Modern scope for vintage radio work

[agardiner]

Hi,

Quick question. I have purchased a new 20MHz DSO oscilloscope. Although a fairly basic model, I bought it from CPC at around £200 rather than a risky cheapo.

The scope is very nice, although I have now realised that X-Y modes are not as good on a DSO. My question is that the scope appears to give off rather a lot of noise; very noticeable when probing around an FM radio. It doesn't appear to be the SMPS, more the LCD.

So, does anyone else here use a modern scope, and therefore should I change mine for a different model, or are modern scopes no good for vintage radio work?

Thanks.

[agardiner]

As a follow up to this, I wonder if my problem is primarily some kind of ground loop. I have noticed that the problem is mainly when the sweep output of my signal generator is connected to my scope. If I disconnect it, then most of the noise disappears. Not perfect, but certainly very usable. It is not the signal generator itself, so must be some kind of noise loop via the grounding.

Any advice please?

[agardiner]

Another follow up. Just tried removing the ground connection on the scope probe; ie, no ground from scope to radio. Radio is otherwise grounded through the sig gen.

Noise is much much lower, so progress and I am on the right lines. Any further help would be much appreciated, as this noise is still too high to have a good trace for FM alignment.

As a side note - the radio I am playing with is a battery one.

[Radio Wrangler]

A digital oscilloscope design has to have the absolute minimum of circuitry before the sampler on each channel. This is all trying to get the highest bandwidth number to use to sell the thing.

Samplers are driven by fast pulses, very fast pulses, and whatever sampling rate they have spectra up to extreme frequencies. A small fraction of the very strong sampling pulses comes whamming OUT of the input sockets of the scope. Up your scope prove cable backwards and gets radiated off into free space. It's a damned jamming generator!

Maybe a buffer amplifier would help, maybe there's enough injected into your ground structures etc.

There in need for linear, analogue scopes still.

Modern network analysers and spectrum analysers sample and digitise their results, but the samplers are well isolated, buried inside them and made no faster than they need to be. They aren't attempting to be general purpose wide bandwidth oscilloscopes competing for a headline frequency number, and a lot of work has gone into screening and preventing RF from fast logic, CPUs and samplers getting out to the stuff under test.

I did the Agilent noise figure analyser. A box with an SMPS, microprocessor, LCD display and all sorts of interfaces, but connected on the front would be something like an amplifier or receiver with an equivalent noise temperature down to within a dozen degrees of absolute zero. Measurements don't come much more sensitive than this. Often extraneous signals from much further away force you to move into a screened room, but you'd still like to take your test gear in with you....

David

[agardiner]

Many thanks David. Excellent information which explains it very well.

One question then. If the sampler is perhaps the primary source of the noise, would a scope with a higher sample rate improve the situation? The one I chose is a 20Mhz model with sample rate of 100MSPS. Other models are available in the same range, including a 50Mhz, 500MSPS and a 100Mhz 1GSPS. I realise they will still produce noise, but wondered how these increased rates would affect the low IF frequencies in a radio. Would it change the effect in a positive way?

Regards,

[Radio Wrangler]

The sampling rate written on the brochure and advert is the headline number. What it can do flat-out.

But when you're digitising slw waveforms, like scans, much lower sampling rates are used, or else the thing's available memory would be filled in far less than your sweep time. So a 'faster' scope is unlikely to fix the problem.

Also, even on the low sampling rate, the rise and fall times and sampling pulse width wo't change, assuring you RFI all the way to astronomical frequencies.

May their houses be cursed unto the seventh generation!

David

[PJC58-Hythe]

I use a Rigol DS1202Z https://www.rigol-uk.co.uk/product/r...-oscilloscope/ and I don't have any issues with noise otherwise it would have gone straight back. I can't see the point in having test equipment that can introduce problems when being a used as a diagnostic tool!

[agardiner]

Thanks David and Peter.

[Keith956]

I have a Siglent scope and haven't seen any significant noise from it, I have checked using a spectrum analyser. By far the biggest source of noise is from a magnifying glass CFL lamp which usually has to be turned off when measuring anything sensitive!

