The 50ohm Termination: How, Why and When?

[Mechatrommer]

1) if i have 2 high speed opamp very close to each other, should i terminate the output of the 1st amp before feeding to the 2nd amp? like in the 1st picture below?
2) how long coax cable then termination is required? or i should always terminate it no matter how short it is?
3) if i have an opamp output directly feed to (1Mohm || 15pF) dso using BNC connector (no coax cable), should i terminate the opamp output (on the pcb) to avoid dso looks capacitive or displaying odd things on the screen? (2nd picture)
4) i read that by using termination, the coax cable will looks purely resistive (the capacitance of the cable and dso can be neglected, at least thats what i understand). can someone please explain in more elaborate (with math formula if necessary) way. why the capacitance become negligible?

5) more confusing combination, eg:
a) 50ohm series at source, 50ohm terminate at load end (1/2 attenuation)
b) 50ohm series and terminate at source, 50ohm terminate at load (1/3 attenuation)
c) 50ohm series and terminate at source, 50ohm series and terminate at load (1/5 attenuation)
d) 450ohm series at source, 50ohm terminate at load end (1/10 attenuation)
e) 950ohm series at source, 50ohm terminate at load end (1/20 attenuation)
f) etc...
why all the different combination and when best to use which? how come using 450ohm (not coax characteristic) also result in very high freq desirable respond?

too much to ask? well i wish i can post more, but this is all for now.

[olsenn]

You don't need to do that, no. The two two 50ohm resistors in the first photo will simply form a voltage divider to half the voltage being fed into the second op-amp. Since there is no gain on part of either op-amp, all the circuit does is provide a buffered output voltage that is at all times equal to one half of the input voltage.

It is common practice to add a small resistor (10 ohms or so) to the output of your final op-amp just to protect whatever device you're connecting it to in case of a voltage spike, but this isn't strickly required. One thing though that it typically done, but is missing from this schematic is to place a high quality film capacitor between your signal input and the input pin on the op-amp. You can add to large resistor to ground as well which adds further protection and noise immunity to the circuit since it acts as a high pass filter to get rid of any 60hz hum or whatever.

[Mechatrommer]

You don't need to do that, no

answer to question 1,2 and 3? the picture is just for example, voltage follower shown, but it can be gain amplifier, inverting, summing etc. the point is to the "mystery of termination". if its on the pcb and long trace, from what i understand, we need termination, but i dont have definitive answer to why/when is that necessary or not. if it is necessary, then i can only understand it is to avoid capacitive effect.

[ejeffrey]

1) if i have 2 high speed opamp very close to each other, should i terminate the output of the 1st amp before feeding to the 2nd amp? like in the 1st picture below?

Generally no.

2) how long coax cable then termination is required? or i should always terminate it no matter how short it is?

If the cable is around 1/10 of a wavelength for sure.

3) if i have an opamp output directly feed to (1Mohm || 15pF) dso using BNC connector (no coax cable), should i terminate the opamp output (on the pcb) to avoid dso looks capacitive or displaying odd things on the screen? (2nd picture)

Possibly.  This isn't actually acting as a terminator (the resistor doesn't have to match any particular impedance), it is just protecting the op-amp from seeing the capacitive load, which could lead to instability.

4) i read that by using termination, the coax cable will looks purely resistive (the capacitance of the cable and dso can be neglected, at least thats what i understand). can someone please explain in more elaborate (with math formula if necessary) way. why the capacitance become negligible?

The cable has both inductance and capacitance.  They work together to give the cable its impedance.  Explaining it with any degree of justice is probably beyond the scope of a forum, check out an electronics textbook or website.

[Harvs]

4) i read that by using termination, the coax cable will looks purely resistive (the capacitance of the cable and dso can be neglected, at least thats what i understand). can someone please explain in more elaborate (with math formula if necessary) way. why the capacitance become negligible?

The cable has both inductance and capacitance.  They work together to give the cable its impedance.  Explaining it with any degree of justice is probably beyond the scope of a forum, check out an electronics textbook or website.

+1  Requires quite a bit of paper to explain this one, and lots of pretty pics of waves traveling in the transmission line.

To say the line will look purely resistive is possibly a bit missleading.  What you will get if the source impedance, transmission line impedance and termination impedance are equal, is zero reflected energy.  Whether this matters to you or not typically depends on the length of the transmission line.  If the line is much shorter than the wavelength of what you're putting down the transmission line (like the <<1/10 ejeffrey stated), then the reflections will occur quick enough to likely not cause any disturbance to the circuit.  If however the line is long enough you'll end up with peaks and troughs in the middle of you're wave.

