Ask any questions relating to the PicoScope hardware or kit contents here.
If you have any questions prior to purchasing the kit post them here.
Pico Staff Member
Pico Staff Member
Posts: 314
Joined: Thu Aug 11, 2011 7:43 am
Location: Cambridgeshire

Re: TA375

Post by Martyn » Fri Dec 28, 2018 6:47 pm

Just a note considering a clock signal in isolation, it consists of 1 high bit, followed by 1 low bit per cycle so a 5MHz clock has 10MBits/s.
Technical Support Manager

Steve Smith
Pico Staff Member
Pico Staff Member
Posts: 816
Joined: Sun Aug 25, 2013 7:22 am

Re: TA375

Post by Steve Smith » Wed Jan 02, 2019 12:40 pm

Hello and thank you for the continued posts.

The older PicoScopes do not reduce the bandwidth when more channels are active (Automotive 3000 and 4000 series)
These scopes behave exactly as the 4425 used in the example waveforms, however their sample rate could not match that of the 4425 and sample rate is king.

There would be concerns with the display of captured data where the input signal approaches or exceeds the bandwidth of the scope for sure.
With serial data this would manifest itself as lowered edges and a more sinusoidal formation to what should be a near perfect squared edge (1 bit)

With that said, if we are predominantly interested in serial decoding and our sample rate is high enough to provide a clearly defined crossing point for the decoder to function (approx. 7-8 samples per bit) then we can get away with pushing the boundaries of bandwidth limitations thanks to software.

If the formation of the input signal and the capture of “noise” (that may be influencing the transmission of serial data) is paramount, then sufficient bandwidth and sample rate are required.

I hope this helps, take care…….Steve

User avatar
Steve Michener
Posts: 6
Joined: Wed Oct 31, 2018 5:56 pm

Re: TA375

Post by Steve Michener » Fri Jan 04, 2019 4:25 am

Thanks Martyn and Steve for the replies, all great information.

Pico Staff Member
Pico Staff Member
Posts: 211
Joined: Tue May 16, 2017 1:02 pm

Re: TA375

Post by ben.martins » Wed Nov 06, 2019 10:53 am

FlexRay with 4823 and TA375

As Steve has already outlined above, the 4225 and 4425 are more than capable of capture FlexRay and successfully decoding the data. With the addition of the 4823 to our scope line up and following a recent customer enquiry, we decided we should see how far you could push the 8-channel scope.

The 4823 has a max sample rate of 80MS/s when using 1-4 channels and as in the above posts, sample rate here is king in order for us to decode successfully. The following captures will describe how we can deliver as many packets of data as possible without using the slow sampling transition mode. You can find more on slow sampling here - https://www.picoauto.com/support/topic19921.html.

FlexRay along with many vehicle networks is a differential signal where we have data pushed high on one side and low on the other. These buses arrive at the various ECU’s which will remove the low from the high to reduce the impact of noise and interference that is expected in the hostile environment of the modern vehicle. If you have followed any of the recent CAN forum topics, case studies and the Practical Pico session on CAN you will have noticed we talk a lot about using Maths and in particular A-B. This allows us to reveal the ‘clean’ signal that the ECU gets to decode, on cancelling out anything that might interfere with the signal.

In order to achieve the 80MS/s sample rate and enough samples on the screen we have found the following settings will achieve the best results in order to decode using both the High and Low bus. This is as follows:
  • 1. Ensure you are using the TA375 Oscilloscope Probes switched to x10 and both are grounded to the vehicle chassis/battery negative
    2. Setup up Channel A and B to reflect the x10 setting on the probes
    3. Time base set to 10ms/div
    4. Sample rate 10MS, you can increase this but it doesn’t change the sample rate
    5. Channel A to FlexRay High Bus
    6. Channel B to FlexRay Low Bus
    7. Activate A-B maths channel
    8. Setup the FlexRay serial decode using the settings Steve has mentioned above, ensure you click the ‘invert’ option
    9. Wake up the bus or start the engine
    10. Run the scope
In the following examples the engine is running and the vehicle is being driven.
FlexRay 10msdiv, 80MSs High, Low and math with decode.png
The properties page can be opened by clicking Views from the top tool bar and will open allowing you to see information regarding the capture. Here we found that moving the sample rate moving to 20ms/div saw the warning error appear in decode table for ‘low number of samples’ and some packets no longer being decoded correctly. This is how we decided that 10ms/div is the best time base to stick to.

Using PicoScope zoom tool we can highlight one packet and visually see the structure. If there was any interference in both H and L this will be eliminated by the maths channel allowing a much cleaning signal to decode on.

Due to the amount of data present you will notice that the number of buffers has been reduced to 15. This will be the case even if you have a maximum number of buffers set much higher. However, we can see from the decode table that in just one buffer there are approx. 570 packets of data. If each buffer is 100ms total time, with 15 buffers we have an overall time of 1500ms = 1.5s. Therefore, in just 1.5 seconds we have captured approx. 8550 packets of successfully decoded FlexRay data.

This is all pretty impressive but we can go one step further. Looking at the structure of the data packets it maybe they are clean enough that we can decode successfully on just one channel. By removing a channel, we can use the additional memory to increase the number of buffers we have without compromising on sample rate. Using the same settings as before but switching off Channel B gave the following results.
FlexRay 10msdiv, 80MSs High bus with decode.png
From the properties panel we can still see that we have the max sample rate of 80MS/s and the structure of the packet is clean enough to decode on. With the time base remaining the same the only thing that has changed is the number of buffers. We have now jumped to 32! Again, as before this is as far as it will go but let’s review the statistics.

As before, each buffer contains approx. 570 packets of data but instead of 15 we have increased to 32 buffers. Time wise we now have 3200ms total captured time = 3.2s and in this brief period PicoScope has captured approx. 18,240 packets of FlexRay data and decoded it successfully!

Here you can quickly filter through the collected data for individual ID’s or data that hasn’t been decoded correctly in order to get some direction if you’re chasing a fault. If you are keen to look further into decoding networks then the property panel is a must and if you haven’t yet found in it PS6 it is located under Views on the top toolbar.

These files end up being massive with one of the captures I have being over 100MB. Please bear this in mind if you are taking many captures as it will quickly fill your PC/Laptop memory!

I hope this helps in some way.


Post Reply