Tektronix TDS3054

[eurofox]

It rain TDS3054 in Australia   

http://www.benl.ebay.be/sch/i.html?_sacat=0&_from=R40&_nkw=TDS3054&LH_PrefLoc=2&_arm=1&_armm=63&_ruu=http%3A%2F%2Fwww.benl.ebay.be%2Fsch%2Fi.html%3F_sacat%3D0%26_from%3DR40%26_nkw%3DTDS3054%26_arr%3D1

eurofox

[nctnico]

I noticed that too. Most have the logo from the Australian army so I guess they are clearing these out. Nice scopes but for that money you can buy something new with more memory. If you just need 500MHz then you can buy a TDS500, 600 or 700 series for peanuts.

[pascal_sweden]

Is the sample rate really 5 GS/s on each channel, even when 4 channels are used at same time?
That is quite impressive!

Of course the memory is not very deep.

On the new Rigol DS4000 series, the sample rate is 4 GS/s, but my understanding is that it is only 1 GS/s if you use all 4 channels. Is that correct?

How come the Tektronix does do it irrespective if you use all channels or not? They don't do any multiplexing than or?

[MarkL]

Yes, the sample rate is 5GS/s on all channels, all on at the same time.  It's not multiplexing, at least not externally between the channels.

It was an impressive scope when it was introduced in 1999 (or thereabouts).

[Nermash]

Ever since Dave's teardown of the TDS3054 I am 95% convinced that Tektronix uses equivalent time sampling as the 5 GS/s figure...

I have no real evidence of that, nor any experience with TDS3054, but looking at many teardowns of the modern scopes (say from 2010 onwards) with 8 bit ADCs at speeds of 2 GS/s or more, I have never seen one without a heatsink on the ADC!

Comparing that  to 10 years older ADC technology in TDS3054, which has 1 ADC per channel, claiming 5 GS/s @ 9 bits!, either Tek had some magic chips 10-20 years ahead of everything else, or IMHO they used marketing wank and went with equivalent sampling speed figure.

I am by no means expert in ADCs, and possibly I am wrong,so  I really would like to hear some resident gurus share their thoughts.

[eurofox]

Ever since Dave's teardown of the TDS3054 I am 95% convinced that Tektronix uses equivalent time sampling as the 5 GS/s figure...

I have no real evidence of that, nor any experience with TDS3054, but looking at many teardowns of the modern scopes (say from 2010 onwards) with 8 bit ADCs at speeds of 2 GS/s or more, I have never seen one without a heatsink on the ADC!

Comparing that  to 10 years older ADC technology in TDS3054, which has 1 ADC per channel, claiming 5 GS/s @ 9 bits!, either Tek had some magic chips 10-20 years ahead of everything else, or IMHO they used marketing wank and went with equivalent sampling speed figure.

I am by no means expert in ADCs, and possibly I am wrong,so  I really would like to hear some resident gurus share their thoughts.

I bough 10 years ago a brand new 300Mhz version from this scope and sold it because I din't like it but of course it is a personal opinion.
I was using before a Tek2465B, I know people will again say don't compare analog with DSO .....

I have a DSO now from Agilent and I'm very happy with it.

[MarkL]

Ever since Dave's teardown of the TDS3054 I am 95% convinced that Tektronix uses equivalent time sampling as the 5 GS/s figure...

I have no real evidence of that, nor any experience with TDS3054, but looking at many teardowns of the modern scopes (say from 2010 onwards) with 8 bit ADCs at speeds of 2 GS/s or more, I have never seen one without a heatsink on the ADC!

Comparing that  to 10 years older ADC technology in TDS3054, which has 1 ADC per channel, claiming 5 GS/s @ 9 bits!, either Tek had some magic chips 10-20 years ahead of everything else, or IMHO they used marketing wank and went with equivalent sampling speed figure.

I am by no means expert in ADCs, and possibly I am wrong,so  I really would like to hear some resident gurus share their thoughts.

How about letting the scope speak for itself?

Below is a screen shot (single_pulse_single_seq.png) of all channels on, single trigger, and to make sure it wasn't cheating, I sent it a single pulse.  This is at 1ns/div in dots mode.  You can count the dots for yourself.  I doubt they're creating any false dots with interpolation or other nonsense.

And putting both the pulse generator and the scope in continuous run you get the second screen shot, continuous.png.  There's virtually no quantization in the time axis, so they're keeping track of the time on an even finer scale.

Call it magic if you want, but I give a lot of credit to the Tektronix engineers who designed it.  They're STILL selling the same acquisition engine after 15 years!

[nctnico]

When designing a DSO fast sampling is really easy. Storing the samples is where you run into a memory bandwidth bottleneck. Which is why the TDS3000 series sample fast but have really short memories.

[David Hess]

Ever since Dave's teardown of the TDS3054 I am 95% convinced that Tektronix uses equivalent time sampling as the 5 GS/s figure...

5 GS/s would be awfully slow for equivalent time sampling but was possible in real time digitizers through interleaving at the time.

I have no real evidence of that, nor any experience with TDS3054, but looking at many teardowns of the modern scopes (say from 2010 onwards) with 8 bit ADCs at speeds of 2 GS/s or more, I have never seen one without a heatsink on the ADC!

Comparing that  to 10 years older ADC technology in TDS3054, which has 1 ADC per channel, claiming 5 GS/s @ 9 bits!, either Tek had some magic chips 10-20 years ahead of everything else, or IMHO they used marketing wank and went with equivalent sampling speed figure.

I am by no means expert in ADCs, and possibly I am wrong,so  I really would like to hear some resident gurus share their thoughts.

Tektronix was making their own custom ASICs including ADCs through their fab which they sold to Maxim who continued to fabricate custom designs for Tektronix.

If I had to make a guess based on the information available and earlier designs, I would estimate that its equivalent time sampling rate is 50 GS/s minimum and that its time to digital converter has a resolution to support a sampling rate twice as high.

[prot_brot]

Just got mine for 350 bucks . Finally an upgrade for my ancient 10MHz analog scope 

The screen has faded and date/time doesn't run when the scope switched off. But all channels works and everything else seems to work just fine.

[MarkL]

I replaced my TDS3054 CCFL tubes because the display was so dim.  Part details here:

  https://www.eevblog.com/forum/testgear/tek-tds-3054b-vs-tds-744a/msg539810/#msg539810

Takes an hour or two.