Oscilloscope for SMPS design?

You need more than the 25MHz bandwidth to see some of the problems in switchers. Good triggering is important. If you absolutely have to, you can make do with low bandwidth scopes. This is doubly so if this hobby stuff you are doing. You can make external circuits with fast comparitors of the like to prove and disprove theories about why the circuit isn't working right.

Also, an AM radio near where you are working on the switcher can help. If th epower switch is oscillation on the edges it will often overpower the station you are listening to.

Where do you live? Country?

Buy a decent used Tek scope on ebay maybe. The 25 MHz cheap analog scopes don't have vertical delay lines, so are hard to use for fast stuff.

What's your budget?

John

eBay Item number: 150019010203 looks quite good as does Item number:

At least they are local to you.

Get a used one from eBay as others have said. What bandwidth and other features you need depends on your on your requirements. If you are working on say mains input SMPS's then a high voltage differential probe will be more valuable than extra bandwidth, it will allow you to safely probe things without killing yourself or blowing something up. You can get by with a 25MHz model if you are working on relatively low frequency stuff, but I'd really be aiming for a 60 or 100MHz.

Dave :)

yep, a good used 100mhz scope for £100 from used suppliers, even less from ebay, dual trace, dual/delayed timebase, anything less and you probably wish for more, and a current probe too, although you can make one quite easily with an ungapped ferrite core with lots of turns on it, a 1k resistor accros it wich goes to the scope and 1 turn for the current sense.

as well as an isolation transformer you probably also need high voltage probes, x100 are ideal, also when looking at saturation voltages you need a voltage limiter so you can have the sensitivity up high enough to measure without the high voltage peaks wich on some scopes can cuase severe distortions, again these can be made.

Colin =^.^=

More hard questions.. I'm an old fart.

The stuff I've worked on has been mainly off-line, above 50W, 100KHz switching frequency, nothing esoteric.

The only time I've had to look at 'high' frequency stuff, before EMC testing (and that only goes up to 30MHz), it has been in the 20MHz-50MHz range and comes from parasitic inductance resonating with parasitic capacitance and might require snubbing.

Then it's useful to be able to measure the frequencies. Sometimes you sort of know what the value is for one, usually capacitance, then you can guess from there. However it's better to add a bit to get two frequencies then do a sum so you really know......

It's quite nice to do it, do the sums, drop in the calculated RC snubber and see the ringing disappear. Of course, more often than not, you can't see which bit of the shit is being shitty and the other times the resistor burns up.

I'd go for a dual trace 60MHz 'analog' scope with a delayed trigger function and get familiar with it. There is a lot you can 'see' with them and I've seen others, and I've done it myself, totally bollocks up simple measurements with digital scopes.

Many of the problems you will come across either aren't really there or you will have designed them in. A 'better' scope won't help you track them down any faster than sitting back and thinking about what might be happening.

I'll stress that one. Just because you have something on the bench behaving like shit there is more than a good chance that any amount of prodding it with your scope isn't going to tell you anything. Back off, leave it alone, sit down elswhere and think. Get LTspice and use it to think for you as well, it has as many channels as your circuit has things that can be measure and works up to light frequencies and beyond.....

Don't forget, lots of SMPS have oscillators/clocks in them. If you want to look at a waveform and are having problems triggering from it you can use one channel on the 'clock' as your trigger and whizzle along the waveform you are measuring using the delay function. Just be careful about the relative grounding of things.

Lot's of other tricks will become apparent as you think about how you are going to make the measurement you really want to make.

A major thing is making proper connections to make your measurements. It always gets mentioned so I'll apologise for doing it.

Try to avoid using the crocodile clip on the end of the dangly bit of wire that comes with your probes. They should come with both springy things you can slip over the end and an adapter that will let you plug them into a BNC connector you solder to your board.

A high voltage differential probe would be good as well and, if you are doing off-line stuff, an isolation transformer and variac are must haves.

Don't disconnect the mains ground from your scope, some bastard will borrow it and put it back on again, BANG!!! Sometimes your test set up may produce wierd results but you'll figure them out.

Burble

DNA

On Sun, 22 Oct 2006 22:21:57 -0700, Myauk top-posted:

No Problem!

Just find some good books on switching power supply design, and save your money for the next three months while you read them! ;-)

Good Luck! Rich

An old tek 7000 series mainframe with 7a13 and 7a26 plugins. If you can get 1 with storage eg 7633 even better. About $150 on ebay.

snipped-for-privacy@aol.com wrote in news: snipped-for-privacy@i42g2000cwa.googlegroups.com:

He will also need a 7B92 or 7B80/85 time base plug-ins with that,and DON'T buy any 7400 or 7500 series mainframe!!! I suggest a 7904,7904A,7704A or 7603;the 7603 will be the least expensive,but only 100Mhz.Use 7A26/7A13/7B53A with it.

legroups.com:

also need a 7B92 or 7B80/85 time base plug-ins with that,and DON'T

Your right he will need a timebase but they usually come with one. almost anything will do. A storage model is a big plus when working on a SMPS, a 7623 7633 for example. 100Mhz should be more than suffcient.

snipped-for-privacy@aol.com wrote in news: snipped-for-privacy@m73g2000cwd.googlegroups.com:

Storage circuitry is more complexity to fail,and the storage CRTs are old and weak.(and HOW is a storage scope useful for working on switchers,that a regular scope would not suffice?)

If one wants storage,then get a digital storage scope,even a TDS model(before a old CRT-storage scope).

can get 1 with storage eg 7633 even better. About $150 on ebay.

This is the sort of thing you should only buy if you really understand what it is and how to use it. It's like buying an odd car - you don't want to do that if you don't know what you are doing.

Jim Yanik a écrit :

Analyzing start-up and fault conditions.

On Oct 24, 6:23=A0pm, Jim Yanik wrote: . A storage model is a big plus when working on

torage circuitry is more complexity to fail,and the storage CRTs are old

They are more complex but not that much, on a tek scope its just a single board, furthermore if it does fail you can fix it unlike a modern digital jobbie.

My 7633 scope is nearly 40 years old and still going strong so its not that weak.

(and HOW is a storage scope useful for working on switchers,that a

Storage lets you run your supply for a few cycles, this prevents things melting during fault finding. You can also run intermittently on full load without the heatsinks fitted.

I have a digital as well but I still use the analoge scope for some things and the op hasnt got the money for one.

suffcient.Storage circuitry is more complexity to fail,and the storage CRTs are old

I've finally fully converted to digital LCD scopes. Last time I used my old 7104, I got confused because all the traces were the same color.

One cool thing about the digitals is that you can freeeze an interesting display, stick an explanatory post-it on the side of the screen, and shoot a picture for posterity.

John

The simple (easy to say, hard to quantify) answer is "whatever you need'. Depends on what frequency range you are working in. If the switching frequency is 1MHz then I think you'll want a 100+ MHz scope. Something with storage. Does not (IMHO) have to be digital. It really depends on 'scope model. I personally prefer analog scope and find lots of the digital scopes not as easy to use (hey, that's just my preferance: I'm an old guy).I must say tho I DON't miss having to drag out a scope camera when I use one of the newer scopes and need to snap a waveform picture.

More important is the brains behind the scope.

M Walter