[kalee20]

I'm wary of fully digital 'scopes unless I have done a sanity check with an analogue 'scope.

It's difficult to beat a beam of electrons, scanning a phosphor screen in real time.

David's points - yes makes sense! And I can well believe that modern electronics can do a pretty impressive job of sampling and digitising the input signal, using modern, affordable mainstream componentry rather than exotica.

That being so, I can also imagine that the difference between a cheap 'scope and an expensive one is that the extra money is spent not on electronics, but on casings - metalwork, screening, shielded compartments for noisy circuitry, maybe a RFI-opaque viewing window over the display. Such hardware doesn't come cheap...

[agardiner]

Originally Posted by Keith956

I have a Siglent scope and haven't seen any significant noise from it, I have checked using a spectrum analyser. By far the biggest source of noise is from a magnifying glass CFL lamp which usually has to be turned off when measuring anything sensitive!

Indeed. My magnifier lamp is terrible for noise. I only use it when needed and have a 2nd lamp for close up bench illumination. It is a fluorescent type and cost a lot of money, but works well and is quiet.

Originally Posted by kalee20

I'm wary of fully digital 'scopes unless I have done a sanity check with an analogue 'scope.

It's difficult to beat a beam of electrons, scanning a phosphor screen in real time.

David's points - yes makes sense! And I can well believe that modern electronics can do a pretty impressive job of sampling and digitising the input signal, using modern, affordable mainstream componentry rather than exotica.

That being so, I can also imagine that the difference between a cheap 'scope and an expensive one is that the extra money is spent not on electronics, but on casings - metalwork, screening, shielded compartments for noisy circuitry, maybe a RFI-opaque viewing window over the display. Such hardware doesn't come cheap...

Totally agree. My scope was inexpensive in the grand scheme of things at just over £200. You can buy similar spec'd scopes for triple that. Trouble is balancing need and cost.

[agardiner]

OK, so thanks for all the input. I have managed to use this information to improve the situation and minimize the noise to the extent that it is usable.

I have also now ordered an inline BNC low pass filter, that should cut off anything above 15MHz, so hopefully this will clear up most of the remaining noise. Once I receive the filter, I will try it out and report back here in the hope that it may be of use to others who own/want to purchase a modern DSO.

[majoconz]

GBP200! Wow! My nice little Philips dual beam goes up to 40MHz and still showing a signal up to 70MHz - and it was NZ$40 - say 20quid - at an auction. It said "Faulty" but I'm still looking for the fault - I think it was 'operator error' as most of the switches were not where you would expect them to be.

[Craig Sawyers]

I bought a nice little Tektronix digital scope a few years ago from a list member here for a couple of hundred quid. TDS2004B. And generally very good it is too. Lightweight, four channel, 60MHz.

But...

I was measuring my analogue crossover, loads of opamps, and darn it - what looked like low level oscillation. Chased my tail for a good two hours trying to find which bit was hooting.

Then for some reason I turned the scope on its end. Oscillation disappeared!

It was coupling from the scope's switched mode supply. So it was not a circuit fault at all.

So these little scopes can sometimes give odd and surprising problems. Even Tektronix ones.

Craig

[Pilot Mariner]

The best feature on my Hameg HM507 is the component tester function. I use this more than any other feature on the scope.

[cmjones01]

Modern electronics is generally noisy, full stop. Doesn't matter whether it's a scope, computer or whatever, all the CPU-controlled gadgetry, switch-mode power supplies and digital screens emit all manner of grot at all sorts of frequencies. Yes, it's tested (or ought to be tested) to comply with the various EMC directives in various parts of the world, but they just set upper limits. All equipment, apart from the very highest-end gear (like the aforementioned noise figure analyser, I suspect) produces easily measurable interference. It's radiated in all directions from the cabinet, and conducted down the mains lead and any other connecting cables. People like me get paid to assess it and make it "good enough" to pass the tests. All this dwarfs, in practice, any effects from samplers on the input of a scope, for example (my favourite digital scope is so old it has a CCD delay on the input before the A/D converter but I digress).

