For the bands 40m and 30m the situation is very simple:
- 10,140MHz crystal available from N4ESS (no change anyway)
- 7,040MHz QRP crystal available
- 7,038MHz QRP crystal, which could serve as a basis for a d.c.-receiver
80m slightly more tricky, but no big deal at all!
- 3,595 = 7,190 / 2
For the bands 160m, 20m, 17m, 12m and 10m, a combination of two crystals will do the trick to generate a frequency in the QRSS range.
The following frequencies (in MHz) are reached by mixing of two fundamental crystal frequencies:
- 1,8372 = 8,8672 - 7,030
- 14,096 = 10,000 + 4,096
- 14,096 = 18,096 - 4,000
- 14,097 = 12,000 + 2,09715
- 18,106 = 8,000 + 10,106
- 24,926 = 14,7456 + 10,180
- 28,125 = 18,000 + 10,125
It seems appropriate to note that the 12m combination under point 6 uses a 10,180MHz crystal which can be found in older crystal-synthesized CB radios.
The 15m band can be reached by means of multplication and division:
- 21,096 = 4,096 + 2 * 8,500
- 21,097 = 2,09715 + 38,000 / 2
Solution #2 looks more promising to me for the following reasons. 38,0 divided by two gets us to 19,0. Mixing products will be 21,097 and 16,903. We certainly would need a band trap for 16,9MHz, which should be doable. As divider I would propose the usual Flip-Flop. The best part is, there are canned oscillators available for 38MHz.