coax cable

[IvoS]

Where I can buy the same coax cable that Rigol uses for cheap DS1052 oscilloscope probes?

[alm]

I would guess cheap/dead probes would be your only source. I've never seen the special low-capacitance lossy coax used for scope probes for sale.

[Short Circuit]

Probably doesn't need to be coax at all since source and termination impedances are practically infinite.
Just looking for 100% shielding and low capacitance should be sufficient, maybe something like HQ phono cable.

[alm]

Are you talking about scope probes? Or is there some alternate application I'm missing? The impedances involved in scope probes are definitely not infinite for any practical purpose at frequencies above DC (above 100 kHz or so), and definitely not resistive. Hence all the ringing when introducing just a small amount of inductance in the ground connection. Just look at the input impedance graph and see where the capacitive part starts to dominate the resistive part. Plus all the fun reflections you would get even if the input impedance was infinite (100% reflected).

[IvoS]

Well, I bought 50ohm BNC to BNC terminated cable (1m long), I am not sure if it was for TV signal or other kind of applications but it was declared 50ohm cable. When using this cable with my low THD 10kHz sinus generator (straight from gen to scope) I noticed significant oscillation, around 82MHz imposed on 10kHz, 610mV p-p. I blame the oscillator, but when I probe the generator output with original Rigol probe, the oscillation disappeared, so the cable did induce some oscillation. I need some low capacitance coax cable (I guess) for this straight connection. I would like to ask you guys one  question, (because I am a dummy). Scope has input impedance of 1Mohm. If I have a source (generator etc.) with low output impedance, can I go to scope directly without probes? In what situation do you use 50ohm termination? Thanks.

[notsob]

Understanding scope probes

Watch this

[IvoS]

Is RG59 coax cable a good cable for instrument connection?

[NiHaoMike]

Is RG59 coax cable a good cable for instrument connection?

Yes, as long as it is properly terminated. Expect up to a few hundred MHz of bandwidth before the losses become significant.

[PA0PBZ]

Is RG59 coax cable a good cable for instrument connection?

Most of the low impedance instrument connections are 50 ohm, RG59 is 75 ohm. Not a logical choice I'd say, take RG58, that is 50 ohm.

[T4P]

There is a good reason for that,
http://groups.yahoo.com/group/PCR1000/message/3130

[alm]

Using RG-59 with a 50 ohm system would indeed create an impedance mismatch. What would the VSWR be at 10 kHz, though? The only reason why you'd need a coaxial cable at all would be shielding, not impedance control. To understand the difference between RG-58 and RG-59, and when and how to terminate, it would be a good idea to read up on transmission lines. With the notion that a scope probe is not a lossless transmission line .

Sounds to me like the sinus is generator is oscillating due to the capacitive load that the cable represents. If it can't drive a few meters of RG-58, then I would add a cable driver buffer at the output. Most audio devices will also represent a capacitive load. Terminating it into the characteristic impedance of the cable may help, since it makes the cable a resistive 50 ohm load, although DC loading will be high.

[kg4arn]

The coax used in the typical 10X passive oscilloscope probe has resistance wire for the center conductor.
This is to dampen the reflections (ringing).
I have never been able to find the stuff for sale.  Anybody seen it for sale??

http://www.slack.com/TE/TekConcepts/TekProbeCircuits.pdf

http://www.dfad.com.au/links/THE%20SECRET%20WORLD%20OF%20PROBES%20OCt09.pdf

[PA0PBZ]

Using RG-59 with a 50 ohm system would indeed create an impedance mismatch. What would the VSWR be at 10 kHz, though? The only reason why you'd need a coaxial cable at all would be shielding, not impedance control. To understand the difference between RG-58 and RG-59, and when and how to terminate, it would be a good idea to read up on transmission lines. With the notion that a scope probe is not a lossless transmission line .

I was not suggesting that you could not use RG59 for the scope, but there is a lot of other stuff (RF Generators, SA's and such) that would need 50 ohm coax, so it just seems a bit weird to get RG59 for the lab.

