USB PC oscilloscope recommendation

[neazoi]

Hi, I am thinking of buying a usb oscilloscope/FFT for the lab, but mubudget is limited. My work is mainly on HF
How about this one?
Hantek 6254BD (datasheet attached)

Prices are 250-350Euros or so.
You get FFT plys a 25MHz waveform generator.
Please tell me if it's worth it or not.

The problem I see is, the input is AC/DC but only 1M impedance. How can I measure in 50R? Will a simple 50R shunt resistor be fine?

 

[Aussie Susan]

I'm not sure I understand the problem with the probe impedance. You want the probe impedance to be as high as possible so that it does not affect the voltage measurement by drawing current.
If you are measuring the voltage across a 50 ohm impedance then the circuit will see the 50 ohms in parallel with the 1Mohm which is 49.9975 ohms - probably not enough of a difference from 50ohms to make a measurable difference.
Also, reading the manual you reference, it shows that the probes have a 10:1 impedance switch which I would interpret as meaning that you can actually get a 10Mohm probe impedance.
I'll not comment of whether it is worth it or not. The spec sheet says 1Gs/s in the hardware and that should be plenty for reliable HF waveform representations, even up to 50MHz. I have a hardware scope that also has 1Gs/s but my interests are more in the VHF area and it starts to struggles about 100MHz (the square waves start to be a bit rounded).
Susan

 

[neazoi]

I'm not sure I understand the problem with the probe impedance. You want the probe impedance to be as high as possible so that it does not affect the voltage measurement by drawing current.
If you are measuring the voltage across a 50 ohm impedance then the circuit will see the 50 ohms in parallel with the 1Mohm which is 49.9975 ohms - probably not enough of a difference from 50ohms to make a measurable difference.
Also, reading the manual you reference, it shows that the probes have a 10:1 impedance switch which I would interpret as meaning that you can actually get a 10Mohm probe impedance.
I'll not comment of whether it is worth it or not. The spec sheet says 1Gs/s in the hardware and that should be plenty for reliable HF waveform representations, even up to 50MHz. I have a hardware scope that also has 1Gs/s but my interests are more in the VHF area and it starts to struggles about 100MHz (the square waves start to be a bit rounded).
Susan

You are right about the impedance things, but you assume that the equipment to be measured is internally matched to 50R. To see how homebrew unmatched equipment behave at 50R, a shunt resistor has to be connected at the input of the scope, so that the measurement is done at 50R. In other words in RF you have to match your devices to be measured to 50R if they are not.
You are actually measuring at 1M, or whatever your scope probe is, but the devices are loaded with 50R. So the end result is measuring at 50R. At least that is what I am thinking of.

My current scope is 500MHz analogue bandwidth and 1Gs/s and I have seen problems with sinewaves near the top edge of the frequency. I do not know at what frequency the actual problems start though since I do not have another more powerful one to compare. I only consider a 250MHz 1Gs/s usb scope, for less size and weight (I have a laptop already), which is important to me.

 

[123jack]

I think I've mentioned picoscope here before. (The make I use)
They have a range online you could use to compare specs and prices with.
They are continually adding software features and provice free ongoing updates.
I've found them very good over time. (I bought my first one about 7 years ago)

 

[Aussie Susan]

You are right about the impedance things, but you assume that the equipment to be measured is internally matched to 50R. To see how homebrew unmatched equipment behave at 50R, a shunt resistor has to be connected at the input of the scope, so that the measurement is done at 50R. In other words in RF you have to match your devices to be measured to 50R if they are not.
You are actually measuring at 1M, or whatever your scope probe is, but the devices are loaded with 50R. So the end result is measuring at 50R. At least that is what I am thinking of.

Of course - if you are measuring an open circuit then all you have is the scope probe impedance and the voltage can be completely unrepresentative of the voltage under load. I must admit that I assumed that you were referring to a 50ohm dummy load.
Nyquist says that you need to sample at twice the frequency you need to reconstruct. However that only applies to sine waves. If you want to reconstruct other shapes then you need to look at the highest frequency that is present in the reconstructed wave and double that for the required sampling frequency. You are working this situation the other way so given a 1Gs/s then you can get a 500MHz sine wave but probably only a 50MHz square-ish wave. That should be fine for HF work (especially if you are looking at SWR etc. waveforms).
Susan VK3ANZ

 

[neazoi]

