Test and Measurement Forum

Picoscope not delivering to specification??

Forum for discussing PicoScope version 6 (non-automotive version)

Picoscope not delivering to specification??

Postby Gustav » Fri May 22, 2015 4:04 pm

Hi

I have a 5244B, or at least that is what I thought I bought...

According to specification 5244B should have a Maximum sampling rate of 250 MS/s (two channels, 8bit) and memory of 512 MS.

When I try to use memory, Picoscope 6 will not go any higher than 25MS/s and 125Ms, for the 5 sec time span that I need.

In all fairness, the sample rate can go up to the specified rate, if the time span is reduced. But I need a 5 sec time span, and I think that it must be a bug in Picoscope 6.
Attachments
picoscope5244B.png
Gustav
Guru
Guru
 
Posts: 63
Joined: Sat Jan 23, 2010 1:33 pm

Re: Picoscope not delivering to specification??

Postby alan » Sat May 23, 2015 5:54 am

Hi,

The 5244B has a maximum sampling rate of 1 GS/s and 512 MS buffer memory. This is shared between the channels so with both channels enabled you have 500 MS/s per channel and 256 MS buffer memory.

Its not possible to sample at the full speed for 5 seconds since 500 million samples per second x 2 channels x 5 seconds = 5 billion samples. This is more than the memory buffer can store (BTW as far as I know, no oscilloscope on sale has a 5 GS buffer memory so no scope can sample at this speed for this long, most are much smaller than the 512 MS you have).

At 5 seconds and 2 channels across the screen your 256 MS per channel can give you 50MS/s sampling max, to get this you will need to use single shot trigger otherwise the scope will use half the memory for the trace you can see on the screen and half to collect the next trace. That is why you are seeing 25 MS/s not 50 MS.

Regards
Alan
alan
Site Admin
Site Admin
 
Posts: 171
Joined: Thu May 25, 2006 8:02 am

Re: Picoscope not delivering to specification??

Postby Gustav » Sun May 24, 2015 6:27 pm

Hi Alan,

Thanks for the hint about using single shot trigger, that did really double the available memory. But there is still a way to go to get the performance of the advertised performance .

With single shot my PS5244B can now delivers 41.67MS/s (not 50MS/s). And it is able to use 208 MS (not 256 MS) of memory.

You wrote "BTW as far as I know, no oscilloscope on sale has a 5 GS buffer memory "... I did not ask for 5GS buffer memory. Just the advertised 256MS.

And even with single shot trigger, the scope is still underperforming some 20% on sample rate and total memory capacity.
Attachments
picoscope5244B.png
Gustav
Guru
Guru
 
Posts: 63
Joined: Sat Jan 23, 2010 1:33 pm

Re: Picoscope not delivering to specification??

Postby alan » Mon May 25, 2015 10:51 am

Hi,

Its a public holiday in the UK today and I do not have a PicoScope 5000 to check with. Sorry for the delay but I will ask someone to check this and reply on Tuesday.

Regards
Alan
alan
Site Admin
Site Admin
 
Posts: 171
Joined: Thu May 25, 2006 8:02 am

Re: Picoscope not delivering to specification??

Postby Martyn » Mon May 25, 2015 12:08 pm

You need to look carefully at the sample interval as this will define how many samples you get for the chosen time across the screen.

The allowable values are defined by the formula defined on Page 18 of the Programmers Manual, https://www.picotech.com/download/manuals/ps5000abpg.en-1.pdf, which gives sample interval options of 1, 2, 4, 8, 16, 24, 32 ns etc These are fixed by the hardware of the scope.

You have chosen a screen view of 5s which means at the 24ns Sample Interval, as shown on the properties panel, you can get 208MS across the screen. The next fastest sample interval of 16ns would give 312MS which is not possible with two channels enabled.

The PicoScope software is designed to give as many samples as possible, for the chosen screen time, and will select a sample interval according to your selections.

If you use the SDK and write your own application you are not restricted to a set time and so can collect samples using the full scope memory.
Martyn
Technical Specialist
Martyn
Site Admin
Site Admin
 
Posts: 2355
Joined: Fri Jun 10, 2011 8:15 am
Location: St. Neots

Re: Picoscope not delivering to specification??

Postby Gustav » Mon May 25, 2015 4:47 pm

OK, I understand. The apparent "underperformance" is caused by
1) a fixed set of possible sample rates
2) a fixed set of possible screen time.
3) The size of the memory buffer.

Obviously, 1 and 3 are real limits, that you can not get around.

In general I am very happy with my PicoScopes (have got two) and PicoScope6. I have no intention of writing an application with the SDK.
Gustav
Guru
Guru
 
Posts: 63
Joined: Sat Jan 23, 2010 1:33 pm

Re: Picoscope not delivering to specification??

Postby Martyn » Wed May 27, 2015 7:49 am

I will put your comments to the development team for discussion.
Martyn
Technical Specialist
Martyn
Site Admin
Site Admin
 
Posts: 2355
Joined: Fri Jun 10, 2011 8:15 am
Location: St. Neots

Re: Picoscope not delivering to specification??