Do some reading on transmission line theory.  You don't necessarily need to go too heavily into the maths, it can be explained quite well by the diagrams and words most books have in them.

[Mechatrommer]

now, there are replies since my last post i'm not aware of. how come i didnt notified? duh! thanks people for answering, i didnt noticed.

3) if i have an opamp output directly feed to (1Mohm || 15pF) dso using BNC connector (no coax cable), should i terminate the opamp output (on the pcb) to avoid dso looks capacitive or displaying odd things on the screen? (2nd picture)

Possibly.  This isn't actually acting as a terminator (the resistor doesn't have to match any particular impedance), it is just protecting the op-amp from seeing the capacitive load, which could lead to instability.

i mean not the series resistor, but terminated resistor to ground. from the datasheet, i read the series resistor is a must if the opamp output is to see anykind of capacitive load. coax cable or dso capacitance in this case.
so... can i omit the resistor to ground? and feed the 50ohm impedance opamp output directly to dso? so i can get rid of 1/2 attenuation opamp output? will it reduce the measuring BW of the scope? noobs question here.

[free_electron]

To get the full bandwidth out of the scope you will need to terminate it at 50 ohm using its internal 50 ohm mode. You need to check what is the maximum vpp you can send into the connector when terminated at 50 ohm. Most scopes do not like more than 4vpp at that setting. You risk burning the terminator.
Read the manual of the scope. Its in there.

Now, you must tune the output impedance of your driver as well. The opamps datasheet will show you the output impedance for a given vpp. You will have to put in a series resistance from opamp output into the coax feed. Your drive opamp needs to be set for a gain of 2. You need to tune the series resistance so that, if you feed 1 vpp into the opamp , you get 1 vpp on the scope.
The series resistance and the 50 ohm terminator form a 1/2 divider. The opamp gain of 2 cancels that ut.

Most simple opamps have an output impedance in the order of a few tens of ohms. That ths3062 is much lower. This value should be listed in the datasheet.

If the ths doesn't pull it switch to an ad811 or ad815. Those beasts can definately do it.
I seem to recall that the ths30 series can drive 50 ohm termination without problems. ( provided you have the right series resistance... . Driving 50 ohm does not mean ' hook 50 ohm to ground'. It means from the output pun to ground you need tot have 100 ohm !
Driving 50 ohm is NOT the same as driving a 50 ohm TERMINATION,

I have the impression that you are in way over your head. You are trying to do something that is quite complex and lack a solid understanding of the things you are dealing with. Making an analog system like you are trying to do is not a matter of taking some junk parts and slapping them together. Also , the 'simulator' aproach is wrong. The problem with simulators is that people blindly believe what comes out of them... For kost of the sims the models are wrong. There are too many factors that are not modelled in simple sims like tina amd ltspice. A really good sim like ADS or Eldo costs a lot of money and it takes years to learn how to use it. And you need accurate models.
Where o work we characterise devices with all kinds of information including parasitics , interconnnects and more. The output of the sim is very accurate to the real finished product. But it takes several labs and 20 people to build the sim models of the parts we use.

Those cheap sims use simplified models. Like in your other question : good luck getting the input bias current info out of the sim. Most opamp models dont specify that. Ltspice has it becaus linear went through the trouble of putting it in , but that info is encrypted. Try the same in multisim and it wont work. Heck , multisim is so stupid that, if you take an opamp , put a gain of 1000 and inject 1 volt it dutyfully says the output is 1000 volt ... With a 5 volt power supply ? Riiiiight ....

[Mechatrommer]

To get the full bandwidth out of the scope you will need to terminate it at 50 ohm using its internal 50 ohm mode

my scope doesnt have internal 50ohm.

Most simple opamps have an output impedance in the order of a few tens of ohms. That ths3062 is much lower

warning: ths3062 is not meant for repetetive signal of more than 900V/us. FWIW.

ths30 series can drive 50 ohm termination without problems

most opamps can, even if you short it to ground direct. the problem is with capacitive load.

You are trying to do something that is quite complex and lack a solid understanding of the things you are dealing with

me know not de transmissioun line.

Making an analog system like you are trying to do is not a matter of taking some junk parts and slapping them together...
Also , the 'simulator' aproach is wrong...

suggestion?

[ejeffrey]

most opamps can, even if you short it to ground direct. the problem is with capacitive load.

No, most opamps cannot drive 50 ohm loads, at least not with anything like the full voltage swing.

