[time-nuts] DAC resolution in GPSDO

[Richard Moore]

On Dec 1, 2008, at 12:09 PM, Warren wrote:

> Message: 4
> Date: Mon, 1 Dec 2008 11:22:42 -0800
> From: "WarrenS" <warrensjmail-one at yahoo.com>
> Subject: Re: [time-nuts] V standards
>
> Concerning the Need for very high resolution Dacs in GPSDO.
>
>> Bruce said: "However there are low noise OCXOs with EFC adjustment
>> ranges of 1E-6 or more, (that would benefit from >20 bit Dacs)

>> Question:  Is the Aging rate of these low noise OCXO units poor  
>> enough
> that you could not use a couple of fixed precision resistor and/or  
> a pot
> for course adjustment and reduce the EFC range of the DAC by say  
> 1/100 ?.
> Are there low noise OCXO (that are being used in GPSDO) whose long  
> term
> drift over say 6 months would need more than 1E-8 or so of   
> 'automatic' turning?

I'm using an Isotemp 134 OCXO that has an EFC sensitivity of around  
1.62E-6 Hz/V near the 10MHz point. It is very stable and also well- 
aged, and its non-adjustable EFC V for 10MHz is near 3.318V. I'm  
running it in Bert Zauhar's FLL GPSDO, which has a PWM DAC with 14- 
bit resolution and an output range of 0 to +5V. This DAC resolution  
is not enough for tight frequency control of the Isotemp, given its  
very coarse Hz/V sensitivity. Combined with the DAC resolution, the  
system granularity is roughly 4.9E-11 Re 10MHz for it's finest step.  
This just isn't quite enough.

> And concerning the 10811A. Should one consider reducing its EFC
> tuning range by say 2 to 20 to ease the requirements of the Dac?
> Even if the EFC tuning range was reduced by just a factor of two,
> It could then be done with just a standard 0 to 5 volt Dac
> instead of the -5 to +5 that I saw suggested by its spec.
> I would think that a well aged unit would be much better than its  
> worse case
> spec of 1E-7 per year, And even if not, I'm guessing  that many of  
> the time-nuts,
> would be very welling to trade off  it needing a have few extra  
> manual adjustments
> in order to get better performance.
>
> WarrenS

Warren, that's exactly what I'm doing with the Isotemp, aiming for a  
division of about 6-10x. It's no problem for folks like us to tweak  
stuff when needed, so that's an effective solution. I'd like to have  
a higher resolution DAC to preserve a large tuning range, but the  
whole idea of this GPSDO was a simple, inexpensive, yet effective DIY  
GPSDO. See www.moorepage.net/gps2.html

Dick Moore