[time-nuts] Fine print on HP 5334B

Dr Bruce Griffiths bruce.griffiths at xtra.co.nz
Fri Feb 23 07:00:21 EST 2007


SAIDJACK at aol.com wrote:
>  
> In a message dated 2/23/2007 05:58:09 W. Europe Standard Time,  
> bruce.griffiths at xtra.co.nz writes:
>
> Current  mode ramp generators using long tailed pairs to switch the 
> charging  current for a capacitor are the time honoured approach for fast 
> ramp  generators.
>
>
> Hi Bruce,
>  
> would you have some sample schematics for a more modern,  faster version of 
> this than let's say the one used in the 5334A etc that  you could share?
>  
> thanks,
> Said
> <BR><BR><BR>**************************************<BR> AOL now offers free 
> email to everyone.  Find out more about what's free from AOL at 
> http://www.aol.com.
> _______________________________________________
> time-nuts mailing list
> time-nuts at febo.com
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>
>   
Said

The following paper illustrates one way of building such an interpolator 
using a custom IC.
http://www.imec.be/esscirc/papers-96/122.pdf

The critical part which limits switching speed in a discrete limitation 
is the inductance of the connection between the two emitters of the long 
tailed pair switches.
It would pay to use a transistor array (e.g. HFA3046 or even better a 
faster version of the CA3054 or CA3102) where the emitters are already 
connected together on chip.
With a modern capacitive charge redistribution ADC, the ADC sampling 
capacitance could form part of the capacitor being charged by the 
current source.
The current source is gated on for the time interval to be measured, the 
ADC conversion is then triggered after waiting long enough for the 
charge on its sampling capacitance to settle. Following the conversion 
the ramp capacitor is discharged. The capacitor reset switch being 
opened just before the current source is turned on.
The required timing is easily generated by a  3 (or more) stage 
synchroniser. The interpolator being used to measure the input

Its actually much easier to use a TDC chip which can achieve a 
resolution of 10ps and a differential nonlinearity of a few picosec as 
long as the delay measured by the TDC is greater than 200ns. Lower 
frequency synchroniser clocks (5-10 MHz) suffice compared to those 
required with a ramp type interpolator of the same resolution.

Essentially a pair of TDCs is used to measure the delay of the start and 
stop channel synchronisers, with a 3 stage synchroniser the measured 
delays by the TDCs can easily be kept in the range of 200-400ns even 
though the time interval between the start and stop edges may be a short 
as a few tens of picosec.

Bruce



More information about the time-nuts mailing list