PicoScope 7 Software
Available on Windows, Mac and Linux
Memory and USB speed confusion
Memory and USB speed confusion
Hi,
Just wondering if someone can clear up some confusion on my side
We actually use some expensive (and oversized) picoscope units at work "as a scope". But now im looking to buy one for myself as i need to do some stresstesting and analysing of digital signals (and signal integrity/shape under external influence). So i will need analog readouts of my digital signals at the same time as i need to capture my digital signals to be decoded to verify:
- If the sent data is correct (digital decoding)
- if the signal is clear and not distorted (analog analysis)
- Preferably trigger once capture first and analyse later as setting up these digital signals to occur will be hard to continuously reproduce to fit a "normal scope operation"
BUT since picoscope is made like a scope i need to be able to one-shot capture the entire signal at a rate any signal distortions are clear. Since i need to analyse low Mhz signals im thinking of getting a 50MHz 2000b or 3000 series scope. However i am very confused:
The 2000 series have a usb2 connection only allowing transfer of 10MS data according to the datasheet, so not the ~20Mhz bandwidth (200MS data) i need, the buffer only allows for 32MS (50MHz model) / 2ch = 16MS = 32ms of data, that means if my consecutive signals are longer than 32ms i will not be able to capture them? or how is this handled/what will happen if i try to capture more? does it just skip samples then? does it just stop recording? Does it just always limit to the 10MS sample rate once you ask more than the buffer size? What if i use the sdk, how does that handle memory or transfer speeds? If that's not possible, why are there 2xxx series scopes with 1GS speeds, you can never transfer those in time can you, it would take 1s to transfer 10ms of data?
The 3000 series has USB3.0 but according to the datasheet still only supports 17MHz transfer, still way too low. Does that mean i can't use either? or i need to grab a way bigger model with way more memory?
Im not shy of a bit of coding, im a programmer as well, so if the SDK can definitely do this then i can make it happen.
Thanks in advance
Just wondering if someone can clear up some confusion on my side
We actually use some expensive (and oversized) picoscope units at work "as a scope". But now im looking to buy one for myself as i need to do some stresstesting and analysing of digital signals (and signal integrity/shape under external influence). So i will need analog readouts of my digital signals at the same time as i need to capture my digital signals to be decoded to verify:
- If the sent data is correct (digital decoding)
- if the signal is clear and not distorted (analog analysis)
- Preferably trigger once capture first and analyse later as setting up these digital signals to occur will be hard to continuously reproduce to fit a "normal scope operation"
BUT since picoscope is made like a scope i need to be able to one-shot capture the entire signal at a rate any signal distortions are clear. Since i need to analyse low Mhz signals im thinking of getting a 50MHz 2000b or 3000 series scope. However i am very confused:
The 2000 series have a usb2 connection only allowing transfer of 10MS data according to the datasheet, so not the ~20Mhz bandwidth (200MS data) i need, the buffer only allows for 32MS (50MHz model) / 2ch = 16MS = 32ms of data, that means if my consecutive signals are longer than 32ms i will not be able to capture them? or how is this handled/what will happen if i try to capture more? does it just skip samples then? does it just stop recording? Does it just always limit to the 10MS sample rate once you ask more than the buffer size? What if i use the sdk, how does that handle memory or transfer speeds? If that's not possible, why are there 2xxx series scopes with 1GS speeds, you can never transfer those in time can you, it would take 1s to transfer 10ms of data?
The 3000 series has USB3.0 but according to the datasheet still only supports 17MHz transfer, still way too low. Does that mean i can't use either? or i need to grab a way bigger model with way more memory?
Im not shy of a bit of coding, im a programmer as well, so if the SDK can definitely do this then i can make it happen.
Thanks in advance
Re: Memory and USB speed confusion
The scope has 2 modes of operation
1. block mode.
All data is captured in the internal memory of the scope and transferred to the pc after the capture is finished. In this case you are always limited to the amount of sample memory in the scope divided by the number of channels. By doing it this way you can capture to as fast as the scope can.
