A simple formula exists for calculation of radiation resistance of a small transmitting loop in free space. The derivation is in most good antenna text books.
The expression depends on an assumption that current around the loop is uniform, so the question is what is the acceptable limit for loop size.
NEC might provide some guidance. A series of NEC-4 models of a octagonal loop of thin lossless wire in free space was constructed with varying perimeter. Perimeter shown is that of a circle of the same area.
Above is a comparison of the two methods of estimation of Rr. To the extent that we trust NEC-4, the graph indicates that error in the simple formula grows quickly for loop perimeter greater than 0.1λ. (The results using NEC-2 are visually identical.)
Many authors set the criteria for a small loop to perimeter<0.3λ, but it is clear that current is not sufficiently uniform for perimeter>0.1λ for estimation of Rr as 31149*(A/λ^2)^2 to 0.1pu error or better.
I bought a USBTiny AVR programmer on eBay for about A$8 posted.
Above, the seller’s pic of the package.
It is almost always the case that the ISP headers on the programmer use the standard pinout published by Atmel, and in that case the supplied ISP cables need to be pinned pin for pin, ie pin 1 to pin 1 etc.
If you look carefully at the pic, the key is towards the top of the pic which means pin 1 on the right hand plug is towards the viewer and pin 1 on the left hand plug is away from the viewer. The cable does not connect pin 1 to pin 1, and as a consequence the package did not work.
There is more than one way to connect these plugs, and above is one way that does connect pin 1 to pin 1, and the cable and USBTiny work. Though the seller has been told of this defect, he continues to sell the item with the pic of the defective cable.
One wonders how many thousands of these things are and will be sold with this defect.
There are a number of USB-RS232 adapters available, among them Silabs, FTDI, Prolific and WCH.
The Silabs CP series adapters have been reliable but not very widely used.
The Prolific adapters are, well, prolific… but given their major compatibility problems and poisoned drivers, they are not a good choice.
FTDI was a good choice until they distributed through Microsoft Windows Update their device poisoning trick to disable chips they thought counterfeit… though they seemed to have backed away from that action.
The new kid on the block is an adapter by Chinese company WCH, chips that appearing in lots of low cost devices, eg Arduino Nano clones with WCH adapters instead of FTDI selling for around A$3+ on eBay whereas the FTDI equipped Nanos are more like A$12.
This article compares the WCH CH340G and FTDI FT232RL.
The Fox flasher MkII is designed to run directly from a 1S LiPo battery at 3.8-4.3V. The battery can be charged by a simple voltage limited charger to 4.1V with a small loss in capacity. This article describes a simple solar charger for such a battery.
The regulator uses a TL431 precision programmable reference in a simple shunt regulator.