New (Old) DSO scope

[crustytv]

Well I have to admit I'm a bit of a fan of Tektronix scopes and thoroughly loved my old analogue 2205 and 2225 . Though they have been consigned to the shelf for over 18 months now after getting my first Digital scope the OWON. Although only 25MHz it proved adequate for the simple task it has been asked to perform, namely helping diagnose faults on my vintage colour telly's.Also as a first foray into the world of DSO not a bank breaker.

However ever since getting the OWON and really liking digital I've hankered after a TEK DSO, of course this has always been a non starter due to the astronomical prices ( have you seen the prices!!! we're talking £1000+ new and that's fairly entry level) of new ones. Second-hand ones are equally very expensive and an even more amazing fact is non workers even command high figures. Recently an opportunity arose to get a "non worker" and although a high price for a potential dud at £120, I felt if I didn't take the plunge now I never would. The logic being although it could be a total waste there was a chance I might be able to repair it....but yep it was a gamble. With no guarantees it would work I set about finding out what sort faults these scopes might be known for and to try importantly track data ( cct) to assist. That's when I found there isn't a great deal out there, oops! It seems a lot of faults were recorded against the attenuator boards, sticky relays and broken BNC inputs due to forceful handling. As for a circuit so far nothing.

I'd also seen a video on the EEVblog where Dave had bagged about 9 TDS 220's at an auction, their only sin was to be out of date, out of calibration and the labs they came from just binned due to that. The auction had to sell as non working but in all cases they did work. I started to dare hope this might be the case of mine but I'm never that lucky and felt sure I was in for a fairly tortuous repair. The last scope I repaired was the 2205 that needed a donor CRT with some rewiring, some fairly mind blowing fault finding for me ( helped by the very good people on UKVRR) and then finally setting it all up.

Anyway it arrived, what is it? The scope is a TDS1002 60MHz 1GS/s and with FFT. I believe the scope dates from around 2002 but might be a little earlier. The scope has a calibration sticker for 2010 along with a PAT test of the same year. Powered up and was presented with a welcome screen, well that's a result in itself! Checking the error log to my absolute delight and amazement, the scope had only ever been power cycled 326 times in approx 14 years and with zero errors, result No.2. Now for some tests, I hooked up the pattern gen to channel 1 then 2, both produced nicely triggered waveforms. I've done various checks with various signal sources and compared the results with both the OWON and my analogue TEK 2225, its all checks out. Also the BNC inputs are rock solid not suffering from any wobble/movement. I think its reasonable to assume this is a very low hour, low use scope.

Compared to the OWON, as you would expect the TEK has so many more bells and whistles to play with along with the FFT ability allowing you look in the frequency domain not just the time domain, a feature the OWON did not have. Of course the big benefit is it being 60MHZ with 1GS/s whereas the OWON was only 25MHz 100MS/s. Itching to put it through its paces so I may just get a rather horrid old 3000 power supply module up on the bench and give a workout. All in all makes for a very lucky buy and a very happy old crusty.

Colour Bar

Hardly used and zero errors

The Crusty Gen feeding in a sine wave to the TDS1002

Time domain view of the signal

Frequency Domain (FFT) mode showing the spectral view of the same signal. Although the scope is only 60MHZ with FFT it can measure frequencies many times higher than that and right down low too.

[Cathovisor]

I think you can count that as a result

Does it have a printer/USB option?

[crustytv]

No it didn't come with the rear expansion module with serial and GPIB etc. Will have to see if they are available on the second-hand market at a reasonable price. Not sure if I really need it or would make use of it though other than a nice to have.

[sideband]

Oooh you lucky whatsit!! We use those and the 220 at work and they are brilliant scopes.

[Cathovisor]

No it didn't come with the rear expansion module with serial and GPIB etc. Will have to see if they are available on the second-hand market at a reasonable price. Not sure if I really need it or would make use of it though other than a nice to have.

TBH it's not worth it, unless someone's written a bit of software that allows you to "print" to a computer screen - someone has done this for the HP audio analyser much loved of Valvebloke that "plots" to a graphic file.

I can't remember if they also had a battery option, which IMO would be more useful.

