开云体育

ctrl + shift + ? for shortcuts
© 2025 Groups.io
  1. Electronics101
  2. Topics

Buffer amplifier for 10MHz ?

 

I am using an SI5351 to generate 10Mhz and 100Mhz digital reference signals via 50ohm 5m long BNC cables but the SI5351
seems to struggle producing a reasonable square waveform at 10Mhz.

What would be the best buffer circuit for this application? and could the same work for 100Mhz ??

Regards,
Dave
wn4isx
 

The SN74LVC1G17 is rated to ~100MHz, most will work up to at least 150MHz.

?

Operation is simple, feed you sloppy signal in and get a clean signal out with extremely sharp transitions, it won't be a perfect 50/50 square wave unless your input is a clean sine wave or triangle wave.

?

They also work great at 10MHz.

?

There are other similar ICs with similar specs. I've used the SN74LVC1G17 and know it works extremely well.

?

It does require 3.3V max Vcc.

?

?

Good luck.

?
 

Looking at the data sheet it can be operated with a Vcc up to 5.5 VDC, max.? You may also need a series resistor at the output pin to reduce/eliminate ringing/overshoot/undershoot on the signal edges.
wn4isx
 

Unless the input signal goes below ground, to a negative voltage, this IC is very tolerant of noisy signals with no ringing or overshoot. A 1k or even a 10k resistor certainly won't hurt the performance.
?
 

开云体育

Thanks for the info ...

On 09/02/2025 11:01, wn4isx via groups.io wrote:

Unless the input signal goes below ground, to a negative voltage, this IC is very tolerant of noisy signals with no ringing or overshoot. A 1k or even a 10k resistor certainly won't hurt the performance.
?
 

开云体育


Hi, as per the advice given in this thread I am using an SN74LVC1G17 to drive a BNC cable but while the unloaded signal of 10Mhz is a nice square wave? with a bit of ringing when I look at the cable signal when connected to a counter by 2m of BNC cable there is a lot of distortion that sometimes makes the counter show twice or three times the frequency :-(? The input to the device is a nice clean square wave between 0v and 4.5v

I suspect the counter has a high impedance input so could this be due to a reflection ??? If so what sort of termination would help???

I want the signal to be used as a reference for the counter.

Thanks Dave


On 09/02/2025 11:01, wn4isx via groups.io wrote:

Unless the input signal goes below ground, to a negative voltage, this IC is very tolerant of noisy signals with no ringing or overshoot. A 1k or even a 10k resistor certainly won't hurt the performance.
?
 

On Sun, May 4, 2025 at 07:35 PM, David Slipper wrote:

Hi, as per the advice given in this thread I am using an SN74LVC1G17 to drive a BNC cable but while the unloaded signal of 10Mhz is a nice square wave? with a bit of ringing when I look at the cable signal when connected to a counter by 2m of BNC cable there is a lot of distortion that sometimes makes the counter show twice or three times the frequency :-(? The input to the device is a nice clean square wave between 0v and 4.5v

Welcome to the world of Signal Integrity and transmission lines and reflections.

I suspect the counter has a high impedance input so could this be due to a reflection ???

Absolutely.? It happens when the coax cable does not see a far-end termination impedance close to its own characteristic impedance.? Then you get a reflection, and ringing.? The faster the edges or longer the cable, the worse the ringing is.
?
Just in case, make sure that everything is grounded (to the common signal ground) everywhere.? Any cable without a ground connection would look bad.? I'm not talking about making connections to earth or anything like that.? Just make sure the cables and the buffer amp share the same grounds as one another and that they actually connect to it.
?

If so what sort of termination would help???

Most likely, the cable's impedance is either 50 or 75 ohms.? Can you read if the cable has an RG-number?? RG-58 is 50 ohms.? RG-59 is 75 ohms.? Most lab equipment uses 50 ohms.? ?Video tends to be 75 ohms.
?
Adding a 50 to 70 ohm resistor across the signal at the far end is probably "good enough" for either type of coax.
?
An alternative is to match the impedance at the driving end, known as "source termination".? To do that, add a resistor in series with the signal, directly at the driving IC's output pin.? You might use something smaller than 50 such as 30 - 40 ohms, counting on there being some 10 to 20 ohms of effective impedance internal to the driving part.? On the other hand, if the driving part happens to be several 74LVC1G17 gates in parallel, then their effective source impedance would be less, so the series resistor would need to be closer to 50 ohms to get a better match.
?
If you do use source termination, and add a termination at the far end too, then the signal's amplitude gets cut in half.? That is probably not an issue unless the counter requires a signal greater than 4 V.
?

I want the signal to be used as a reference for the counter.

Check to see if the counter's reference input does not already have a 50 ohm internal termination.? It is possible that its reference input is terminated (with 50 ohms) but its signal input is high-Z.
?
Andy
?
?
 

I agree, if you have connections with a BNC termination, the easiest way is to put a 'T' at the counter with a 50 ohm BNC load attached. Driving a cable with a square wave without termination will produce all kinds of ringing and distortion.

