Topics

Bandwidth vs rise time of digital 'scopes: Behavior of the TDS3054


 

Inspired by recent postings in this group, I've decided to describe my experience with a TDS3054 and a TDS3054B.

These 'scopes have a published bandwidth of 500 MHz on all channels.
The TDS3000-series Service Manual (071-0382-01) on page 1-3 mentions a *calculated* rise time of 700 ps.
This complies with BW = 350/tr (BW in MHz, tr in ns).
350 is the accepted number for analog 'scopes (Gaussian fall-off). Apparently, Tek didn't want to use the larger (400 - 450) value that makes sense for digital 'scopes with their much steeper fall-off.
I measured both TDS 'scopes to indeed have about 700 ps rise time on all four channels, with only a few percent overshoot, in line with the rise time as per the specifications (although mentioned there as calculated).

Two interesting things:

- Both TDS 'scopes have their -3 dB points clearly above 600 MHz on all four channels! If this applies to "all" samples of this model, it indicates that 400 - 420 would be more applicable than 350 but Tek decided in their sales documentation to do as if the number 350 that everybody knows from the analog age would apply, calculating BW from tr. The assumption "Nobody is interested in BW, rise time is what counts" wouldn't be surprising in these digital times.

- Tek sold a TDS3064 having a BW spec. of *600 MHz* with V/div. settings that nicely match my checking results for the two TDS3054's!

AFAIK, checking and (automatically) calibrating frequency behavior in these 'scopes is done by adjusting the step behavior.

Any comments?

Raymond


Bob Albert
 

What do you mean, adjusting the step behavior?

On Sunday, January 31, 2021, 02:51:21 PM PST, Raymond Domp Frank <hewpatek@gmail.com> wrote:

Inspired by recent postings in this group, I've decided to describe my experience with a TDS3054 and a TDS3054B.

These 'scopes have a published bandwidth of 500 MHz on all channels.
The TDS3000-series Service Manual (071-0382-01) on page 1-3 mentions a *calculated* rise time of 700 ps.
This complies with BW = 350/tr (BW in MHz, tr in ns).
350 is the accepted number for analog 'scopes (Gaussian fall-off). Apparently, Tek didn't want to use the larger (400 - 450) value that makes sense for digital 'scopes with their much steeper fall-off.
I measured both TDS 'scopes to indeed have about 700 ps rise time on all four channels, with only a few percent overshoot, in line with the rise time as per the specifications (although mentioned there as calculated).

Two interesting things:

- Both TDS 'scopes have their -3 dB points clearly above 600 MHz on all four channels! If this applies to "all" samples of this model, it indicates that 400 - 420 would be more applicable than 350 but Tek decided in their sales documentation to do as if the number 350 that everybody knows from the analog age would apply, calculating BW from tr. The assumption "Nobody is interested in BW, rise time is what counts" wouldn't be surprising in these digital times.

- Tek sold a TDS3064 having a BW spec. of *600 MHz* with V/div. settings that nicely match my checking results for the two TDS3054's!

AFAIK, checking and (automatically) calibrating frequency behavior in these 'scopes is done by adjusting the step behavior.

Any comments?

Raymond


 

On Mon, Feb 1, 2021 at 12:05 AM, Bob Albert wrote:


What do you mean, adjusting the step behavior?
I realize "step response" is the term commonly used. During adjustment/calibration, step response is made to follow the applied electrical voltage step as accurately as possible. Some overshoot is commonly accepted.

Raymond


Ed Breya
 

Another thing you may want to compare is the "effective bits" vs frequency. Ed


 

On Mon, Feb 1, 2021 at 01:02 AM, Ed Breya wrote:


Another thing you may want to compare is the "effective bits" vs frequency.
Sure, important and interesting. Is there a special relevance to the point I was trying to make? AFAIK, the architecture and hardware are the same across the whole family, incl. the TDS3064 600 MHz version.
Any suggestion re. how to find out? Basic resolution for these 'scopes is 9 bits. Max. RT sample rate is 5 GSa/s.

Raymond


n49ex
 

I have a couple of TDS3054Bs and have had one TDS3064A, and have done rise time and -3dB measurements on them. Just as others have mentioned, I saw around 700pS rise times and 600MHz -3dB points - on both the '54 and '64! Exactly the same! So, my guess is that Tek somewhere along the line decided to sell a 600MHz labelled version of what was essentially the same electronics because somebody said they could do so (legitimately by some measure), but then had second thoughts or maybe the engineers told the marketing folks to back off!

Reinhard


Ed Breya
 

Hi Raymond, I meant to look at effective bits as a general view on all DSOs. I don't know what the detailed differences are among the TDS scopes. I actually find it quite confusing with all the various models, since I'm not very familiar with them. One thing I do know, is that I "upgraded" my TDS754 (I think) a while back, to perform as a TDS784, using some mods I found online. As I recall, some BW limiting caps needed to be deleted from the amplifier circuits, and some jumpers needed to be changed to make the brain think it was the higher-end model. It appeared to work.

Ed


Jean-Paul
 

Ed and Raymnond:

HAve several Yokogawa scopes, with 500 MHz BW will try out and post images.

The transient response on analog scopes can be tweaked for best risetime but sime overshoot, and usually several small ripples.

Unclear if digital scopes can be adjusted for this.

Bon journee,


Jon

PS: the number bits referred to as "effective number of bits " ENOB " may include a few "marketing bits" (:-:) as per Bruce Jackson, Apogee Electronics


Jean-Paul
 

Hello all just checked the Yokogawa DL7440, with Leo B 40 ps BNC pulser, get 570 pS RT, OS ~ 14.5%

Using HP RG gen 8640B , Gen Rad RG/6U cable and 20 db atten, getting 3 db at 607 MHz.

0.35/Tr gives 614 MHz so very close.

Will post screenshots in photos soon, under DL7440 album

Bon journée,

Jon


Jean-Paul
 

Correction, the cable used for BW test,

GenRad type 874 connectors, 75" RG214/U

(Not RG/6u 75 Ohm !)

Jon