[Lloyd]

I used to use these when I worked in mobile phone repairs! I've been after one for ages, but never see them at the right price. I stupidly missed out on a potentially free one when the lift company I worked for closed down, they were letting us take whatever we wanted, I asked for an ultrasonic cleaner and a bench power supply, but thought it would be greedy to ask for the scope too. I bet it went in the skip with everything else

Regards,
Lloyd

[Cathovisor]

There are occasions though when an analogue scope is best. An XY trace when sweeping an IF strip comes to mind.

[BriancuffRest In Peace]

That didn't come to mind because of the BT8161 did it Mike??

[Cathovisor]

That didn't come to mind because of the BT8161 did it Mike??

Might have done....?! (we don't have a whistling innocently emoticon!)

Actually, it was in connection with a Dynatron radio chassis with variable selectivity I need to align; plus I don't have the space for things like a Knott Polyscanner or a R&S Polyskop!

The thing is, my LeCroy only works in 'dot' display mode with a fixed sample clock of 1Ms/sec in XY mode. It does however still offer variable persistence, which is useful at slow sweep speeds.

[crustytv]

Might have done....?! (we don't have a whistling innocently emoticon!)

Do now

[Cathovisor]

[crustytv]

Meanwhile back to the TDS1002

TBH it's not worth it, unless someone's written a bit of software that allows you to "print" to a computer screen - someone has done this for the HP audio analyser much loved of Valvebloke that "plots" to a graphic file.
I can't remember if they also had a battery option, which IMO would be more useful.

Well after a bit of research it appears Tektronix did just that. Its called *OpenChoice it came with the +module but I believe it was/is available to download from their site.

Only problem is the serial/GPIB interface card/cables are about as expensive as the scopes when new. Unless one turns up el cheapo (doubtful) then that's a option I won't be having. There isn't a battery option.

*OpenChoice ™ PC
• TDSPCS1 OpenChoice PC Communications Software included with each Module
Communications
• Seamless connection from oscilloscope to PC through GPIB and RS-232
Software
• Transfer and save settings, waveforms, measurements, and screen images
• Includes a Windows desktop data transfer application in addition to convenient Microsoft Word and Excel Toolbar Add-ins

+Module
•Printer Port Centronics-type:Parallel
•Printer Capability: Graphics File Formats – TIFF, PCX (PC Paint Brush), BMP (Microsoft Windows), EPS (Encapsulated Postscript) and RLE,Printer Formats – Bubble Jet, DPU-411, DPU-412, DPU-3445,Thinkjet, Deskjet, Laser Jet, Epson Dot (9- or 24-pin)* 2 , Epson C60* 2 , Epson C80.Layout – Landscape and Portrait
•RS-232 Programmability: Full talk/listen modes. Control of all modes, settings and measurements. Baud rate up to 19,200. 9-pin DTE RS-232 Cable (012-1651-00) included
•GPIB (IEEE std. 488-1987): Full talk/listen modes. Control of all modes, Programmability settings and measurements.

[Michael Watterson]

OWON was only 25MHz 100MS/s.

So more like a 3MHz to 6MHz scope, compared to analogue.

This DSO is about a minimum spec for a digital scope. So a good find. Pity it has no interface. Try Helmut Singer (Germany), online. I've bought quite expensive stuff and cheap stuff there.

X - Y operation is where all the cheaper DSO fail. I've no idea how good this one is at X Y.
XY is good for filter alignment and also four quadrant component testing.

Make a noise source by 6MHz to 50MHz band pass filter on a zener, then amplified. Feed filter or IF with that and use FFT/Spectrum mode to align filters/IFTs etc.

Serial better option than GPIB unless you get a bargain on a PC GPIB too, as they are very expensive. Even the GPIB cables are dear now.

[Cathovisor]

I was lucky enough to get a reasonably cheap and genuine HP/Agilent PCI GPIB card so it shouldn't cause any problems as some cards are wont to do. Primarily, to interface to my HP audio analyser to provide the graphical 'plot' function when I do frequency response and distortion tests - there is a nice free suite of programs for this.

It'll be interesting to see what kind of a display I get from the LeCroy in XY mode when I do a sweep with it. Maybe I should dig out the Kikusui display scope I have that partners with a Glensound frequency response measuring set I own.

[Mark Hennessy]

OWON was only 25MHz 100MS/s.

So more like a 3MHz to 6MHz scope, compared to analogue.

