The closest I had come was a P712+8N->P1 regulator. The input signal could come in any time, but the drive signals (testing whether an input signal had come in in the preceding interval of 712+8N ticks) had to be synchronized mod 8.

Suddenly, building a candidate P1 replacement for the old period-8N and period-60N universal regulators seemed like an insurmountable opportunity:

Code: Select all

`x = 130, y = 179, rule = LifeHistory`

67.2A$66.A.A$60.2A4.A$58.A2.A2.2A.4A$58.2A.A.A.A.A2.A$61.A.A.A.A$61.A

.A.2A$62.A2$75.2A$66.2A7.A$66.2A5.A.A$73.2A2$22.2A$15.2A5.2A$15.2A3$

17.2A17.2C5.A19.2A$17.2A17.C6.3A14.A3.A$11.2A21.C.C9.A13.4A$11.2A21.D

C9.2A$60.2A$60.A$57.2A.A$56.A2.A$57.2A$42.2A$42.2A3$64.A$29.2A3.A29.

3A$8.A20.A3.A.A31.A$8.3A19.A3.A.A29.2A$11.A6.3D10.A3.A.A.2A14.A$10.2A

8.D8.A.4A2.A.2A14.3A19.2A$19.3D6.A.A3.A.A21.A18.A$28.A.A2.A2.2A.2A16.

2A16.A.A$29.A3.2A2.A.A35.2A$37.A.A10.2A$38.A11.2A6.2A$58.2A$5.2A101.

2A$6.A101.A$6.A.A101.A$7.2A81.2A14.5A$88.A2.A13.A$88.2A.A.A12.3A$21.

2A67.A.2A15.A$21.A.A27.2A37.A15.4A$23.A26.A2.A2.2A30.3A10.2A3.A3.2A$

23.2A26.2A2.A.A29.A13.2A4.3A2.A$.C51.2A16.2A14.2A20.A.2A$2.C4.2A44.A

17.A37.A$3C4.2A41.2A.A2.A15.3A33.2A$50.A.2A.A.A16.A$54.A.A$51.2A2.A

44.2A$49.3A.2A45.A$48.A28.2A22.3A$25.A23.3A.2A21.A.A5.2A17.A$23.3A9.

2C.C12.A.2A21.A7.2A$22.A10.C6.C34.2A$23.A37.2A$22.2A3.2A3.C5.E2.C19.

2A7.2A17.A$24.2A2.A7.3E2.C28.A14.2A.A.A$22.2A2.2A4.C2.E32.A.A13.A.A.A

.A$23.A.A6.C2.2E4.C26.2A11.A2.A.A.A.A.2A$22.A2.A55.4A.2A2.A2.A$23.2A

8.C6.C44.A4.2A$35.C.2C44.A.A$48.2A33.2A$19.D28.2A$20.2D$19.2D3$24.2A

38.A$24.A38.A.A$25.3A35.A.A$27.A36.A$21.3D41.3A$23.D43.A$22.D3$58.2A$

28.2A29.A$28.A30.A.A$22.2A5.3A17.A10.2A$21.A2.A6.A15.3A$22.A.A21.A$

21.2A2.2A19.2A$23.2A2.A3.2A48.2A$21.2A3.2A4.A48.2A$22.A9.A.2A$21.A11.

A2.A$22.3A9.2A10.D$24.A19.D.D2.2A$45.2D2.2A$.3E$3.E$2.E$58.A3.2A$6.2A

49.A.A3.A$6.2A48.A.A3.A48.A9.2A$52.2A.A.A3.A49.3A7.2A$22.2A28.2A.A2.

4A.A5.2A18.A24.A$22.A33.A.A3.A.A5.A16.3A23.2A14.A$20.A.A29.2A.2A2.A2.

A.A2.3A16.A40.3A$20.2A31.A.A2.2A3.A3.A18.2A38.A$41.2A10.A.A15.2A53.2A

$41.2A11.A17.A$6.2A64.A.2A$5.A.A65.A2.A$5.A68.2A$4.2A83.2A13.2A$89.2A

13.2A$37.A11.2A$36.A.A10.2A$28.A3.2A2.A.A$27.A.A2.A2.2A.2A$18.3D6.A.A

3.A.A$9.2A8.D8.A.4A2.A.2A$10.A6.3D10.A3.A.A.2A52.2A.A6.2A$7.3A19.A3.A

.A56.2A.3A3.A.A21.2A$7.A20.A3.A.A63.A2.A23.A.A$28.2A3.A58.2A.3A2.2A

25.A$93.A.A31.2A$81.2A10.A.A$81.2A11.A$29.D11.2A$29.3D9.2A$29.D.D24.

2A$31.D23.A2.A$56.2A.A$59.A$59.2A$10.2A21.DC9.2A$10.2A21.C.C9.A$16.2A

17.C6.3A$16.2A17.2C5.A3$14.2A$14.2A5.2A$21.2A8$76.A$76.3A$79.A$78.2A

7$88.2A$81.2A5.A.A$81.2A7.A$90.2A2$77.A$76.A.A.2A$76.A.A.A.A$73.2A.A.

A.A.A2.A$73.A2.A2.2A.4A$75.2A4.A$81.A.A$82.2A!

This regulator should be driven by at least a period-1177 drive gun. That is to say, test gliders should come in on the white glider's lane (from the NW) no less than 1177 ticks apart. Other than that there's no periodicity requirement for the test input. Advance the white glider two ticks to see the difficult case that's handled by the almost-never-used extra eater catalyst.

The signal input is the yellow glider from the southwest. After it destroys the blocking eater, the next test glider will produce an output signal, and rebuild the eater. I've left open two possible output paths, marked with damaged eaters. Either or both can be used.

The next test signal should be sent about 1177*2 = ~2400 ticks later, for absolute safety. There's an optional eater (yellow, circled) in the center that can absorb most test-glider timings, but a few of them would cause trouble -- and it's not really part of the design to lose test signals like that, anyway.

This is just a early alpha version, of course. There must be ways to rewire this to rebuild the eater in less than 1000 ticks -- maybe quite a bit less. Anyone feel like improving the bounding box or the recovery time?