120ohm imbedance line to 50ohm oscilloscope

[veeti]

Hello

I’m starting to debug/reverse engineering CAN bus,
and i’m not sure how to deal with line imbedance.

Can I just make custom cable with BNC connector which is ”shunted” with 120ohm resistor and take test lead from there?

I allso have 2pcs MCP2515 connected to arduino and raspberry pi - and have been able to make data moving. But i would like to see signals allso in a scope screen.

br: veeti

[jonpaul]

balanced differential 120 Ohm TP line.
Ideal is a differential probe

As drawing the BNC will short one side to ground

Use two 10X probes, set-up scope  for differential CHA-CHB

Unawares if Chinese scopes have such a function (or INV B) as do most Tektronix scopes or old analog 

Alternative Use a pulse transformer or differential line receiver

See Tektronix Circuits Concept book Oscilloscope Probe Circuits
Also théorie and practical differential balanced transmission

Bon courage

Jon

[ataradov]

Don't overthink it.

To check signal integrity just use one 10x probe on each side to see that signals are symmetrical. No need for math functions.

To decode just pick any side and assume that the other one is complimentary.

And decoding is far better with a logic analyzer anyway.

[pcprogrammer]

The original poster shows a Siglent SDS1102X-E in his picture which as far as I can tell, has selectable input impedance of 50Ohm and 1MOhm.

I'm no expert here, but think the 1MOhm option with a normal 10x probe should be used instead of 50Ohm option with a normal 10x probe?

And for the 50Ohm setting an active probe is needed?

[ataradov]

Yes, you need 1 M input and 10x probe.

[veeti]

Thanks for fast response.

But if I use high impedance and probe - why there wouldn’t be reflection back to the line?

Standard says max 30cm line without termination, and now i would have 1Megaohm termination with 1meter cable.

[ataradov]

Your line is terminated by the actual termination resistors on the bus. The bus would not even notice your 1 MOhm  "load" even with 1x probe.

[exmadscientist]

Don't overthink it.

People grossly overestimate how much this stuff matters.

Yes, when conductors are long compared to the bit rate / edge rate you want to do it right. But how often are you doing gigabit links or kilometer cables? And if you're using a scope, if you get it wrong, you can see it on the scope, it's not like it'll be waiting to bite you later. (Though you can get bit if you forget that having the probe there changes things a little.)

This also shows up in controlled impedance on PCBs. You just don't need it for a lot of things, "best effort" (take your stackup, calculate your widths, use 'em, and forget about it) is good enough up through even USB 2.0 High Speed, as long as the rest of the environment is gentle enough. Which it usually is.

[radiolistener]

Can I just make custom cable with BNC connector which is ”shunted” with 120ohm resistor and take test lead from there?

yes, you can use 120Ω termination resistor, but in such case you're needs to use high impedance input (1 MΩ).
If you use 50 Ω input on 120 Ω termiantor, the actual load will be 1/(1/50+1/120) = 35.29 Ω.

If you're needs to use it with 50 Ω input, then you're needs to measure it through 120-50= 70 Ω resistor (connected in series with 50 Ω oscilloscope input).

Another way is to use attenuator, see attachment for 10 dB and 20 dB attenuators. The 20 dB one can be used with 1:10 settings on oscilloscope to see actual voltage levels.

[ataradov]

But that's not how CAN works. CAN bus is terminated by 120 Ohm resistors on two ends. That's all. Your scope should not add more "termination".

You are just making it complicated for no reason at all. Just stick a scope on the bus and see that it all works and you just see what happens on the bus.

[radiolistener]

But if I use high impedance and probe - why there wouldn’t be reflection back to the line?

if you use high impedance probe, your CAN line needs to be terminated. It is already terminated if it's connected to another receiver with 120 Ω input. If you're needs to measure line which is not connected to anything, then you're needs to put 120 Ω termination resistor with high impedance input or use oscilloscope with proper input impedance.

If you're want to measure 120 Ω line which is already terminated by another receiver, then oscilloscope with 50 Ω input is not suitable for you, because when you tap your 50 Ω probe to already terminated 120 Ω line, the point where you tap the probe will have 35 Ω impedance and as result it leads to wave reflection.

