as well as its assembler instruction set (over 100 instructions, compared with fewer than 40 for the 16F series PICs). In addition, the printed circuit board has a number of jumpers that allow different configurations to be selected.

And, more interestingly, the PCB has a "complex programmable logic device." A CPLD is an integrated circuit with a collection of logic elements, gates, counters, dividers and the like, which are soft-connected internally, i.e., these elements may be assigned to input and output pins and connected internally via software commands. These devices are intriguing combinations of hardware and software, but for the purpose of learning how a DSP is programmed, it adds another layer of complexity. TI's University DSK uses the CPLD to allow the TMS320VC33's various pins to be connected to the outside world and to other ICs on the board in different fashions, depending on the CPLD's configuration program parameters. Fortunately, I can go a long way with the default CPLD configuration. 

Over the next two or three days, my goal is  to write an assembler program to initialize the on-board A/D and D/A converters and to read the A/D value and echo that value to the D/A. This will run at 48 KHz sample rate. Once that is running, I can be a bit more adventuresome, and try filtering. There are sample programs that include this sort of echo function, but I'm not quite at a point where I'm ready to chop down the sample programs.
 

Z100 Updates

I've updated the Z100 LED Tuning Indicator manual and schematic. Both are available via the Z100 page, or the Documents page, or by clicking the links in the first sentence.

Changes to the Z100 manual are mostly in the area of merging the two update sheets into the document. The update sheets cover options for installing the LEDs and the gray film.  After building a couple more Z100s for purchasers, I've found a better way to install the film over the entire lens, as described in the revised manual. If you have built a Z100 and wish to modify how you have installed the film, but need another film strip, drop me an E-mail and I'll send you an extra piece of film. Having the entire lens covered with the film helps the Z100's apperance as the printed circuit board edge and downward projecting leads have reduced visibility.

Texas Instruments TMS320VC33 DSP Development Kit

I've experimented the last few days with Texas Instruments' $150 TMS320VC33 DSP Development Kit (the "University DSK"). The kit is based on TI's TMS320VC33 DSP chip, running at 150 MHz clock, offering 75 MIPS processing power. The kit includes stereo A/D and D/A converters, extra memory and other features. The TMS320VC33 is a 32-bit floating point device, and is a good fit of price versus performance for the application I have in mind. More information on it is available at http://focus.ti.com/docs/toolsw/folders/print/tmdsdsk33.html.

My thought is that a new panadapter design would benefit from adding a DSP behind a wide crystal filter, to provide multiple bandwidth options, clearing the way for increased span width selection and a faster sweep. (Faster sweeps also require a different display technology, and I'm looking at that as well.)

To make a long story short, I'm disappointed with the University DSK board, or, I should more correctly say, I'm disappointed with the documentation and supporting software, as the board hardware seems to work well. 

The University DSK's interface is via the parallel port. I initially thought it would work with a USB-to-printer adapter, but that turns out to be incorrect, as the associated software plays tricks with the parallel port that are only supported with a genuine hardware parallel port. (Parallel port transfer is used because it can be much faster than RS232, the only other access method commonly available when the protocol was developed 20 years ago.)

The Gateway laptop computer I use with Swordfish and Delphi to develop software has only USB ports, so I installed the DSK's software on an older Gateway 500 MHz Pentium laptop with a genuine hardware parallel port, and Windows 2000 operating system. Then I discovered that about two-thirds of  the demonstration programs supplied with the DSK will only run under Windows 95 or Windows Me. This fact is stated in the DSK's "read me" documentation and seems to be correct. Although these programs run in a DOS box, there is apparently enough difference between W2K (and XP) and earlier versions to break the code.

Although the DSK comes with software tools—assembler, debugger, C compiler, linker and the like—the tools have mostly not been updated since they were developed around 1995. Even by 1995 standards, however, the tools are primitive, and are DOS command line programs. No editor, let alone an "Integrated Development Environment" or IDE. Some automation is possible via DOS batch files, but Borland offered a DOS-based IDE with its early Turbo Pascal compilers in 1981-82 that is head and shoulders above TI's tools.

Some of the supplied tools are Windows 2K/XP compatible, such as the Windows debugger, so in theory it is possible to develop code for the University DSK board with the supplied software, but it will be a painfully slow process.

In order to program the University DSK with a modern IDE-based compiler or assembler, it's necessary to purchase two additional items:

1. A JTAG adapter. The standard way to program code into TI's DSP is via JTAG, which is a four/five-pin interface added to a chip, designed so that multiple chips on a board can have their JTAG lines daisy chained together.This way, the "JTAG Emulator" need only connect to a single JTAG port to have access to all chips on the PCB. The standardized JTAG connector is a 14-pin header and the electronics that connects that 14-pin header to your PC is called a JTAG Emulator. (It's far more than just a 14-pin adapter, of course, as it must convert data from your PC into the format expected by the PCB and also convert data received from the PCB to a format expected by the software running on your PC.) JTAG emulators connect to the PC via the parallel port, PCMCIA card slot, an ISA slot or, most recently, via the USB, depending on the emulator model. JTAG emulators from a major supplier are priced at about $1.5K for the least expensive unit, up to $4K. (I've found Chinese knock-off JTAG emulators for as little as $300, but the story on the street is that they are of questionable quality and may work with some chips but not others.) [In addition to the parallel port interface that works with the specially supplied software associated with the University DSK, the board also contains the standard JTAG interface.]
2. A copy of TI's Windows-based, IDE-style Code Composer software, v 4.12. Runs about $1.5K, although it's available for $600 if you have a university mailing address and thus qualify for TI's educational discount.