[free_electron]

To get the full bandwidth out of the scope you will need to terminate it at 50 ohm using its internal 50 ohm mode

my scope doesnt have internal 50ohm.

let me guess. Wing wang pong scope ? It's not a scope if it doesn't have an input switchable between 1meg and 50 ohm...

You can get around it by putting a 50 ohm resistor yourself as close as possible to the bnc. Plug

ths30 series can drive 50 ohm termination without problems

most opamps can, even if you short it to ground direct. [/quote]
you are kidding right ? That is not driving. That is shorting. Driving a load means being able to do full vpp across that particular load. If you short the output you have no usable output !

me know not de transmissioun line.

you dont need a transmission line ! You are dealing with 10MHz sinewaves .. Come back if you have edge rates in the nanosecnod range... Then we'll talk transmission lines.


Also , the 'simulator' aproach is wrong...[/quote]suggestion?
[/quote] soldering iron ?

I dont know of dave reads this one but here's an idea for a new t-shirt

Front shows an old school soldering iron under 45 degree angle. Text above says ' i have a soldering iron ' ( white text ) text below says 'AND I'M NOT AFRAID TO USE IT' (red text )

[Time]

wing pang pong scope?  lol, how delightfully offensive.

Lots of entry level/hobbyist scopes don't have 50 ohm options.  i.e. the Tektronix TDS2000 series which goes for about $2k these days I believe

[free_electron]

Lots of entry level/hobbyist scopes don't have 50 ohm options.  i.e. the Tektronix TDS2000 series which goes for about $2k these days I believe

simply install a 50 ohm fedd through terminator at the end  of the bnc cable. problem solved. By tre way, that tektronix is a toy. I can get used scopes for half that price that do 1GHz...

[Mechatrommer]

yeah! that a hell of offensive statement! . my original question is... a fast opamp... have a series 50ohm, that is a must all right... now the question is... the 50ohm to gnd after the series 50ohm.. do i need it? because my scope doesnt have internal 50ohm. i dont need "external 50ohm inline", since i'm intending to connect the pcb direct to scope, "external inline terminator" is pointless imho, since i can do it in pcb... now... for that short distance "direct pcb to scope... do i need "50ohm to ground" in pcb? (acting as "internal 50ohm inline" TO GROUND?, and of course after the "a must series 50ohm just after the opamp output")? if not... that series 50ohm combined with 15pf scope internal, will act as low pass right? duh! is my question confusing enough. i guess so since i'm noob, i may come out with nonsensical question.

soldering iron ?

and then what?... taking some junk parts and slapping them together?

ps: "["quote"]" must be terminated with "["/quote"]" (excluding double apostrophe) no! thats not an offense! thats a... friendly reminder
need several cycle for editing

[vk6zgo]

To get the full bandwidth out of the scope you will need to terminate it at 50 ohm using its internal 50 ohm mode

my scope doesnt have internal 50ohm.

let me guess. Wing wang pong scope ? It's not a scope if it doesn't have an input switchable between 1meg and 50 ohm...

In over 40 years in Electronics,mostly in TV & Broadcast Transmission,I've never used an Oscilloscope with a switchable 50 Ohm termination.
Out in the real world,you quite commonly work with 75 Ohm systems as well,so you will need through terminations for that impedance,why not for 50 Ohms as well?
 

You can get around it by putting a 50 ohm resistor yourself as close as possible to the bnc. Plug

ths30 series can drive 50 ohm termination without problems

most opamps can, even if you short it to ground direct.

you are kidding right ? That is not driving. That is shorting. Driving a load means being able to do full vpp across that particular load. If you short the output you have no usable output !

me know not de transmissioun line.

you dont need a transmission line ! You are dealing with 10MHz sinewaves .. Come back if you have edge rates in the nanosecnod range... Then we'll talk transmission lines.

Gee! I'll have to tell the blokes down at the MF AM Broadcast sites,that they don't need a transmission line!


Also , the 'simulator' aproach is wrong...[/quote]suggestion?
[/quote] soldering iron ?

I dont know of dave reads this one but here's an idea for a new t-shirt

Front shows an old school soldering iron under 45 degree angle. Text above says ' i have a soldering iron ' ( white text ) text below says 'AND I'M NOT AFRAID TO USE IT' (red text )
[/quote]

OK,Shafri gets a bit overextended sometimes,but he's a nice bloke,& he is learning by doing!

[Mechatrommer]

you are kidding right ?

yeah i guess so . which one is better? a stable 0V, or an oscillation?

OK,Shafri gets a bit overextended sometimes,but he's a nice bloke,& he is learning by doing!

from time to time, i kept thinking that, i need that "a piece of paper with golden mark on it" to be eligible to participate going back to school is not an option for me anymore. so what hope do i have? being a lumberjack is another option

[free_electron]

Pieces of paper with golden stars are overrated. Waaaaaay overrated..