When making a measurement, the key is how to minimise this noise's effect on the measurement. Earth loops are a real problem, I find, because every piece of equipment is squirting some variable amount of noise down its earth conductor, so their (earthed) metal housings are never at quite the same voltage, as far as RF is concerned. If you're using a scope probe to look at a circuit which is itself connected to a signal generator, this means that the scope probe's clip ground ends up being at a different voltage to the scope's housing (and thus the ground on its input socket). Now there's a current flowing along the screen of your scope probe cable. The cable has inductance and resistance, so there's thus a voltage drop, which appears in series with the signal you're trying to measure. Result? A noisy trace.

This becomes intolerably bad when, as is often the case in modern electronics development, there's also some USB-attached widget (like a microcontroller programming dongle, or serial adapter) also connected to the circuit. The host computer is earthed, and thus so are all the USB things, but there's a significant current flowing along all those USB cables in the form of the widgets' power supply current and any common-mode leakage from their data connections. Now the earth loop noise gets really bad. On my bench, if I'm trying to make an analogue measurement on such a circuit with any degree of precision, I have to physically unplug all those USB-attached connections.

It gets even more fun when there are heavy currents flowing. Some projects I work on involve currents of 500A or more, and characterising transients of that magnitude is really hard when it turns out that the pulse you thought you measured was actually just induced current flowing in the shield of the cable connecting your current probe amplifier to the scope...

There are solutions: isolating test equipment from ground can help, but has to be done with great care and awareness of the safety consequences. I've toyed with the idea of powering various items of equipment on my bench from a centre-earthed 55-0-55V site transformer and removing the mains earth from their housings. Using battery-powered equipment also works well, and I suspect USB isolators would be helpful but I've not tried them. It's possible to fix it altogether but usually only for a known type and bandwidth of signal. For example, balanced audio connections work well because it's easy to make electronics with good common-mode rejection at audio frequencies. And complex RF and microwave setups avoid noise effects by operating over a narrow band of frequencies using filters. Otherwise we just have to be aware of the issue, understand its effects and put up with it.

Chris

[agardiner]

Great write up Chris, thank you.

Well, the low pass filter has arrived and been tested. No surprise that it doesn't solve the problem. However, it has allowed me to conduct some more tests, and I have been able to measure the noise element. It is approx 600Hz at around 177mV P-P. I say approx, as the noise frequency is varying all the time between around 400Hz and 800Hz; depending on when you grab the sample will depend on the exact frequency measured. It is coming from the scope itself.

Interestingly, the noise is far worse (results as above) when the signal generator is turned off, or disconnected from the 2nd scope channel. When connected and switched on, the noise level is approximately half.

Given that the sample frequency is 100MSa/s and that a typical switch mode is many Khz, this leaves a bit of a mystery as to where this noise is exactly coming from. Any ideas?

[GMB]

I have been able to measure the noise element.

How?

You need to decide if the noise you are seeing on the scope is real or not.
Always remember that there is no such thing as "ground". So the box of the scope may well be being driven by internal noise sources and by connecting to a tuner where the aerial is connected to earth you feed these noises onwards into the tuner input.

Where to do connect your probe? Hopefully the scope is showing you the difference between its two points of contact. Easy to forget that the "signal" you see may actually be because your earth reference is a place of much noise pickup.

You can try some ferrite clamps on your probes to see if that helps. People forget that being coaxial does not stop noise flowing on the ground as it has the option of flowing on the outside of the screen.

[agardiner]

So the scope is connect to just before the detector on an FM receiver. Feeding in 10.7MHz on my generator allows me to clearly see the waveform, but the trace is noisy and I can hear the noise on the receiver output.

I am going to try making up a high pass filter instead, with a suitable high impedance, to see if it helps. Will update when I have tried.

[McMurdo]

If it's the Multicomp branded 20mhz scope, it looks like an Owon, lots of reviews of them in the usual places.

https://www.amazon.co.uk/SDS1022-Osc...03667562&psc=1