I'm considering two alternatives.

• The EzDSP, which is similar to the University DSK in that it runs the TMS320VC33 DSP chip, but with a built-in JTAG interface (still uses parallel port access, but the JTAG emulator lets it communicate with Code Composer). It comes with a restricted version of Windows-based Code Composer v.4.12. (Restricted to being used with a PCB having the on-board JTAG.) The EzDSP is available from Digital Spectrum for $519. http://www.spectrumdigital.com/product_info.php?cPath=43_91&products_id=83
• TI's TMS320C6713 DSP Starter kit, http://focus.ti.com/docs/toolsw/folders/print/tmdsdsk6713.html, available for $395. It includes on-board JTAG emulation and USB interface. It also comes with a restricted version of Windows-based Code Composer. However this board uses a more expensive DSP, running at 1 GHz clock. For the purposes I have in mind, a TMS320C6713 is not an optimum selection.

More later in June as I will have a chance to use the ExDSP board then.
 

All Z91 owners have now been shipped a replacement panel set (except for one Z91 that was shipped with the good panels, and for Z91's sold with Z90 conversion kits.)
 

I shipped the one assembled Z100 ordered at Dayton today. That completes all orders I received at Dayton.

I've also added a new page, Dayton 2007, with a few photos of the Hamvention and my comments.
 

Made it back from Dayton. It's a 10 hour drive to Clifton, and as much as I enjoyed the show and meeting about 20% of my domestic Z90/91 customers, it's good to be home. I'll post more Dayton photos tomorrow, perhaps as a separate page.

I also wish to congratulate the guys running the "Four Days in May" events. The banquet was exceptionally well done, and pre-drawing the smaller prizes was an excellent idea. (I won a small prize, as a matter of fact. A toroid assortment from Kits and Parts. I donated a Z100 kit as a banquet prize as well.)

All the Z100 kits ordered at the show are now shipped. One customer ordered an assembled Z100 and that won't be done until tomorrow.

I've sent notices with the tracking number to all Z100 customers for whom I could find an E-mail address, so if you have not received a message from me, please drop me an E-mail and I'll give you the information.

Also, I've changed the Z100 manual supplement, based on comments from two early builders. The kits shipped today have the updated supplements. I plan to revise the manual over the next couple days, but a couple of customers from VE3-land that I met at Dayton do not have these supplements. Or, if you were the Z100 winner at the FDIM banquet, you do not have the most recent supplement.  If they will drop me an E-mail, I'll send the most recent copies. Also, if you have not yet started building your kit, drop me a message and request the most recent supplement. (The supplement applies to building, not operating, so if your kit is already built, don't worry.)

The revised manual will be posted for downloading, when finished, of course.
 

This morning was occupied by helping Larry set up the Telepostinc table. This evening, I visited the QRP Amateur Radio Club, International's Four Days in May "Meet the Authors" event. I had intended to attend several presentations but did not get around to it.
 

I've begun receiving feedback from the first Z100 builders and I've added a new page Z100 Builder's Notes to make their comments available to other builders. You can reach this page by clicking on the link above, or from the Z100 page. In order to keep the clutter down, I'm not going to make it a  top level page. (And, I've been contemplating re-arranging the site to put all Z90/91 pages under a single sub-page, and probably will do that over the next few weeks.)

Made it to Dayton, total time just under 10 hours and distance driven 501.3 miles according to my GPS receiver. The Homewood Suites has wireless internet access and I've logged in successfully.  The first three hours were pleasant enough, at least to the extent driving on an interstate highway is ever pleasant. Then the rain started as I passed through the Appalachian mountains. It stopped around Wheeling WV and the rest of the trip was overcast, but dry.

The long term weather forecast calls for dry weather during the Hamvention, but on the cool side. I've been to Dayton under the old schedule (in late April, if my memory is correct) when it was bitterly cold and wet. Generally, the mid-May schedule has been decent, although often showery.

It's been five or six years since I've been to a Dayton Hamvention and the area around the hotel (near Wright-Patterson Air Force Base) has gone from open fields to houses, offices and hotels.

I'm packing up for Dayton and will be out of easy E-mail contact until Sunday evening, May 20th. At Dayton, I can be found at Booth 517, or via cell phone. The cell phone number is Area Code 201, followed by 757 2121.

I'm staying at the Homewood Suites near Wright-Patterson AFB.

All Z100 kits have been shipped and the purchasers given tracking/confirmation numbers, and one was delivered this afternoon.

I am bringing the laptop and digital camera, so you may see a photo or two of Dayton posted here.
 

