Modern scope for vintage radio work

[agardiner]

Yes, I think it is a CPC own badged Owon scope.

[Cruisin Marine]

I find modern scope's nothing less than a complete pain.

[Katar83]

Adrian,

I have a digital and analogue scope here, happy to bring them with me with the Murphy radio that we discussed, so you can confirm if the noise is down to scope or something else. Drop me a message if you're interested, don't use mine very often but they are both working well.

PS

I had an owon vds1022 earlier and found the manufacturer software for it lacking. There was a 3rd party software for it that was much better but I've sold it in the end and bought a Tektronix, much nicer scope.

[agardiner]

That is very decent of you, many thanks. I may take you up on the offer, depending on time constraints! Plus of course there is Retrotech this weekend......

Meanwhile, I have started watching an interesting video about scope noise from Dave Jones at EEVBLOG. Apparently, some of the issue is that analogue scopes don't show the noise that is actually present, rather than digital ones being noisy. You can watch it here if interested: https://youtu.be/Znwp0pK8Tzk

Note; I am not criticising analogue scopes at all. As someone has already pointed out, it can be a case of the right tool for the right job.

[agardiner]

OK, I had a bit of time this afternoon, so have carried out more tests regarding the noise issue with this modern scope. I tried the following.....

1) High pass filter on the scope input - 100KHz low cut off
2) Low pass filter on the scope input - 15MHz high cut off
3) Plugging in scope with earth disconnected from the mains lead
4) Combinations of above.

None of these measures makes much difference, although disconnecting the earth from the mains makes matters worse.

The interesting fact is that the problem only manifests itself on the VHF band. I can use the scope for AM work without any issue. I can also use it successfully for VHF work if I shut down the sets local oscillator and inject IF directly; this produces a very usable S curve and a fairly clean IF waveform.

So, suitable workarounds have been established. Interesting though that the noise coming from the scope is in the VHF band. From the way the scope behaves and the way the noise changes depending on what's on the screen, it would appear to be primarily a screen driver noise. I wonder if there might be a way of filtering it in the scope itself

[GMB]

It sounds to me like you have bought a non CE compliant scope. Radiation in the VHF and up is typical of badly designed modern electronics.
This is why we have the CE (or UKCA) regulations!

Sadly some foreign manufacturers think the requirement is to just write CE on it, skipping the expensive certification tests. I would consider complaining to the supplier as "not fit for purpose".

[cmjones01]

Even a CE compliant piece of equipment doesn't generate no noise. It just means that it doesn't exceed a certain limit when used as intended. Everything gets tested with the bare minimum of wiring attached to it to represent "real use", because that makes it more likely to pass. I can guarantee that anything with an LCD screen on it, whatever marks it has (CE, UL, UKCA, whatever) will emit interference at various frequencies.

The CE and UKCA self-certification process means that the device meets the required limits, not that those limits are good enough for working on the insides of FM radios.

Chris

[Radio Wrangler]

Some scopes deliberately scatter their sampling timing to effectively increase their resolution with repetitive signals. An analogue interpolator is used to measure the scatter offset on each sample so that it can be displayed at the right time value. Some scopes just run a super fast digitiser without any monkey business, but some of these play pseudo-random scattering tricks when they need to discard samples in order to conserve enough trace memory for the sweep time. This is used to prevent the creation of regular artefacts from a regular sample dumping sequence.

These pseudo-random processes act to scramble the spectrum of any sampling noise radiated or conducted from the scope. On a radio, it sounds like random noise.

David

[agardiner]

So from everyone's suggestions, you could probably reduce the noise by housing the scope in a metal cabinet? Not that I am going to do this! Just curious.

Very little noise is actually coming from the probe, most is just radiated from the scope itself. It is both interesting and someone irritating, that so much modern electrical noise affects the AM band, and I am lucky to have a good AM signal where I am, yet a piece of test equipment more useful for FM work should affect the FM band! Typical.

[cmjones01]

Originally Posted by agardiner

So from everyone's suggestions, you could probably reduce the noise by housing the scope in a metal cabinet? Not that I am going to do this! Just curious.

Putting it in a metal cabinet might reduce the noise, or it might make it worse. Welcome to the wonderful world of EMC (Electro-Magnetic Compatibility) testing. Manufacturers, most of them at least, do go to a lot of effort to get their products through the tests successfully and the tests are taken seriously. They're also expensive, so not something you want to keep repeating. The problem is that controlling RF emissions from equipment and giving it immunity from the RF emissions from other things is simply hard.

To give some examples: LCD screens tend to be quite noisy, because they're a physically quite large area covered in a matrix of wires and transistors all of which are constantly being scanned at high frequency - a typical pixel clock might be 50MHz, with the horizontal clock at 40kHz. That matrix forms a lovely array of antennas which inevitably end up resonating at some harmonic of the clock frequencies.

Getting rid of noise emitted from digital and analogue circuits inside the box is also tricky. You can filter things and ground things, but all those ground connections have inductance and resistance, so they're not perfect. It can easily turn out that ground wires or metal housings actually turn in to antennas and make matters worse. And the whole thing is likely on the end of a long mains lead, as well as other connections, all of which have their effects, usually making emissions worse.

I dealt with one case where a well-proven design involving a modern microprocessor passed all the EMC tests, and then its PCB was enlarged a bit to accommodate a modification. Suddenly a load of spikes appeared in its emission spectrum, so it failed the tests. It turned out that the spikes were at harmonics of the resonant frequency of one dimension of the PCB, which happened to coincide more-or-less with the processor's clock frequency.

Speaking of spikes, even a single spike over the limit will fail the test. They're often at frequencies used in the circuit, such as a bus clock. So what do we do to fix them? One answer is to wiggle the clock frequency around a bit, make it noisy, so as to spread its energy over a broader bandwidth. That reduces the height of the spike so the test passes, but the total emitted energy remains the same and may even be more objectionable when heard as interference on a radio, for example.

Then there's the matter of immunity: the equipment shouldn't do anything odd when exposed to RF radiation within certain limits. This also turns out to be hard. I've recently been involved with an electric vehicle charger, which if course has to work with whatever cable is plugged in to it, and we have no control over its length or screening properties. The cable carries both power and low-level control signals. It's very likely that someone will be using a radio transmitter nearby - think of all the radio equipment fitted to vehicles, everything from CB radios, through amateur radio equipment, to 5G phones. It's all going to be near that cable, so the tests simulate this. All that RF energy ends up finding its way in to the charger when it's picked up on that cable, and it's not allowed to affect the charger's operation. It's surprisingly easy to thoroughly upset a 12 volt control signal by injecting a few watts of RF in to it. In the case I was involved with, the components added to solve one EMC problem actually made immunity worse, so we had to find another solution. In these cases, it can end up mattering which side of a PCB an earth wire is attached to. That's the level of detail we're talking about.

I hope I've explained a bit what a CE or UKCA conformity mark really means. It doesn't mean that the product is perfectly quiet at RF. Conversely, if a product emits RF noise, that doesn't mean its conformity mark is fake. Better EMC performance, in excess of that required by the regulations, is one of the things that may differentiate a £20,000 oscilloscope from a £200 one.

Chris

[agardiner]

Wow. I had no idea that it was that complicated, but having spent my career in service rather than design, it is not surprising. Thank you Chris for such a comprehensive explanation. Can't even begin to think of the complications designing life dependant systems such as aircraft then!

On that note, I will happily put up with a little bit of interference in the FM spectrum!