This is a notification regarding a minor software flaw discovered in the PicoDopp DF, in August 2007. If you have purchased a PicoDopp DF and you believe your PicoDopp DF exhibits the problems described here, you should contact me ( Bob Simmons ) about fixing the problem.

In August 2007, a minor software flaw was detected in the PIC code for PicoDopp DF, when bench tests were done on a VHF receiver with extremely narrow bandwidth. ( A Hamtronics R302 unit ) The result of this flaw was poor sensitivity to DF audio "tone", requiring a very high setting of the volume control on the reciever, to get a DF reading. Furthermore, sensitivity to weak signals ( 50% or more noise in the audio ) was reduced.

A detailed public announcement about the problem was made on the Yahoo PicoDopp public forum on 13 August 2007. The forum has over 150 members, most of them PicoDopp owners and users. After one week, ( when this web page was created ) no-one else has reported similar problems, so this may very well be a "non-problem", which is only evident in receivers with EXTREMELY narrow I/F bandwidths.

In any event, the repair of the flaw ( if required ) involves the replacement of the PIC chip on the PicoDopp MAIN board. Two resistor changes are also reccomended, but not essential.

The proper way to identify the problem requires some oscilloscope measurements of the reciever audio, but most people don’t have an oscilloscope, or they don’t know how to use one.

If you find that you must use a very high setting of the reciever volume control ( to get a PicoDopp DF reading ) then POSSIBLY this problem is the cause. This is especially likely if the receiver you are using has a bandwidth of less than 20 KHz, but it also can be caused if you are using a very small antenna, ( like the MiniAntenna ) with antenna spacing much less than ¼ wavelength between antennas. ( = lower DF tone in the speaker audio )

Please note that the audio sensitivity of the PicoDopp is deliberately designed to be fairly "low", so that it will not "trigger" on noise, if a signal is not present on the channel. The sensitivity level was selected so that the PicoDopp would not generate any DF messages on "pure noise", even if the speaker volume is set to its maximum possible value. ( = about 0.5 watts on most radios )

If you decide this problem applies to you, there are two possible ways to fix it, and you may elect to use either one.

If you think your soldering skills are pretty good and you think you can safely replace the PIC chip on the PicoDopp MAIN board, I’ll send you a new PIC chip, and you can perform the replacement yourself. The chip is a 28-pin SOIC device, with .050 inch pin spacing, which isn’t too hard to deal with, using ordinary soldering tools. If you elect to do this, I’ll be able to get a new chip off to you QUICKLY, and the time required to perform the repair will be minimized.

If you think your soldering skills are NOT adequate to perform this repair, I will provide a new PicoDopp MAIN PC board, ( complete ) for you to install. There will be no need for you to return the original PC board, you can just trash it. If you elect to do this, ( i.e. get a complete PC board ) the time required for me to provide a new board will be greater than the time required to replace the chip, alone. Replacement PC boards should start to become available by 1 September 2007.

Contact me ( Bob Simmons ) via e-mail with details of your purchase, ( name, address, purchase date ) and your preferred method of dealing with the problem. ( new chip or new PC board ) If you need further help to confirm that the problem really applies to you, I’ll try to suggest some further tests that are within your means. If there is still some doubt, I’ll probably just send a new PC board or a new PIC chip, and you can try it.

The problem involves a DSP "trick" I decided to employ in the PicoDopp code... the switching filter for the PicoDopp DF was activated only ½ of the time, to improve the Doppler SNR. ( that was the intention, anyway ) The filter was enabled immediately after any "new" antenna was selected, and disabled about 300 microseconds later.. this equals about ½ of the time alloted for each antenna while scanning. ( 573 microseconds / antenna x 4 antennas = 2.29 milliseconds = 436 Hz scan rate )

The idea was to sample only the Doppler pulse and ignore the 2^nd ½ of the time interval for each antenna... the Doppler pulse is created immediately after each antenna is switched, but due to time delays in the filters of the reciever, the pulse is not actually FM-detected until some time later... This is called "group delay" and it is never important in voice recievers, so the topic is somewhat obscure, among people familiar with "voice" radios.

For wideband recievers, the delays are small and insignificant, and the PicoDopp performs well. For really good recievers with very narrow passbands, the delays are greater, and as a result, a good portion of the Doppler pulse energy is simply being ignored by the software. ( see figure for graphic explanation of the problem )

To fix this problem, the sampling window has been increased from 50% to 100%, so that the design is no longer sensitive to variations of group delay, which will depend on the type of receiver employed.

The minimum receiver bandwidth useful for the PicoDopp DF probably is about 10 KHz... bandwidths less than that will cause group delays so great that there is insufficient time for each Doppler pulse to achieve completion before the next antenna is selected... in this case, a slower antenna scan rate is the proper solution, but that is beyond the scope of the original PicoDopp design, and would require some customized ( but simple ) PIC code changes.