This week will be occupied by the Dayton Hamvention. I'm shipping all Z100 kits for which payment has been received this morning. (And the last Z91 kit as well.) Next week, I'll pack and ship the replacement Z91 panel set, as well as any Z100 orders that might arrive while I'm at Dayton.

I've added a series of photos to the Z100 page showing the hardware evolution from my initial breadboarding to the finished printed circuit boards.

All parts arrived on schedule, and all Z100 kits for which I have received payment go out tomorrow. It took most of Thursday and Friday to sort through the parts and individually pack them into kits. Since the Z100 is a simple kit, with no surface mount components, the packaging is less elaborate than I used in the Z90/91 kits. No stages, as the entire build takes about 2.5 hours.

I built three kits over the weekend, one to show at Dayton and two for customers ordering completed units. All three went well and the final PCBs have no errors. I found a few things to fix or expand in the instruction manual and plan to do that later today. When finished, I'll post the revised PDF copy of the manual as well.

Curious what 1700 LEDs look like? Here are 1,000 red, 500 yellow and 200 green LEDs ready to be put into kit packages. And, yes, I counted every one that went into the kit bags. I won't guarantee that I didn't make an error, but I tried my best.

The only parts remaining for the Z100 kits are the printed circuit boards, scheduled for delivery tomorrow. A large box of parts from Mouser arrived today and I've spent the evening putting holes in enclosures. In two hours and thirty minutes, I completed 11 enclosure sets, a bit under 15 minutes per enclosure.

I should meet my objective of shipping all Z100 kits for which I have received orders on Monday. It's possible, of course, that I've messed up the PCB, as I did make one small changes after the last prototype. If so, it would be highly embarrassing, to say the least.

I've signed up for the QRP Amateur Radio Club International banquet at Dayton and also will have a Z100 (and also a Z90, Z91 and Z1502) at the vendor's exhibit.

I've posted the other news of the day—the 3rd attempt at Z91 panels arrived from Emachineshop today. I have not gone through all the box, but I think there are enough good panels to meet all the Z91 needs. I'll begin mailing out replacement Z91 panels the Monday after Dayton, or more likely, the Tuesday after Dayton, as I'll need at least a day to recover from the 10 hour drive between Clifton VA and Dayton OH.
 

I've revised  the Z100's Assembly and Operations manual and posted the revised file. It may be read by clicking here or via the Z100 page.

 If you wish to order a Z100 via PayPal, you can compute the amount due (the kit plus shipping) and send the amount via PayPal to orders@cliftonlaboratories.com. I've verified this link is working and receives PayPal payments.

One batch of Z100 parts arrived today, and I expect the remainder to arrive tomorrow and Friday. Absent any last minute problems, this will put the first kit shipments out Monday.

Error in PayPal address -- I had a typographical error in the PayPal link on my Z100 page from the time it was posted until 0600 EST today. It's now fixed, but if you sent an order via that address during that time, please cancel it and resubmit to the correct address: orders@cliftonlaboratories.com.

It's now working and my apologies to anyone inconvenienced by my error.
 

At long last, the Z100 CW tuning aid kit page is available by clicking here or via the navigation link at the left.

The first Z100 production parts showed up this afternoon—the laser-cut clear Plexiglas front panel. Based on the shipment schedule, all remaining parts should arrive by Thursday or Friday, so kits can go out early next week--if there are any orders, of course!

The Z100 parts orders are all placed and should arrive over the next few days. I'll assemble a Z100 page tomorrow (promise), with pricing and ordering information. Today, I built two simple fixtures to hold the case for machining. I will mill the access holes in the case bottom and end using the fixtures and my milling machine.

The Z100 manual is now about 50% longer than before, as I've added many photos and drawings. Here's one example.
 

As usual, the April Updates have been moved to an archive page reachable via the links at the top of this page, or by clicking here.

I've spent most of the day revising the Z100 manual and in the process found a small error in the most recent prototype board. Not enough to justify a new prototype run, so I'll fix the layout tomorrow and get the board order in.

Shown below is an assembled Rev 2 board. I made an error assembling it and only put two yellow LEDs in. Normal configuration is two center green LEDs, flanked by two yellow LEDs on each side and the remaining 18 LEDs are red.

All the reddish LEDs are deep red, although the camera flash makes some look orange.

The Z100 thus permits you to quickly and efficiently tune a CW signal so that you are "zero beat" with the received frequency. And, unlike simple tuning indicators that show GO/NO GO, the Z100 informs you how far off frequency you are and in what direction you must tune for zero beat.

For RTTY and other digital mode operators, the Z100 will let you tune a digital signal in correctly without having to keep one eye on your computer's software frequency display and the other eye on your transceiver.

Since this is a digital device, you can store 16 center frequency settings, such as one for CW, one for low tone RTTY, one for high tone RTTY, one for PSK, etc.

And, the Z100's firmware is open source. I'll put it up on the Z100 page (when I get around to making the page, that is) and you can modify it to your liking. You don't like 25 Hz per LED? Change it.

The Z100's firmware is small enough that Swordfish's free compiler accommodates the code. And, since the 18F2620 PIC supplied with the Z100 has bootloader software, you won't need a programmer. All that is required is a USB-to-TTL logic level adapter cable. It costs $20 from Mouser.