[time-nuts] Re: James Miller/Brooks Shera 10 MHz GPSDO

[Mervyn Thomas]

Thanks Ed - another thing I'll try.

Right now with the 3V cell in the EFC line the osc output freq is 
holding within the limits of not illuminating either the Low or Hi freq 
LED's and the actual 10MHz measured on an HP 5335A with a GPS locked 
Time Base is showing0.03 of a Hz high

Merv

On 2/09/2022 8:58 am, ed breya via time-nuts wrote:
> Since it appears there's no nice and easy way to access the OCXO 
> coarse tuning cap (there almost certainly is one in there, whether 
> it's accessible or not), a practical fix is a simple circuit mod (add 
> one resistor) to the DAC circuit.
>
> I looked up the AD1861, and see that it has the typical built in opamp 
> and feedback resistor for current to voltage conversion. Also, 
> apparently, depending on hookup and the data formatting, it can be 
> arranged to go between +/- 3V output. The simplest way to fix the 
> output range issue, without changing the basic hookup, any software, 
> or VCO gain factors, is to just add pure DC offset current at the 
> summing node.
>
> If you need the output voltage to be more plus, then pull some current 
> from the summing node into the -5V supply through a resistor. The FS 
> current from the DAC seems to be around 1-2 mA, so a value for R in 
> the few to tens of k-ohms range can move it significantly. Just figure 
> out what the scaling actually is, and set it up for enough offset to 
> keep the range good for a while - you can adjust again as it ages. Two 
> things that will be different from original is that the external 
> offset R won't track the internal feedback R thermally, and using the 
> minus 5V for a reference may not be as clean as you'd want. The data 
> sheet I found doesn't say much about the internal reference, but I'm 
> guessing it's a +1.25 or 2.5V bandgap. Or, it may simply be the +5V 
> supply.
>
> Since it's overall in a closed loop feedback system, the offset R 
> tracking is probably irrelevant, but the noise on the -5V may be an 
> issue, depending on the PLL characteristics. First prove the concept 
> with what you already have, then refine it as needed.
>
> Good luck.
>
> Ed
>
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe send an email to time-nuts-leave at lists.febo.com