Higher speed coax test cable

[king.oslo]

Hello there,

I recently bought a second hand Agilent 33120A. Now I want to test all the functions.

I connect the unit to my scope through a short,cheap coax test cable (bnc to bnc) from china. I measured its capacitance to 40pF using Agilent U1733C.

At 15mhz squarewave, the scope is seeing a 15mhz sinewave. I suspect this is because of the capacitance of the wire.

What kind of test leads do I need to actually be able to use fequencies this high?

Thank you for your time.

Kind regards,
Marius

UPDATE: The scope is the DS1052E.
I dont own a 50 ohm terminator. What is that?

[thb]

15 MHz is almost DC  I guess the function generator is broken.

[alm]

15 MHz is not considered a high frequency for common coax, even when taking the odd harmonics of the square wave into account. A short piece of RG-58 should be fine. Make sure you terminate it into 50 ohms at the scope end for best signal integrity. Ideally an internal terminator in the scope or feed-through terminator, but a tee and normal terminator will also work at these frequencies.

Just poking a 10x probe into the output connector and touch the ground ring to the BNC shell should give a decent representation of the signal.

[PeterG]

Try testing with scope leads or test with a 50 ohm terminator.

Just an idea.

Regards

[thb]

From my experience a missing 50 Ohm termination will approx. double the amplitude, but not change a square wave into a sine. Not even closely at 15 MHz. It would be interesting to know if it's a perfect or crappy sine signal on the scope.

[bingo600]

My Rigol DG1022 is spec'ed to 20MHz sine , but only 5Mhz Square.
Could it be somewhat the same on the Agilent ?

Either Rigol FSCK'ed up in the design , or maybe it's common that the square harmonics are degrading the output (making the square round).
But then again the price of the Agilent , ought to enable them to insert a highspeed comparator and generate a square from the sine.

You could try to convert the sine to a square with an AC14 , and then have a look on the scope.

Did you remember to terminate the Scope end of the BNC with 50-Ohm ?

/Bingo

[vk6zgo]

What kind of Oscilloscope are you using?

[alm]

The Agilent 33120A is specced for square waves up to 15 MHz with t_r < 20ns, which is not exactly a perfect square wave. It looks more like a distorted sine. You should see a difference when you switch it to sine mode. Try measuring the rise/fall time, it should be < 20 ns into 50 ohms (usually also < 20 ns without termination if your cable is not too long). Of course you need a scope with a rise time well below 20 ns to measure this (50 MHz bandwidth should be fine).

You can get a much faster square wave from the sync output, although with a fixed amplitude and DC offset. This one can produce ~1 ns rise times and definitely needs termination to prevent ringing.

[king.oslo]

I don't own a 50 ohm termnination. Is this a 50R pull down resistor at the input of the scope?

The scope is the DS1052E.

At 15MHz, the squarewave doesn't look like a perfect sine wave on the scope, but it looks very transformed. As I turn the frequency down, it begins to look more and more like sqarewave. At 1MHz, it looks about square.

What are your thoughts?

Thanks.M

[jahonen]

Lumped capacitance of any transmission line is pretty much meaningless parameter since it has close relationship with the line impedance and line length. If you look at something like Sucotest 18, you'll notice that even that has something like 90 pF per meter, yet it performs very good even at 18 GHz.

Attenuation at higher frequencies is primarily due to dielectric and conductor (resistive) losses, not the LC parameters. Even lousy coax like RG-174 should perform ok for the 15 MHz square wave. Something else must be the problem.

Regards,
Janne

[larry42]

20ns Risetime => 50MHz, i.e. the 5th harmonic of 15MHz.

And it's a 50MHz scope. I'd say it's as expected.

My HMO1524 (150MHz) and 33521A also don't output a nice square wave at 30MHz.

[alm]

My impression is that the generator is within spec. A 15 MHz signal with 20 ns rise/fall times will consist of a 20 ns rising edge from 10% to 90% of the amplitude, 13 ns of a 'flat' part above 90% of the amplitude, 20 ns falling from 90% to 10% and again 13 ns below 10%. Not that far from a sine wave.

A 50 MHz scope will have a rise time of ~7 ns, and should be able to measure 20 ns rise times fairly accurately.

[vk6zgo]

I don't own a 50 ohm termnination. Is this a 50R pull down resistor at the input of the scope?

The scope is the DS1052E.

At 15MHz, the squarewave doesn't look like a perfect sine wave on the scope, but it looks very transformed. As I turn the frequency down, it begins to look more and more like squarewave. At 1MHz, it looks about square.

What are your thoughts?

Thanks.M

You  wouldn't normally refer to it as a "pull-down" resistor,but yes,it is a 50 Ohm resistor connected across the input of the 'scope input.
A cable terminated in its characteristic impedance no longer looks like their measured capacitance & inductance.