Dan Kahn
On Sunday, May 4, 2025 at 11:18:01 PM EDT, Andy via groups.io <ai.egrps@...> wrote:


On Sun, May 4, 2025 at 07:35 PM, David Slipper wrote:

Hi, as per the advice given in this thread I am using an SN74LVC1G17 to drive a BNC cable but while the unloaded signal of 10Mhz is a nice square wave? with a bit of ringing when I look at the cable signal when connected to a counter by 2m of BNC cable there is a lot of distortion that sometimes makes the counter show twice or three times the frequency :-(? The input to the device is a nice clean square wave between 0v and 4.5v

Welcome to the world of Signal Integrity and transmission lines and reflections.

I suspect the counter has a high impedance input so could this be due to a reflection ???

Absolutely.? It happens when the coax cable does not see a far-end termination impedance close to its own characteristic impedance.? Then you get a reflection, and ringing.? The faster the edges or longer the cable, the worse the ringing is.
?
Just in case, make sure that everything is grounded (to the common signal ground) everywhere.? Any cable without a ground connection would look bad.? I'm not talking about making connections to earth or anything like that.? Just make sure the cables and the buffer amp share the same grounds as one another and that they actually connect to it.
?

If so what sort of termination would help???

Most likely, the cable's impedance is either 50 or 75 ohms.? Can you read if the cable has an RG-number?? RG-58 is 50 ohms.? RG-59 is 75 ohms.? Most lab equipment uses 50 ohms.? ?Video tends to be 75 ohms.
?
Adding a 50 to 70 ohm resistor across the signal at the far end is probably "good enough" for either type of coax.
?
An alternative is to match the impedance at the driving end, known as "source termination".? To do that, add a resistor in series with the signal, directly at the driving IC's output pin.? You might use something smaller than 50 such as 30 - 40 ohms, counting on there being some 10 to 20 ohms of effective impedance internal to the driving part.? On the other hand, if the driving part happens to be several 74LVC1G17 gates in parallel, then their effective source impedance would be less, so the series resistor would need to be closer to 50 ohms to get a better match.
?
If you do use source termination, and add a termination at the far end too, then the signal's amplitude gets cut in half.? That is probably not an issue unless the counter requires a signal greater than 4 V.
?

I want the signal to be used as a reference for the counter.

Check to see if the counter's reference input does not already have a 50 ohm internal termination.? It is possible that its reference input is terminated (with 50 ohms) but its signal input is high-Z.
?
Andy
?
?
 

开云体育

Many thanks all that fixed it.

On 05/05/2025 14:17, Dan Kahn via groups.io wrote:

I agree, if you have connections with a BNC termination, the easiest way is to put a 'T' at the counter with a 50 ohm BNC load attached. Driving a cable with a square wave without termination will produce all kinds of ringing and distortion.

Dan Kahn
On Sunday, May 4, 2025 at 11:18:01 PM EDT, Andy via groups.io <ai.egrps@...> wrote:


On Sun, May 4, 2025 at 07:35 PM, David Slipper wrote:

Hi, as per the advice given in this thread I am using an SN74LVC1G17 to drive a BNC cable but while the unloaded signal of 10Mhz is a nice square wave? with a bit of ringing when I look at the cable signal when connected to a counter by 2m of BNC cable there is a lot of distortion that sometimes makes the counter show twice or three times the frequency :-(? The input to the device is a nice clean square wave between 0v and 4.5v

Welcome to the world of Signal Integrity and transmission lines and reflections.

I suspect the counter has a high impedance input so could this be due to a reflection ???

Absolutely.? It happens when the coax cable does not see a far-end termination impedance close to its own characteristic impedance.? Then you get a reflection, and ringing.? The faster the edges or longer the cable, the worse the ringing is.
?
Just in case, make sure that everything is grounded (to the common signal ground) everywhere.? Any cable without a ground connection would look bad.? I'm not talking about making connections to earth or anything like that.? Just make sure the cables and the buffer amp share the same grounds as one another and that they actually connect to it.
?

If so what sort of termination would help???

Most likely, the cable's impedance is either 50 or 75 ohms.? Can you read if the cable has an RG-number?? RG-58 is 50 ohms.? RG-59 is 75 ohms.? Most lab equipment uses 50 ohms.? ?Video tends to be 75 ohms.
?
Adding a 50 to 70 ohm resistor across the signal at the far end is probably "good enough" for either type of coax.
?
An alternative is to match the impedance at the driving end, known as "source termination".? To do that, add a resistor in series with the signal, directly at the driving IC's output pin.? You might use something smaller than 50 such as 30 - 40 ohms, counting on there being some 10 to 20 ohms of effective impedance internal to the driving part.? On the other hand, if the driving part happens to be several 74LVC1G17 gates in parallel, then their effective source impedance would be less, so the series resistor would need to be closer to 50 ohms to get a better match.
?
If you do use source termination, and add a termination at the far end too, then the signal's amplitude gets cut in half.? That is probably not an issue unless the counter requires a signal greater than 4 V.
?

I want the signal to be used as a reference for the counter.

Check to see if the counter's reference input does not already have a 50 ohm internal termination.? It is possible that its reference input is terminated (with 50 ohms) but its signal input is high-Z.
?
Andy
?
?

1 - 9 of 9