@vk6zgo : any decent scope has the capabilit to terminate its input at 50 ohm. The reason is to maintain signal integrity all the way , through the connector to the digitizer.
Even the old tektronixes with vacuum tubes already had that.

As for you transmission line. You are dealing with a modulated signal at high power.
It is not frequency that matters but edge rate. Like i said a 10 MHz sine wave doesn't need transmission lines ( we are talkin stripline or microstrip on a pcb over a distance of an inch here between two opamps) . We'll talk microstrips when your edge rates are much higher. No need to overengineer and seek nonexistent problems. Problems are very well capable of finding you all by themselves. I've done enough boards for signal integrity in my life and gone through enough EMC quals to know what matters and when it matters.

Emc is not the same as signal integrity. Neither are power losses. The question is what are you shooting for.

[vk6zgo]

Pieces of paper with golden stars are overrated. Waaaaaay overrated..

@vk6zgo : any decent scope has the capabilit to terminate its input at 50 ohm. The reason is to maintain signal integrity all the way , through the connector to the digitizer.
Even the old tektronixes with vacuum tubes already had that.

As for you transmission line. You are dealing with a modulated signal at high power.
It is not frequency that matters but edge rate. Like i said a 10 MHz sine wave doesn't need transmission lines ( we are talkin stripline or microstrip on a pcb over a distance of an inch here between two opamps) . We'll talk microstrips when your edge rates are much higher. No need to overengineer and seek nonexistent problems. Problems are very well capable of finding you all by themselves. I've done enough boards for signal integrity in my life and gone through enough EMC quals to know what matters and when it matters.

Emc is not the same as signal integrity. Neither are power losses. The question is what are you shooting for.

My friend,I've used more Oscilloscopes,(mainly HP & Tektronix), than you've had good breakfasts! 
None of them had 50 Ohm terminations included.
Yes,all of them could have an external termination added,but none of them had it as part of the normal input circuitry.
By the way,what is this digitiser?---Not in old vacuum tube Tektronix 'scopes!

"Edge rate",unless I am very much mistaken is a $5 word meaning "rise time".
The minimum rise time of a signal is intimately related to the highest frequency content of the signal.
So how can you say it is not frequency that matters?

OK,granted,for short distances you can get away with things,& as you said,at 10MHz,an inch or so doesn't matter!
As far as modulation & high power,that doesn't really matter,what does matter,as you commented,is the length of the transmission line.

[jahonen]

I think that is easier to avoid that frequency thinking altogether when dealing with digital signals and just think about how much physical space will the signal edge take on a transmission line. This works whether we have widely separated fast edges or very high repetition rate signal. It works even with single transitions where the frequency and harmonics thinking breaks down.

So for example if we have a 1 ns edge and put that onto a stripline structure where signal velocity is approximately about half of that of free space, 1.5e8 m/s. So 1 ns edge will take 1 ns * 1.5e8 m/s = 15 cm length of transmission line (yes, you can put several of those edges simultaneously on the line if the line is long enough and/or edge spacing is small enough!). And now if edge length starts to be 10%…20% of total trace length, then one should start worrying about reflections.

For RF things are a bit different. Although frequencies are lower for that 10 MHz system, but the system is more sensitive to reflections. Digital system may well operate happily with just 10 dB return loss (or SWR of 1.74 ~1.9 or |S11| = -10 dB) but that is probably too little for RF power transmission. Imagine a 1 kW transmitter which gets 900W 100W return power due to mismatch.

Regards,
Janne

[free_electron]

My friend,I've used more Oscilloscopes,(mainly HP & Tektronix), than you've had good breakfasts! 

wanna take a bet on that ? besides its not how many , its what you do with them 

just looking around here there's at least 7 sitting around and i have used all of them in the last couple of months , and they're all different.
The tek7405 only has 50 ohm input... same for the lecroy 7200 . 50 ohm only input. Agilent 90000 has APC7 connectors ( called sexless connectors since the are surface mating. there is no male or female. twist connector on using a torque wrench... its that precise. then again... it does 20 Ghz bandwidth ...

The scopes in my hobbylab at home ( i got a few there too :p ) : all switchable from 1meg to 50 ohm and back, includes my 20 years old agilent 54645d.
My first scope i bought was a german made hameg 604. full analog. 50 ohm switchabel inputs. followed by a philips pm3360. guess what : 50 ohm switchable inputs. i haven't even seen or used a scope that does not have a 50 ohm termination option.