As alm has pointed out,the "square" wave from this device isn't too spectacular in the first place,particularly when you get up a bit
in fundamental frequency.
If it was a good square wave,you would be seeing the limitations of your Oscilloscope,

[ejeffrey]

Almost all function generators have 50 ohm source termination.  If done properly, that means that the reflected wave is mostly absorbed after the first reflection.  Even without termination at the scope you should see too much distortion of the waveform, certainly not at 15 MHz.  Even if you did, you wouldn't see a square wave turn to sine, you would see ringing on the rising and falling edges.  Probably you are seeing the limiting bandwidth of the generator.

[BravoV]

I thought the OP's scope DS1052E does NOT have 50 ohm termination input mode, cmiiw.

[alm]

The sync output should be able to exercise the limitations of your scope, although it's not specified for either rise time or aberrations. In the past I measured it to have a 1 ns rise time into 50 ohms on a 500 MHz scope.

[JuiceKing]

My impression is that the generator is within spec. A 15 MHz signal with 20 ns rise/fall times will consist of a 20 ns rising edge from 10% to 90% of the amplitude, 13 ns of a 'flat' part above 90% of the amplitude, 20 ns falling from 90% to 10% and again 13 ns below 10%. Not that far from a sine wave.

A 50 MHz scope will have a rise time of ~7 ns, and should be able to measure 20 ns rise times fairly accurately.

I have this same function generator and this is exactly what I see. Despite the fact that this function generator is used widely in university labs for introductory electronics courses and the consternation you'd expect around the high-frequency squarewave output you see, there is paradoxically no discussion about it to be found. (I looked.) And, no pictures. So, here are some. How do they compare with what you see?

You do need 50-ohm termination on your scope if you want decent waveforms. I am using a pass-through terminator in these examples.

- Ken

[JuiceKing]

The sync output should be able to exercise the limitations of your scope, although it's not specified for either rise time or aberrations. In the past I measured it to have a 1 ns rise time into 50 ohms on a 500 MHz scope.

It never occurred to me that this would be usable in this way, but you're right! On my 100Mhz scope, I'm seeing a rise-time between 3-4nsec. If you can live with the fixed amplitude, this is a better signal.

[king.oslo]

Hello everyone, thanks for your relies.

Ken Schwarz, that was very useful. My waveforms look very much like yours. I measure 20-21ns rise time.

But I have no input termination, so I cannot compare with input termination screenshots.

Which imput terminator do you recommend?

Kind regards,
Marius

[JuiceKing]

But I have no input termination, so I cannot compare with input termination screenshots.

Which imput terminator do you recommend?

I bought mine from probemaster.com.

http://www.probemaster.com/product_info.php?cPath=7&products_id=268

You want the 50-ohm, pass-through type (P/N 1024). You can get it with 0.1% tolerance (P/N 1024A) for more money, but I'm not sure how valuable that really is. If you want to check calibration of your signal generator, I think you'll want the 0.1% version--at least the HP service manual calls for this specification. Passthrough terminators with 0.1% tolerance are surprisingly hard to find.

- Ken

[king.oslo]

http://www.ebay.com/itm/390435505932?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649#ht_1522wt_956

I bought two of them. Is that going to do?M

[SeanB]

I use whaty I have a lot of, old 10base  t pieces and the 50R terminators for the end of the cabling.

[bingo600]

I use whaty I have a lot of, old 10base  t pieces and the 50R terminators for the end of the cabling.

I use the same ...
A BNC-T and some old 10base-T 50-Ohm terminators.

Cheap

[SeanB]

Got plenty from upgrading the network from coax to first CAT3 then now Cat5 cable. I reused a lot of the old coax to wire up video cameras, often it was only a case of extend the cable up to the camera and put in power. Still have a few crimp connectors around, and still have the crimp tooling.

[alm]

You want the 50-ohm, pass-through type (P/N 1024). You can get it with 0.1% tolerance (P/N 1024A) for more money, but I'm not sure how valuable that really is. If you want to check calibration of your signal generator, I think you'll want the 0.1% version--at least the HP service manual calls for this specification. Passthrough terminators with 0.1% tolerance are surprisingly hard to find.

You would probably also use a power meter, since scopes will not usually have 0.1 % amplitude accuracy. The 0.1 % spec is only for DC, and not very meaningful for high frequency, since the max. VSWR is 1.2. Most instruments and cables are not specced for 0.1 % tolerance in impedance either. Tektronix had a terminator that was 50 ohm +/ 0.1 %, but it was only specified up to 100 kHz or so. Was used for some low frequency amplitude calibration.

Got plenty from upgrading the network from coax to first CAT3 then now Cat5 cable. I reused a lot of the old coax to wire up video cameras, often it was only a case of extend the cable up to the camera and put in power. Still have a few crimp connectors around, and still have the crimp tooling.

10base2 is 50 ohm, video is 75 ohm. Fortunately analog video signals are low frequency, so impedance mismatches aren't usually a problem at moderate distances.