Over-simplification alert

Digital oscilloscopes work in several modes. The main two are real-time sampling and equivalent-time sampling. There are variations on the theme, naturally...

When in real-time sampling mode, then yes, the higher the sample rate, the better. But you only really need this for single-shot captures.

For repetitive waveforms, equivalent-time sampling builds up a "picture" of the waveform over many sweeps, giving excellent results.

When you think about it, an analogue scope inherently requires a repetitive waveform to give you a stable trace, so there's no "loss" when using a DSO in ET sampling mode. Obviously ignoring analogue storage, which has limits on write speed, etc.

Quite a good article here: http://www.tek.com/document/application ... e-sampling

The real power of a DSO is for single-shot capturing. For that, sample memory is as important as sample rate. Old 'scopes are severely limited in their memory (fair enough - it has to be fast to store the samples, and that's expensive). These Teks only have 2,500 sample points, which severely limits the usefulness of it when doing digital investigations. The Rigol DS1054Z has 24Mpts, which is amazing - the ability to freeze and pan and zoom on a signal is just brilliant. I don't know how they do it for the money.

As well as the Rigol, I have an old Fluke Combiscope; both an analogue and digital 'scope at the touch of a button. The digital side was a bonus; I didn't buy it because of that. Now, despite being only 100MS/s, it's a 100MHz 'scope that works beautifully for repetitive waveforms. Single-shot captures are quite limited - the panning and zooming facilities are practically non-existent (hardly any memory, but still more than the Tek, which surprised me!), so it's essentially just a way to freeze the current waveform. But, the measurements are useful - that's another big advantage of DSOs.

[Cathovisor]

My LeCroy WaveAce 2024 is 1 Mpts/Ch (2 Mpts interleaved). So some way behind the Rigol you mentioned!

The other features I find useful on a DSO are infinite persistence - catch an intermittent jitter on a waveform; and the noise reduction capability. Plus the ability to pick out one burst in an eight field sequence!

I think the Tek Chris has bought is already showing its worth; there are times though when you really don't want that 'sampled look' on a waveform and the 2445B (or the Hitachi) comes out

[Mark Hennessy]

My LeCroy WaveAce 2024 is 1 Mpts/Ch (2 Mpts interleaved). So some way behind the Rigol you mentioned!

All these numbers are confusing

Your 'scope is 1GS/s (samples per second); that's the same as the Rigol. The interleaving feature is interesting - I wonder if that means that there are 2 ADCs shared between the 4 channels? On the Rigol, there is only one ADC, so the sample rate comes down as you turn on channels. In practice, it only needs the higher rates at higher timebase speeds anyway.

Where the Rigol does score here is the memory depth. It's 12Mpt by default, or 24Mpt if you pay for the upgrade (or hack it!). Whereas according to website the LeCroy appears to be only 24kpt. I'm pretty sure that's a mistake. The brochure suggests 1Mpts/channel, which is more like it, but still quite a way behind the Rigol.

But nothing is ever straightforward! The Rigol uses commodity DDR chips to get the memory size at such low cost, but these are relatively slow. There's every chance the LeCroy is using faster memory, but cost means a limit on the size. But the tradeoff here is memory size vs acquisition speed, and perhaps the Lecroy will do more updates per second? Certainly, the Keysight 'scopes are very fast, but have limited memory depths because rather than separate memory chips, the sample memory is integrated in the same ASIC that deals with everything else.

Update rate vs memory depth is yet another balancing act with DSOs that we need to understand on top of the analogue "basics". Another example of a "new" DSO-only thing is the acquisition mode. I wouldn't claim to be fully automatic in these areas yet, but it's true that DSOs have more ways to lie to you if they choose!

At least modern DSOs seem to be free of aliasing now. We have an older OWON that aliases for fun - feed it a 1kHz sine wave and it'll show you almost any frequency you like. This is because the ADC's sample rate is tied to the timebase speed rather too closely - the limited sample memory making that necessary. No idea if that applies to Chris's model, but it's certainly a trap for the unwary. That's where having a dedicated hardware counter built into the 'scope is a good idea (the Rigol manages that). The more basic models can only measure frequency based on analysing the samples in memory, which is no good if they're mostly aliases!

[Cathovisor]

I wouldn't claim to be fully automatic in these areas yet, but it's true that DSOs have more ways to lie to you if they choose!