[alm]

I concur that RG-59 is an odd choice for the lab unless you happen to have it available. My choice for generic coax cable would also be RG-58. RG-174 if diameter/bending radius is an issue, and higher quality cables for long cables or higher frequencies. RG-58 is fine for most frequencies where you would still use BNC connectors, however. All this talk about losses at microwave frequencies seems odd in the context of a 10 kHz low distortion sinus generator.

The coax used in the typical 10X passive oscilloscope probe has resistance wire for the center conductor.
This is to dampen the reflections (ringing).

Most systems involving sufficiently high frequencies that reflections become an issue (wavelength is shorter than ten times the electrical length of the cable) just use low-impedance terminated transmission lines (for example RG-58 terminated into 50 ohms). These behave much better than scope probes. They tend to be quite flat across a wide bandwidth (easily into the many GHz with the right cable and connector), have better signal integrity, have a very constant impedance over this bandwidth, and are easy to simulate.

The impedance of a scope probe will vary by about six orders of magnitude across its bandwidth of at best a few hundred MHz. Note that aberrations in transient response can usually be in the order of 1-5%, and up to -3 dB in the frequency domain. An oscilloscope is not a low-distortion instrument. You'll also note in the Tektronix Circuit concepts book that the termination circuit for high-bandwidth probes gets quite complex to compensate for all of the aberrations.

It's only at high impedances that transmission lines get tricky. Try doing the math for the dimensions of a cable with a 1 Mohm characteristic impedance . Hence the use of lossy coax and other tricks to keep the effects of impedance mismatches down.

[kg4arn]

Most systems involving sufficiently high frequencies that reflections become an issue (wavelength is shorter than ten times the electrical length of the cable) just use low-impedance terminated transmission lines (for example RG-58 terminated into 50 ohms).
....
It's only at high impedances that transmission lines get tricky. Try doing the math for the dimensions of a cable with a 1 Mohm characteristic impedance . Hence the use of lossy coax and other tricks to keep the effects of impedance mismatches down.

I agree.  My post was meant for the OP. 

[IvoS]

Using RG-59 with a 50 ohm system would indeed create an impedance mismatch. What would the VSWR be at 10 kHz, though? The only reason why you'd need a coaxial cable at all would be shielding, not impedance control. To understand the difference between RG-58 and RG-59, and when and how to terminate, it would be a good idea to read up on transmission lines. With the notion that a scope probe is not a lossless transmission line .

Sounds to me like the sinus is generator is oscillating due to the capacitive load that the cable represents. If it can't drive a few meters of RG-58, then I would add a cable driver buffer at the output. Most audio devices will also represent a capacitive load. Terminating it into the characteristic impedance of the cable may help, since it makes the cable a resistive 50 ohm load, although DC loading will be high.

Thank you for your response. I just simply can not add a buffer because it would null and void the whole circuit application. My low sine gen has THD of 1ppm, it's built based on AN67 (page 62) from Linear technologies. I will buy a bulk of RG58 and bunch of BNC connectors and crimp them myself. I just need some reliable BNC LAB interconnect in different lenghts. RG58 should do it. The cable I was using before was some cheap BNC cable and I think it's even 75ohm cable.

[Mechatrommer]

i suggest get RG174 instead of RG58. esp if you have a tight bench. RG58 is big cable about 5mm OD, not really comfortable to work with. my 2cnts, just in case if you havent noticed.

[T4P]

But RG174 is known as rubbish coax isn't it? How about RG316? Those are used for NIMs (Nuclear Instrumentation Modules ... scary)
Scratch that. Nope, been looking on the wrong datasheet. It's much better than RG174

[Mechatrommer]

But RG174 is known as rubbish coax isn't it?

how do you define rubbish for coax cable? i have here 2 type of 50ohm coax, one is smaller RG174, one is 3x25 cable, RG58 by the look of the diameter. please tell me what method to prove RG174 is rubbish than the bigger coax.