By the way,what is this digitiser?

a digitzer is a circuit found in any non-stone age dinosaur technology scope 

"Edge rate",unless I am very much mistaken is a $5 word meaning "rise time".

some peole call it that. depends on britisch english vs american english. tomatoooo tomato . poteeeeto potaaato.

The minimum rise time of a signal is intimately related to the highest frequency content of the signal.

That is a correct definition. Unfortunatley it is commonly quoted as
'the rise time of a signal is related to the frequency' ( note the omission of minium and maximum ) and then it becomes a blatant lie !

[Mechatrommer]

so... with all this very "informative" posts (i mean it), the actual point of 50ohm terminator is... to avoid reflection. right? all i have to worry is that... if the trace length is more or less than 1/10 of wavelength? so say, if assuming the signal from my opamp to the next passive/active element inside the scope is less than 1/10 wavelength, then... i dont need the 50ohm teminator within my pcb (which will be connected directly to the scope, 1Mohm 15pF impedance) right? only when the the trace/coax length is more than 1/10 wl, then only then i will need to terminate the line, right?

[vk6zgo]

My friend,I've used more Oscilloscopes,(mainly HP & Tektronix), than you've had good breakfasts! 

wanna take a bet on that ? besides its not how many , its what you do with them 

just looking around here there's at least 7 sitting around and i have used all of them in the last couple of months , and they're all different.
The tek7405 only has 50 ohm input... same for the lecroy 7200 . 50 ohm only input. Agilent 90000 has APC7 connectors ( called sexless connectors since the are surface mating. there is no male or female. twist connector on using a torque wrench... its that precise. then again... it does 20 Ghz bandwidth ...

Sounds a bit like the German Dezifix & "Spinner" connections from 20-30years ago,but a lot smaller

The scopes in my hobbylab at home ( i got a few there too :p ) : all switchable from 1meg to 50 ohm and back, includes my 20 years old agilent 54645d.
My first scope i bought was a german made hameg 604. full analog. 50 ohm switchabel inputs. followed by a philips pm3360. guess what : 50 ohm switchable inputs. i haven't even seen or used a scope that does not have a 50 ohm termination option.


The fact remains that there are many thousands of Oscilloscopes  still in service throughout the world which do not have 50 Ohm termination,either fixed or switchable.

By the way,what is this digitiser?

a digitzer is a circuit found in any non-stone age dinosaur technology scope 
Ohh,you mean an ADC!  
But you mentioned old vacuum tube 'scopes,so my comment was a bit "tongue in cheek"

"Edge rate",unless I am very much mistaken is a $5 word meaning "rise time".

some peole call it that. depends on britisch english vs american english. tomatoooo tomato . poteeeeto potaaato.

Funny,all the American publications I've read use "rise time",except Application Notes where they seem to have a language of their own!

The minimum rise time of a signal is intimately related to the highest frequency content of the signal.

That is a correct definition. Unfortunatley it is commonly quoted as
'the rise time of a signal is related to the frequency' ( note the omission of minium and maximum ) and then it becomes a blatant lie !

No,it's still correct,but it is useless without the other information.

 You made a sweeping generalisation:
"It's not a scope if it doesn't have an input switchable between 1meg and 50 ohm..."
& I disagreed,so it does no good to enumerate the 'scopes at your work,etc,your original comment is incorrect.
Ask a few people working in general Electronics to show you their Oscilloscopes.

[steve_w]

I'll show you mine if you show me your's ;-)

I don't really think we are really helping Mechatrommer, he just wants to know about electronics why not stick to answering the questions he asks?

BTW Mech good questions I am learning from the answers as well. 

[jahonen]

The scopes in my hobbylab at home ( i got a few there too :p ) : all switchable from 1meg to 50 ohm and back, includes my 20 years old agilent 54645d.

54645D at my work does not have 50 ohm inputs, and neither Agilent spec sheet http://cp.literature.agilent.com/litweb/pdf/5968-2610EN.pdf mentions one?

Regards,
Janne

[vk6zgo]

I'll show you mine if you show me your's ;-)

I don't really think we are really helping Mechatrommer, he just wants to know about electronics why not stick to answering the questions he asks?

BTW Mech good questions I am learning from the answers as well.

Steve,I think you are right!

free_electron & I are carrying on a similar argument two threads down,maybe we should take it there.

What do you think about that idea,free_electron?

Or will we just take a break?

[Tube_Dude]

My first scope i bought was a german made hameg 604. full analog. 50 ohm switchabel inputs....

Hi

The Hameg 604 doesn't have switchable 50 Ohms inputs.