VCO part repair:
VCO output (one that goes to band 2 converted) should be about 0-1dBm. If not the U2 is probably faulty.
VCO amplifier is 3 stage amplifier:
u3 - MAR-8, Q6 - rf bjt, and the big transistor mounted on the body.
Enable VCO amp (test 1 0, 1 8 4 3, 3 f(reset)).
Output after u3 should be about +7dBm. Output of the BJT transistor should be about +17dBm.
If VCO amp is disabled (test 10, 1843, 37) last 2 stages are turned off, and but U3 stays enabled.
In my case U3 was faulty. As MAR-8 is very hard to obtain I went for easier to obtain RF amplifier - SPF5043Z. I replaced U3 with U2. This way the VCO amplifier is in the same specs(or similar) as before. And i placed SPF5043z in U2 position.
SPF5043Z is bit different: Input voltage is max 5Volt. I placed little LDO in SOT23-5 package on free ground plane. Wired 12volts to input, and output directly into the L7 inductor through 0ohm resistor. R14 (100ohm resistor) was removed.
I also had to replace R32 (200ohm resistor) with 50ohm, to get the signal strength a bit higher. At -6dBm output the counter selftest (test 01) would not count 200MHz.
Here is picture of repaired VCO output:
http://robot.mysteria.cz/IMG-20200220-WA0002.jpegMess is just flux residue, ground planes are hard to solder, as they are placed on the aluminum case.
I tested the counter to be mostly in specs. -30dBm signal at 6GHz through 1 meter of RG316(~ -2dB) + 2x N to sma (~ -2dB?) ~36dBm sensitivity.
Second counter would not lock VCO to correct frequency. It turned out to be MC12013 on the converted control board. These ECL logic circuits are almost unobtainable (china virus stopped postal service) or for very high price (>20eur + s/h). I suggest scrapping from old equipment. I scrapped mine from 14GHz SSPA from the same era. MC13013 with MC10131 form 40/41 dual modulus prescaler. If someone knows some 40/41 dual modulus prescaler which can go up to 550 MHz please let me know. (SP8793 or MC12016 are not good alternative)
I hope this will help someone in future :-).