Of course - if you are measuring an open circuit then all you have is the scope probe impedance and the voltage can be completely unrepresentative of the voltage under load. I must admit that I assumed that you were referring to a 50ohm dummy load.
Nyquist says that you need to sample at twice the frequency you need to reconstruct. However that only applies to sine waves. If you want to reconstruct other shapes then you need to look at the highest frequency that is present in the reconstructed wave and double that for the required sampling frequency. You are working this situation the other way so given a 1Gs/s then you can get a 500MHz sine wave but probably only a 50MHz square-ish wave. That should be fine for HF work (especially if you are looking at SWR etc. waveforms).
Susan VK3ANZ

I mainly consider sinewaves and their harmonics, and not really interested about the fast switching edges of square waves.
You know HF sinewaves and their harmonics, that can go as high as 100-150MHz or so max.

Of course, when you buy an instrument wou plan for future work as well, but in the current time, this is my work and space and weight is more important.

Will a USB 1Gs/s 250MHz analogue bandwidth scope be able to do accurately enough measurements for my purpose?

- - - Updated - - -

I think I've mentioned picoscope here before. (The make I use)
They have a range online you could use to compare specs and prices with.
They are continually adding software features and provice free ongoing updates.
I've found them very good over time. (I bought my first one about 7 years ago)

I have seen of the pico scopes as well. But they are way too expensive for their characteristics. Look at the price of a 3000 series 200MHz one, it is 4x times the Hantek!

 

[c_mitra]

You are right about the impedance things, but you assume that the equipment to be measured is internally matched to 50R..

To measure voltage, you need a probe that has as high impedance as possible. You can even get a 100X probe and that comes handy when you have to measure high voltages...

To measure current, you need a probe that is in series with the current source and a modern oscilloscope is not useful for this purpose.

You need to match the impedances (input Z= output Z) when you need to maximize the transfer of power. With an oscilloscope you are mostly measuring the voltage using a high impedance probe.

I do not understand what you mean by "but you assume that the equipment to be measured is internally matched to 50R"- what is this 50R coming from?

 

[neazoi]

I do not understand what you mean by "but you assume that the equipment to be measured is internally matched to 50R"- what is this 50R coming from?

Commercial RF gear is internally matched to 50R (let's assume that this is true for all the gear's frequencies). So I suppose one can measure the voltage accross it with the 1M scope directly?

However in homebrew equipment, one need to load the RF with a 50R shunt resistor, so as to ensure operation at 50R. Then you can measure with the stope 1M accross this 50R and then your measurements will be true for 50R impedance.

Are these points right?

 

[c_mitra]

Are these points right?

I suggest you to read the article:

https://www.qsl.net/va3iul/Impedance_Matching/Impedance_Matching.pdf

this is a good one, from a technical standpoint.

 

[neazoi]

I suggest you to read the article:

https://www.qsl.net/va3iul/Impedance_Matching/Impedance_Matching.pdf

this is a good one, from a technical standpoint.

Thanks, but too much theory for me. All I am asking is practically, how to measure with the 1M scope input, a circuit (for example an oscillator that has a capacitor coupling on it's output) that has an unknown impedance and I want to measure it on 50R.

All the examples I have seen on the net, connect a 50R shunt resistor to ground (or a feedthough 50R terminator), so that the measurements are done at 50R, despite the fact that the scope has an input of 1M. Is that accurate enough?

It is important to do the measurements on 50R for my RF homebrew circuits, so I have to decide if this scope solution is adequate.

 

[c_mitra]

Basically scopes are voltage devices, they measure potentials. And they do the job well.

You can measure voltages at any point on a circuit using a scope. Because of the high impedance (most scopes has an input of 1M in parallel with 10 or 20pF but they most common probes are 10X) they do not load the circuit much. But perhaps you will not be able to see the voltage waveforms on a crystal's terminals or the oscillations in a tank circuit. My scope works only upto 100MHz but then I work only sparingly....

There is no need to use any special resistor in addition to the standard probe to measure potentials on a RF circuit board (if the voltage and frequency are within the limits...)

 

[neazoi]

There is no need to use any special resistor in addition to the standard probe to measure potentials on a RF circuit board (if the voltage and frequency are within the limits...)

You are saying that because the scope won't signifficantly load the circuit and that is true. However, since I do not know the output impedance of the circuit under test and since I need the circuit to operate with defined characteristics on 50R, won't a 50R shunt resistor needed at it's output?
This way my measured voltages will refer to a resistive 50R load.

Am I wrong?

