I have a BK Precision 1801 Frequency Counter which I wanted to test its accuracy. I used a NEO-7M GPS module to output a 10MHz frequency when there is a satellite lock. I found that my unit was off by several hundred hertz. Adjusting the on board trim capacitor (C202) I was able to bring it to read 10,000,000Hz. After about an hour and a half the unit was reading 5Hz lower. Adjusting back to 10MHz and coming back after 3 hours the unit was reading 5Hz higher.

The manual indicates a temperature stability of better than +/- 10ppm (0-50c ambient). I believe the variations I am seeing above are due to this (it uses a 10MHz crystal).

The manual also indicates that "..If greater measurement accuracy is required than afforded by the internal time base, the instrument can be modified for use with a higher precision external time base. Modification instructions include provision for a switch to select either the internal or external time base..." (I could not locate any modification instructions in the manual)

Looking at the schematic, I believe the modification needed (for a permanent external time base [no switch]) would be to cut the trace between points A and B, and jumper A and C. A 10 MHz reference signal would then be fed into point D. I was thinking to use a TCXO. See attached schematic with modifications circled.

Does this sound correct?

Full Schematic: http://bama.edebris.com/download/b&k/1801/bk_1801.pdf

 

I realized that the link I provided earlier doesn't have the full schematic.

Please see: http://ftb.ko4bb.com/getsimple/inde...ounter_-_40MHz_Operator_Manual_Schematics.pdf

With regards to the previous stated modifications, would a TCXO with TTL output be fine?

 

I ended up making the modifications discussed in my earlier post and all seems to work well. After adjusting the TCXO the unit starts off at about 8-9hz higher (from what I recall) and settles to almost exactly 10,000,000hz (GPS) after about an hour or so of warm up. The 5V rail was a little low (around 4.85 to 4.75V) and changed when the devices input voltage changed (120V to 110V) so I powered the TCXO from a dedicated 7805.

 

YES, it looks OK. One alternative to the change would have been to add a heater to hold the crystal and time base circuit at a constant temperature. But that would want to be on all the time, wasting power. Some systems use that.
But now a question, since I believe that the 10mHz signal from a GPS receiver is controlled by the 1 Hz signal, are you 100% that the GPS signal was "perfectly stable"? And it is also possible that I got it backward, but I think that is what I read.

 

YES, it looks OK. One alternative to the change would have been to add a heater to hold the crystal and time base circuit at a constant temperature. But that would want to be on all the time, wasting power. Some systems use that.
But now a question, since I believe that the 10mHz signal from a GPS receiver is controlled by the 1 Hz signal, are you 100% that the GPS signal was "perfectly stable"? And it is also possible that I got it backward, but I think that is what I read.

I was thinking about using an OCXO but I think I would need some alternate way to power it as I don't think the current power supply could handle it (seeing that is already under 5V).

For the GPS I used the following circuit outlined at

which is from derived from http://www.scullcom.uk/design-build-a-gps-locked-frequency-standard-part-4/ . Louis Scully mentioned on his website the following:
".. I have compared it with my rubidium standard which would mean to within a few parts in 10 to the power 11."

 

I was thinking about using an OCXO but I think I would need some alternate way to power it as I don't think the current power supply could handle it (seeing that is already under 5V).

For the GPS I used the following circuit outlined at

which is from derived from http://www.scullcom.uk/design-build-a-gps-locked-frequency-standard-part-4/ . Louis Scully mentioned on his website the following:
".. I have compared it with my rubidium standard which would mean to within a few parts in 10 to the power 11."

A temperature controlled heater should always be powered from a separate supply because the voltage will certainly vary as the heater switches off and on, or even just varies the heat output. That is one of the reasons that they are not always used. So you are quite right about the power supply concern. I had not considered that when I suggested it, but yes, a temperature controlled time base crystal heater needs a separate power source, which does not need to be smooth power or regulated, though. But your package may not have room for the additional supply.