Postby springburo » Mon Jun 22, 2015 10:05 am

I have the same issue with a 5443A it is specified as having 64ms, so I assume with 4 channels active this is 8ms/ channel in 8 bit mode. Using 4 channels, 20ms/div and single shot trigger the properties shows a mximum of 5,000,000 samples at at rate of 25MS/s. Even going to a single channel the limit seems to be 5,000,000 samples at 25MS/S

How Can I access the rest of the memory ? :?:
springburo
Active User
Active User
 
Posts: 13
Joined: Mon Jun 08, 2015 5:24 pm

Re: Picoscope not delivering to specification??

Postby Mark_O » Tue Jun 23, 2015 10:41 am

springburo wrote:How Can I access the rest of the memory ? :?:

Change to a slower timebase. Or increase the sample rate.

At 20 ms/div, that's 1/5 sec full screen. At 25 MS/s, that's 5 MS, exactly what you are seeing.

(Oh, and ms is milliseconds, and MS is mega-samples. Also "a 5443A it is specified as having 64ms, so I assume with 4 channels active this is 8ms/ channel in 8 bit mode", that would be 16 MS/channel, not 8. It only drops to 8 MS/chan in 12bit+ modes.)

Example1: changing from a 20 ms to 50 ms/div timebase increases the acquisition time to 1/2 sec, and thus 12.5 MS of the 16 MS available per channel.

Example2: At 25 MS/s the sample interval is 40 ns. You could decrease that to 32, 24, or 16 ns (incremental steps of 8 ns). At 16 ns, that's 62.5 MS/s, and thus 12.5 MS would again be utilized.

As long as the display area is locked to 10 horizontal divisions, there will always be some slack space (unused sample memory), unless the combination of timebase and sample rate just happens to line up properly.
Mark_O
Advanced
 
Posts: 44
Joined: Fri Oct 03, 2014 5:58 am

Re: Picoscope not delivering to specification??

Postby springburo » Wed Jun 24, 2015 3:46 pm

sorry about the bad syntax.
I tried setting the max number of samples to 10MS with 4 channels, single trigger and 5ms/div. The properties window shows 1562500 samples and 31MS/s. If I reduce the time base the sampling rate goes up, but the number of samples is 1250000.

I tried your examples of increasing the time base to 50ms, set the number of samples to 10MS, single trigger 4 channels, press start and the number of samples is set back to 2MS

I do not seem to be able to increase the number of samples above 1562500 , everytime I press start, the number of samples in the picoscope menu is reduced to 2MS. What am I doing wrong? :(
springburo
Active User
Active User
 
Posts: 13
Joined: Mon Jun 08, 2015 5:24 pm

Re: Picoscope not delivering to specification??

Postby Martyn » Wed Jun 24, 2015 7:21 pm

Do you have any Maths Channels, Filtering, of Resolution Enhancement enabled as this will reduce the number of samples to 2MS.

For information the new Bets software has upped this figure to 5MS whilst also improving the processing speed.
Martyn
Technical Specialist
Martyn
Site Admin
Site Admin
 
Posts: 2355
Joined: Fri Jun 10, 2011 8:15 am
Location: St. Neots

Re: Picoscope not delivering to specification??

Postby Mark_O » Thu Jun 25, 2015 1:02 am

Martyn wrote:Do you have any Maths Channels, Filtering, of Resolution Enhancement enabled as this will reduce the number of samples to 2MS.

Good to know! So using a Maths channel to achieve interchannel time-correlation (for example) is not very productive. Someone could spend a lot of money to get a large memory capacity, only to wind up neutralizing it completely.

It may be worth pointing out that this type of processing can be disabled during live acquisition, and then re-enabled after stopping acquisition (during analysis), whilst retaining the full memory depth. At least, I believe that is the case.

For information the new Bets software has upped this figure to 5MS whilst also improving the processing speed.

This had me scratching my head for a second. Who is "Bets"? :) Then I realized you must have meant "Beta".
Mark_O
Advanced
 
Posts: 44
Joined: Fri Oct 03, 2014 5:58 am

Re: Picoscope not delivering to specification??

Postby bennog » Thu Jun 25, 2015 6:57 am

Mark_O wrote:It may be worth pointing out that this type of processing can be disabled during live acquisition, and then re-enabled after stopping acquisition (during analysis), whilst retaining the full memory depth. At least, I believe that is the case.


This is the way I use the math functions (and also serial decoding functions)

It would be nice to have an option to disable math and serial decoding during live capture and automatically re-enable after pressing the stop button.

This would save me a lot of time !

Benno
bennog
Guru
Guru
 
Posts: 79
Joined: Mon Nov 26, 2012 9:16 am

Re: Picoscope not delivering to specification??

Postby springburo » Thu Jun 25, 2015 7:06 am

Thanks with the beta version I can get the 5MS. It seems strange that then maths affects the memory so much. Is this because the maths is done in real time instead of post processing.?
Are there plans to give the full memory with the maths function. ?
springburo
Active User
Active User
 
Posts: 13
Joined: Mon Jun 08, 2015 5:24 pm

Re: Picoscope not delivering to specification??

Postby springburo » Thu Jun 25, 2015 7:14 am

I agree an option for the maths to be real time of post processing would be a good work around and cannot be too difficult to do.
Can this be done using macros, so when the trigger is enabled, the maths is turned off and then after capture reanabled ?
springburo
Active User
Active User
 
Posts: 13
Joined: Mon Jun 08, 2015 5:24 pm

Next

Return to PicoScope 6 for Windows

Who is online

Users browsing this forum: No registered users and 0 guests