 

[c_mitra]

What I am saying that you do not need any extra 50E resistor to measure the voltage on a scope. If the 50E resistor is part of the board, let it be. But you do not need anything extra. For example, I do not understand the meaning of the sentence:

I do not know the output impedance of the circuit under test and since I need the circuit to operate with defined characteristics on 50R

The potentials at a given point in a board can be tested with a scope probe without any extra resistor. If you need to use the circuit with a 50E load, let it be there...

For example, many RF circuits use an antenna for power (some will say signal) transmission and reception and the circuit must be matched with the antenna, the antenna must be matched with the empty space...

These matchings are traditionally done at 50E and you will find it rewarding to study how the antenna are matched with empty space (and 50E impedance)

 

[neazoi]

What I am saying that you do not need any extra 50E resistor to measure the voltage on a scope. If the 50E resistor is part of the board, let it be. But you do not need anything extra. For example, I do not understand the meaning of the sentence:



The potentials at a given point in a board can be tested with a scope probe without any extra resistor. If you need to use the circuit with a 50E load, let it be there...

For example, many RF circuits use an antenna for power (some will say signal) transmission and reception and the circuit must be matched with the antenna, the antenna must be matched with the empty space...

These matchings are traditionally done at 50E and you will find it rewarding to study how the antenna are matched with empty space (and 50E impedance)

Measuring 1vpp at 50R is different than measuring 1vpp at 1M. By shunting a 50R, you measure the power of an oscillator at a 50R load.
Am I missing something here?

 

[schmitt trigger]

Thanks, but too much theory for me. All I am asking is practically, how to measure with the 1M scope input, a circuit (for example an oscillator that has a capacitor coupling on it's output) that has an unknown impedance and I want to measure it on 50R.

All the examples I have seen on the net, connect a 50R shunt resistor to ground (or a feedthough 50R terminator), so that the measurements are done at 50R, despite the fact that the scope has an input of 1M. Is that accurate enough?

When I used to work with RF, I always used a 50R feedthru with a female BNC on one end, a male BNC on the other.
We had some expensive Tektronix scopes which had an internal (selectable) 50 ohm termination. But the danger of forgetting about it and leaving it on was too great, so everyone around the lab always used those terminations instead.

Similar to these:
**broken link removed**

 

[neazoi]

When I used to work with RF, I always used a 50R feedthru with a female BNC on one end, a male BNC on the other.
We had some expensive Tektronix scopes which had an internal (selectable) 50 ohm termination. But the danger of forgetting about it and leaving it on was too great, so everyone around the lab always used those terminations instead.

Similar to these:
**broken link removed**

What is my equipments I want to measire is internally matched to 50R? This will parallel with the feed through resistor, resulting in 25R is that right? So do I have to use this feed through resistor in this specific case??

 

[c_mitra]

Measuring 1vpp at 50R is different than measuring 1vpp at 1M. By shunting a 50R, you measure the power of an oscillator at a 50R load.
Am I missing something here?

I do not see how you can get the power of an oscillator from a single point measurement at a 50R load; you will not even know the voltage!

Oscilloscopes shine when asked to show fast changing voltages!

 

[neazoi]

I do not see how you can get the power of an oscillator from a single point measurement at a 50R load; you will not even know the voltage!

Oscilloscopes shine when asked to show fast changing voltages!

I am sorry I do not understand what you are saying
Can you please explain?

 

[ads-ee]

Maybe you should look at this article. No where in this scenario are you running the scope with 50 ohm inputs, you are measuring voltage and therefore want 10Mohm input impedance to minimize any effect the scope may have on the circuit.

- - - Updated - - -

And here is an article on current sensing.

 

[neazoi]

Maybe you should look at this article. No where in this scenario are you running the scope with 50 ohm inputs, you are measuring voltage and therefore want 10Mohm input impedance to minimize any effect the scope may have on the circuit.

- - - Updated - - -

And here is an article on current sensing.

I think that there is a misunderstood here. I always measure at 1M probe, no doubt.
The 50R feed through resistor is a part of the circuit in fact and not part of the scope.
As I see it, it is used to load the oscillator to 50R, so that measurements are done with the oscillator loaded at 50R.
Then any voltage measurements with the 1M scope probe will actually refer to 50R.
Is that so confusing?

What I am saying is that if a commercial module is to be measured, there might be no need to put the 50R feed through resistor, because these modules are already matched to 50R, in other words their output impedances are 50R. So paraleling another 50R (the feed through) will give voltages at 25R load, not 50R.