[time-nuts] Questions about TAC frontend, and some measurements

[Bruce Griffiths]

Attila Kinali wrote:
> On Sun, 23 Dec 2012 22:45:40 +0100
> Fabio Eboli<FabioEb at quipo.it>  wrote:
>
>    
>> Il 2012-12-23 07:42 Bruce Griffiths ha scritto:
>>
>>      
>>> The classic TAC using current mode switching is similar to the
>>> attached circuit schematic.
>>>        
>> http://pastebin.com/EkgqmgfE
>>      
> I have a couple of small questions about this circuit.
>
> Why are Q2 and Q13 driven by Vth? As there are inverted versions
> of the Discharge and Ramp signals available, wouldn't it make sense
> to use those to drive Q2 and Q13? I guess it would enhance switching
> speed.
>
>    
Only if the 2 complementary signals have closely matching propagation 
delays.
This is usually true with ECL logic but not necessarily true with CMOS 
logic.
Also the reverse emitter base breakdown voltage of faster transistors 
will be exceeded with 10V pp differential drive.
> What is the reason behind the emitter followers Q1 and Q9?
> Respecitvely, why shouldnt R3/R4, R7/R8 be connected directly to V+/V-?
>
>    
Avoidance of saturation and ensuring sufficent headroom for the current 
sources/sinks.

> Is there a special reson why the current source around Q9 is set to 20mA
> and the one around Q4 to 10mA? Is it because Q14/Q15 are driven by a 20mA
> current source while Q18/Q19 by a 10mA source?
>
>    
No special reason although if the long tailed pair driving a subsequent 
longtailed pair has a significantly lower tail current than the driven 
pair the second pair will switch more slowly.
> Am i correct, that the only current source whos value really matters is
> the one around Q11? If so, wouldn't it be beneficial to use a stable
> reference voltage (probably coupled with the ADC reference) to be used
> in an opamp based current source against GND or V- and a current mirror
> (cascode or wilson) to drive Q18/Q19 (while leaving the other LED based
> current sources as they are, including Q16)?
>
>    
Ideally the discharge switch current source should equal twice the 
charge switch current source to ensure equal currents (and ideally equal 
voltage drops across) in the clamp diodes when the ramp capacitor is 
fully discharged..

A mirror uses extra matched transistors that can be avoided if an opamp 
and reference is used to replace the LED's together with resistor 
isolated feedback from the current source emitter.

The ADC reference isn't always accessible particularly with an ADC 
embedded within a microprocessor.
> 			Attila Kinali
>
>    
Bruce