Also note, when you use high impedance probe, it has bandwidth limitation. So, if you want to see more high frequency details, it's better to use 50 Ω input with proper attenuator. But it will works when your line is not connected to anything else, this way will not work if your line is already terminated with another receiver. When the line is already terminated, the only way is to use high impedance probe.

In simple words, if you're needs to catch transmission between devices, then just use high impedance probe. If you're want to measure parameters of device output which is not connected to anything else, then you're needs to use proper terminator resistor with high impedance oscilloscope input or use proper attenuator with 50 Ω oscilloscope input.

The main idea is that endpoint of line needs to have proper 120 Ω resistance (for 120 Ω line) which include resistance of the oscilloscope probe. If oscilloscope probe has too low resistance you can add resistor in series to get actual 120 Ω. If oscilloscope probe has too high resistance, you're needs to add resistor in parallel to get actual 120 Ω. This is how it works.

Also note that resistor should not be inductive. Some kind of resistors can have too high inductance and can't be used. Use RF grade resistors or at least film type with no winding for such purposes. SMD resistors bellow 200 Ω usually works ok up to 50-100 MHz and even more.

[Wallace Gasiewicz]

If you wish to terminate your 120 ohm line into the 50 ohm input of your  scope, you can add a 70 ohm resistor in SERIES with the input. (not to ground). I think radiolistener stated this.
You could make up some sort of connector with the 70 ohm resistor in it. As  pointed out, you need a resistor that is not too reactive.
I think you could put a regular axial lead resistor in the center pin of a BNC connector and jam it into the center coax wire and solder it. It might just fit into the BNC, if you trim the center just right.....

If you wish to terminate the 120 ohm output into the high impedance scope input, use 120 ohms to ground.

However if your 120 ohm signal is already terminated into another device, use 10X probes with the high impedance scope input. Lots easier.

[Manul]

But if I use high impedance and probe - why there wouldn’t be reflection back to the line?

Standard says max 30cm line without termination, and now i would have 1Megaohm termination with 1meter cable.

I think what is confusing you (and seems to be not explained by others) is that scope probe is NOT your typical coax cable. It has special construction and it is made to be deliberately a lossy transmission line (center wire is highly resistive). This transmission line is carefully crafted to have wide and flat frequency response and minimal reflection impact.

This is of course the reason why we can use it to probe things quite universally. If your scope probe would me made from standard 50 Ohms cable and matching termination at the scope end, it would be useless for most of everyday circuit probing, causing loading. And if unterminated at the scope end, it would also be useless because of all reflections, ringing and garbage frequency response.

[David Hess]

But if I use high impedance and probe - why there wouldn’t be reflection back to the line?

The line is already terminated at each end in the circuit.  A high impedance x10 probe can be attached anywhere because it will only present like 10 picofarads of capacitance.

[jonpaul]

Bonjour: The CAN bus is similar to RS-232, RS-485, AES/EBU, with a differentledrive to a twisted pair lik Ethernet cable UTP or TP.

Most of the TP cables are designed for Zo 100..110..120 Ohms Differentl, ie the ter reisitor is line to line NOT line to return.

Probing the TX or RX sides is easy with 10X probes into a 1M scope.

At these rates a 50 Ohm scope and cax is NEVER required.

Two probels can be used diffreretially if the scipe has CHA - CHB function.

But one seldom see a hug imbalalnce so just one probe, moved from L1 ro L2 is enough to see the leves and edges.

As the RS-485 RX IC has some hysterysis, noise is somewhat rejected.


See the CAN bus spec, RS-485 TX and RX IC specs and excellent app notes of Belden Cable by Steve Lampen for balanced  TP cable and transmission. .

Vehicle cable runs are typically 1m to 5m but theenvironment is VERY noise with EMI and high amplited transients.

Finally i have written AES , SMPTE and NAB papers since 1995, on the topics of digital audio transmission and cables and transformers, the baud rate is 6.144 MB to 24.576 MB.

Ciao!

Jon