Hi Speed Comparator operation


David GM6BIG
 

Hi Rich,

there were a couple of mistakes on the PCBs depending on the version.
All from memory, and still waking up this morning !

Missing ground on the little 5v regulator.
O/P from the op amp to wrong pin on next chip.
Think I biased up the opamp o/p after coupling cap to centre the waveform better.
One of the divider tracks was wrong too.
(missing track or wrong legend on PCB).
Feedback for the divider chain needed adjusting (I had used an alternative MMIC).
Maybe not obvious, but a track(s) needs cut depending on the configuration, and wires run in.
Sure I used an alternative opamp in the end, the original options hard to get and quite expensive.
Paul has corrected these over the years, but all worth checking - trust nothing !

Will check later to see if I have made any notes on these and my emails to Paul, see if I can be more precise.

Cheers, DAvid  GM6BIG

On 20/05/2021 11:04, G8ZHA via groups.io wrote:
I have built a couple of W1GHz projects: a GPS locked 10MHz osillator and a 100MHz PLL board. Each of his circuits uses a LT1116 hi speed comparator to lift the oscilator signal up to 5V CMOS logic levels to feed the dividers.

On the 100MHz PLL PCB, there is a 3.3V osc, feeding a MC12080 Div by 10 prescaler, AC coupled into the non-inverting input ofthe LT1116. The inverting input is tied directly to ground. I can see a nice square wave signal into this pin, about 1.5V pk-pk, centered on 0V. The outputs, which are open circuit, are very distorted, see photos.

I thought that this may be because the signal is outside the common range of 0 - 2.5V for the LT1116, so I biased the input to the middle of the voltage range but was not successful.

Am I missing something obvious?

I also built the 10MHz version, which feeds the osc straight into the comparator. I hadn't got a LT1116, so used a LT1016 instead, which has a common mode input range of 1.25 tp 3.75V range. Again, I biased the input signal to the middle of the voltage range, but saw the same distored oyutput.

(in the UK, the LT1116 is not available from RS etc)

I am at a loss to understand what is going wrong.

Rich G*ZHA
--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus


Rien Eradus PA0JME
 

https://www.analog.com/media/en/technical-documentation/data-sheets/lt1016.pdf

Op 20-5-2021 om 15:18 schreef Rien Eradus PA0JME via groups.io:

A bit further datasheet shows

Input Impedance and Bias Current

Input bias current is measured with the output held at 1.4V. As with any simple NPN differential input stage, the LT1016 bias current will go to zero on an input that is low and double on an input that is high. If both inputs are less than 0.8V above V, both input bias currents will go to zero. If either input exceeds the positive common mode limit, input bias current will increase rapidly, approaching several milliamperes at VIN = V+.

Op 20-5-2021 om 12:04 schreef G8ZHA via groups.io:
I have built a couple of W1GHz projects: a GPS locked 10MHz osillator and a 100MHz PLL board. Each of his circuits uses a LT1116 hi speed comparator to lift the oscilator signal up to 5V CMOS logic levels to feed the dividers.

On the 100MHz PLL PCB, there is a 3.3V osc, feeding a MC12080 Div by 10 prescaler, AC coupled into the non-inverting input ofthe LT1116. The inverting input is tied directly to ground. I can see a nice square wave signal into this pin, about 1.5V pk-pk, centered on 0V. The outputs, which are open circuit, are very distorted, see photos.

I thought that this may be because the signal is outside the common range of 0 - 2.5V for the LT1116, so I biased the input to the middle of the voltage range but was not successful.

Am I missing something obvious?

I also built the 10MHz version, which feeds the osc straight into the comparator. I hadn't got a LT1116, so used a LT1016 instead, which has a common mode input range of 1.25 tp 3.75V range. Again, I biased the input signal to the middle of the voltage range, but saw the same distored oyutput.

(in the UK, the LT1116 is not available from RS etc)

I am at a loss to understand what is going wrong.

Rich G*ZHA


Rien Eradus PA0JME
 

A bit further datasheet shows

Input Impedance and Bias Current

Input bias current is measured with the output held at 1.4V. As with any simple NPN differential input stage, the LT1016 bias current will go to zero on an input that is low and double on an input that is high. If both inputs are less than 0.8V above V, both input bias currents will go to zero. If either input exceeds the positive common mode limit, input bias current will increase rapidly, approaching several milliamperes at VIN = V+.