[IvoS]

Thanks guys. Yes, I was also looking at the smaller diameter cable, and I can imagine it's sometimes better to handle and manipulate the thinner cable on the bench. The differences I see between RG58 and RG174 is that RG58 has less attenuation on higher frequencies (higher bandwidth) and more tight braided shielding. RG58 can be 98% braided shielding vs. RG174 is like 90% shielding.

[Mechatrommer]

by "reading" it, its like a bit "rubbishier" but so far i cant see any significant loss in RG174 for my use, but well i guess your milage will vary, for eg if you are working with -120dB signal or up to GHz range, then the loss spec might be of your concern. or maybe if you need very accurate THD measurement then you should be aiming for the best, for example semi rigid or rigid coax or even proper "waveguide"

[T4P]

I don't mean it's really unusable but if you're migrating above 450MHz (which powerful chips run at) it is a problem when you're already dealing with small signals and only worsened by the RG174 loss because 74db/100m loss is VERY significant and only worse if you're dealing with RF.

I don't say it's bad for the OP but it's okay to use RG174 as a beginner, it's fine. I use RG174 too because that's all i could find at the brick and mortar shops, i'll use RG58 or RG-400 or RG-142 and also some RG-316 for short stuff

[T4P]

That's why i said it's okay for him if he's not migrating upwards to RF
Less loss is always good because you will be further away from the noise floor and don't need to resort to funny tricks

[Mechatrommer]

what was it? sorry.

[jahonen]

Yes, RG-174 is quite lossy compared to something like semi-rigid coax types. It depends if loss is acceptable or not. For a precision measurement, neither RG-58, RG-174 or RG-316 is good enough. Here is one measurement I made on a self-made SMA-SMA RG-174 cable, of about 1 meter of length.

http://koti.mbnet.fi/jahonen/Electronics/Measurements/Spectrum/CoaxConn_003.PNG (I hate that my R&S FSV7 does double the screenshot pixels, AFAIK a feature that can't be switched off. Otherwise, a very nice machine.)

Reference is a Huber&Suhner SucoTest semi-rigid cable. For normal coax types, it seems that one must basically make the cable thicker to reduce losses. That might be due that AC resistance is quite high due to the skin/proximity effect and only way to reduce that is to increase the surface area of conductors. But something like 10 mm thick coax cable is pretty clumsy for casual lab use.

Regards,
Janne

[Mechatrommer]

yeap! if you are like Janne, you'll be in 3 conditions:
1) RG174 will not be enough for you. read 7GHz app and up... and
2) you've long enough playing around with coax and hence...
3) you wont ask noob questions like i am here and there.

as i mentioned in my deleted post after reply #21, some people add even further attenuation (loss) at the coax tip to gain BW in general purpose probing and be happy. so it'll depend on what you are doing. messing with big or semi rigid coax on a tight bench is certainly not for me for now. YMMV.

[IvoS]

Yes, RG-174 is quite lossy compared to something like semi-rigid coax types. It depends if loss is acceptable or not. For a precision measurement, neither RG-58, RG-174 or RG-316 is good enough. Here is one measurement I made on a self-made SMA-SMA RG-174 cable, of about 1 meter of length.

http://koti.mbnet.fi/jahonen/Electronics/Measurements/Spectrum/CoaxConn_003.PNG (I hate that my R&S FSV7 does double the screenshot pixels, AFAIK a feature that can't be switched off. Otherwise, a very nice machine.)

Reference is a Huber&Suhner SucoTest semi-rigid cable. For normal coax types, it seems that one must basically make the cable thicker to reduce losses. That might be due that AC resistance is quite high due to the skin/proximity effect and only way to reduce that is to increase the surface area of conductors. But something like 10 mm thick coax cable is pretty clumsy for casual lab use.

Regards,
Janne

Oh my goodness! That's in GHz region! I don't have to worry about that at all  , I'll be happy if I get some day to MHz region. Now seriously. I am after THD with low frequency signals, that's all I want. Here is the THD spectrum graph where I could get so far with my oscillator. The measuring method (sound card) is my limit point of measuring. This sinus oscillator should be much better than that!