There are times when I reach for the analogue 'scope because I don't believe what the digital one says!

I do think though the record length is only 24kpts.

[Mark Hennessy]

I do think though the record length is only 24kpts.

I've just looked at the "Fact Sheet", which shows a table of specifications per model. Now I see that you're right

So, 1 or 2Mpts is available on the 1000-series models, but once you get to the 2000-series models, it drops to 12 or 24kpts. Shows that you have to look behind the "Up to 1Mpts" headline...

Quite surprising in a way, but I guess that suggests that something similar to what I was saying earlier is happening - bespoke (fast) hardware compared to generic DDR memory on the cheaper models. But given how different the innards of the 1000 and 2000 series models seem to be, it's a little surprising that they put them in the same datasheet.

Whatever the numbers, they're nice machines...

[Cathovisor]

They do things a little different to the Tektronix way - video triggering can be fun - but one thing that is interesting is the ability to drive the scope and capture waveforms over a network.

I'm slowly getting used to the foibles as I'm new to DSOs but we'll get there

[Dr Wobble]

Digi scopes do indeed look tempting, did you find yours easy to drive Chris? There are several TDS scopes on ebay at the mo for around £200, which is a reasonable price. Don't have 200 quid though and have more than enough scopes needing attention.

Andy.

[Mark Hennessy]

Digi scopes do indeed look tempting, did you find yours easy to drive Chris? There are several TDS scopes on ebay at the mo for around £200, which is a reasonable price. Don't have 200 quid though and have more than enough scopes needing attention.

Andy.

Whatever you do, do not spend £200 on a second-hand DSO when you can buy a Rigol DS1054Z for ~£300 brand new with warranty, probes, USB/Ethernet, intensity grading (what Tek call "digital phosphor"), extremely deep sample memory, etc, etc. DSOs have really changed a lot in the last few years. If you don't have £300, then it's worth waiting until you do. Honestly.

Though if you find one like Chris did for £120, then that's a different matter

At the bottom end of the market, be really careful about anything other than the DS1054Z - there are lots of competing models that come nowhere close to what you get from the Rigol.

All that said, if you mostly do audio, you probably don't need a DSO. A good analogue 'scope is actually much better for audio work. The FFT functionality of a DSO might seem tempting to an audio person, but they are universally no good for audio because DSOs only have 8 bit ADCs - you'd get much better results from your PC soundcard (even a crap one) and free software.

Where DSOs really shine is non-repetitive stuff. Anything digital is an absolute joy to troubleshoot with a DSO with a deep sample memory.

[Cathovisor]

Some audio-frequency traces from my DSO - which show the limitations of an 8-bit ADC. Both are of a 440Hz sine wave from my Ascel function generator.

Waveform at normal sweep rate

Waveform at slower sweep rate

The quantising is very obvious in the lower speed sweep; the upper one shows how the sweep speed affects the number of samples per second.

FFT function on the Ascel at 465kHz.

This last trace shows the FFT function in use.

[peterscott]

There are times when I reach for the analogue 'scope because I don't believe what the digital one says!

The digies are lovely but you certainly have to rotate the sweep rate control just to check that you aren't looking at an aliased picture.

Peter

[Mark Hennessy]

That's definitely true!

Some DSOs alias more than others. With the Rigol, you have to try really quite hard, but some others do it for fun

There's usually a way to get around it though: using the hardware frequency counter.

All DSOs can display frequency as a measurement (along with countless other parameters). But, these measurements are based entirely on the waveform shown on the screen - the samples held in memory. Normally, that's not a problem, but if there's aliasing because the timebase speed is inappropriate (or the signal is clipping, for amplitude measurements), then that's where things go bad...

The hardware counter is a separate circuit that measures the frequency in exactly the same way a dedicated frequency counter would, and always gives accurate results, whatever is on the screen.

On my Rigol, it's enabled all the time. It hardly takes up any space on the screen, and there's no reason not to leave it going. In addition, I often have frequency enabled as a measurement, and providing those two numbers broadly agree, then I know there's no aliasing.

Not all DSOs have a separate HW counter, so that's something that's worth looking out for when buying. It's not a show-stopper as dedicated counters are plentiful and cheap, but that's another box to accommodate, and another set of connections/terminations to consider. Of course, if your 'scope has a Y-output, that's a good use for it.