Op 20-5-2021 om 12:04 schreef G8ZHA via groups.io:

I have built a couple of W1GHz projects: a GPS locked 10MHz osillator and a 100MHz PLL board. Each of his circuits uses a LT1116 hi speed comparator to lift the oscilator signal up to 5V CMOS logic levels to feed the dividers.

On the 100MHz PLL PCB, there is a 3.3V osc, feeding a MC12080 Div by 10 prescaler, AC coupled into the non-inverting input ofthe LT1116. The inverting input is tied directly to ground. I can see a nice square wave signal into this pin, about 1.5V pk-pk, centered on 0V. The outputs, which are open circuit, are very distorted, see photos.

I thought that this may be because the signal is outside the common range of 0 - 2.5V for the LT1116, so I biased the input to the middle of the voltage range but was not successful.

Am I missing something obvious?

I also built the 10MHz version, which feeds the osc straight into the comparator. I hadn't got a LT1116, so used a LT1016 instead, which has a common mode input range of 1.25 tp 3.75V range. Again, I biased the input signal to the middle of the voltage range, but saw the same distored oyutput.

(in the UK, the LT1116 is not available from RS etc)

I am at a loss to understand what is going wrong.

Rich G*ZHA


Rien Eradus PA0JME
 

Hi I am not familiar with both devices but took a glance at de 1016's datasheet. It tells me that inputbias current is about 10uA.  If Vss is at ground there can be no bias current at all unless input consists of pnp pair. Maybe you have to lift both inputs a bit to allow for biascurrent to flow and decouple the inverted input. That is what the datasheet shows providing bias through a decoupled resistor from the output. Tieing a i/p to ground is never a good idea unless a device has both positive and negative supplies.

Best regards Rien PA0JME

Op 20-5-2021 om 12:04 schreef G8ZHA via groups.io:

I have built a couple of W1GHz projects: a GPS locked 10MHz osillator and a 100MHz PLL board. Each of his circuits uses a LT1116 hi speed comparator to lift the oscilator signal up to 5V CMOS logic levels to feed the dividers.

On the 100MHz PLL PCB, there is a 3.3V osc, feeding a MC12080 Div by 10 prescaler, AC coupled into the non-inverting input ofthe LT1116. The inverting input is tied directly to ground. I can see a nice square wave signal into this pin, about 1.5V pk-pk, centered on 0V. The outputs, which are open circuit, are very distorted, see photos.

I thought that this may be because the signal is outside the common range of 0 - 2.5V for the LT1116, so I biased the input to the middle of the voltage range but was not successful.

Am I missing something obvious?

I also built the 10MHz version, which feeds the osc straight into the comparator. I hadn't got a LT1116, so used a LT1016 instead, which has a common mode input range of 1.25 tp 3.75V range. Again, I biased the input signal to the middle of the voltage range, but saw the same distored oyutput.

(in the UK, the LT1116 is not available from RS etc)

I am at a loss to understand what is going wrong.

Rich G*ZHA


G8ZHA
 

I have built a couple of W1GHz projects: a GPS locked 10MHz osillator and a 100MHz PLL board. Each of his circuits uses a LT1116 hi speed comparator to lift the oscilator signal up to 5V CMOS logic levels to feed the dividers.

On the 100MHz PLL PCB, there is a 3.3V osc, feeding a MC12080 Div by 10 prescaler, AC coupled into the non-inverting input ofthe LT1116. The inverting input is tied directly to ground. I can see a nice square wave signal into this pin, about 1.5V pk-pk, centered on 0V. The outputs, which are open circuit, are very distorted, see photos.

I thought that this may be because the signal is outside the common range of 0 - 2.5V for the LT1116, so I biased the input to the middle of the voltage range but was not successful.

Am I missing something obvious?

I also built the 10MHz version, which feeds the osc straight into the comparator. I hadn't got a LT1116, so used a LT1016 instead, which has a common mode input range of 1.25 tp 3.75V range. Again, I biased the input signal to the middle of the voltage range, but saw the same distored oyutput.

(in the UK, the LT1116 is not available from RS etc)

I am at a loss to understand what is going wrong.

Rich G*ZHA