2. streaming mode
All data is transferred to pc memory at a maximum rate depending USB2 or USB3 (10MS/sec or 17MS/sec) divided by the number of channels. In the SDK you can go faster in the scope specs is stated max streaming in picoscope and in the SDK.
You have no memory limits in this mode. I have done weeks of sampling at 10MS/sec looking for a specific problem and only save data to disk if the problem occurs (10 sec before and 10 sec after)
Benno
1. block mode.
All data is captured in the internal memory of the scope and transferred to the pc after the capture is finished. In this case you are always limited to the amount of sample memory in the scope divided by the number of channels. By doing it this way you can capture to as fast as the scope can.
2. streaming mode
All data is transferred to pc memory at a maximum rate depending USB2 or USB3 (10MS/sec or 17MS/sec) divided by the number of channels. In the SDK you can go faster in the scope specs is stated max streaming in picoscope and in the SDK.
You have no memory limits in this mode. I have done weeks of sampling at 10MS/sec looking for a specific problem and only save data to disk if the problem occurs (10 sec before and 10 sec after)
Benno
Re: Memory and USB speed confusion
So if i understand it correctly i can't use either, even in SDK since neither makes 200MS/s in streaming mode even in SDK mode?
So does picoscope have a scope with bigger memory that is under 1k? I can only seem to find really specific scopes with more memory? If i would go for 1s of data i would already need 200MS memory?
And this block mode, can i use this with the SDK? and how do i use it with the software, as i don't see a way to switch? (or i am blind
)
So does picoscope have a scope with bigger memory that is under 1k? I can only seem to find really specific scopes with more memory? If i would go for 1s of data i would already need 200MS memory?
And this block mode, can i use this with the SDK? and how do i use it with the software, as i don't see a way to switch? (or i am blind
)Re: Memory and USB speed confusion
on the GitHub examples are streaming and blok examples for a lot of different languages.
https://github.com/picotech
P.S. If I where you I would stay away from the scopes with less than 32MS sample memory, but of cource I can not see in your financial situation.
Benno
https://github.com/picotech
P.S. If I where you I would stay away from the scopes with less than 32MS sample memory, but of cource I can not see in your financial situation.
Benno
Re: Memory and USB speed confusion
Honestly i would love to see what scope you would recommend, €€€ under 1k is not a problem but the issue is more i don't really need the high bandwidth, just the faster transfer, and as far as i can find models in 2/3 series for normal usage i see none with high transfer or much memory, even 64MS from 3203D only captures 160ms of data. To get 1s i would need 400MS memory, which only the 1.4k 3206D has, that also has a way too high bandwidth i don't need, which most likely is most of the cost...
Re: Memory and USB speed confusion
What sample rate have you in mind for your needs ?
If I read your question right you need a MSO, I advice 4 analog channel + 16 digital channel.
3403D MSO USB3 (17 MS streaming PS) 125 MS SDK
Be aware that when using the SDK there is now way you can view the dat in picoscope software or use serial decoding (you need to write this yourselves)
But not knowing the data rate of your digital signal you need to analyze (usually you need about 5-10 samples per bit to analyze the problem)
So if you want to analyze a 400kbs i2c bus you need 4MS/sec sampling
If you analyze a 1Mbps CAN bus then you need 10MS/sec sampling.
I personally have the 3403D MSO and have never had a bus that I could not capture in streaming mode.
P.S. A other important item the 2000 and 3000 scopes are all 8 bit scopes (as most scopes from other vendors). That means if you are measuring 1V/div (-5V to +5V) the smallest signal you can see is about 40-50 mv and that is usually noise so practically you can only see differences of about 0.1V in the -5V to +5V range.
If you want higher resolution samples you need to look at the 4000 or 5000 series scopes.
Benno
If I read your question right you need a MSO, I advice 4 analog channel + 16 digital channel.
3403D MSO USB3 (17 MS streaming PS) 125 MS SDK
Be aware that when using the SDK there is now way you can view the dat in picoscope software or use serial decoding (you need to write this yourselves)
But not knowing the data rate of your digital signal you need to analyze (usually you need about 5-10 samples per bit to analyze the problem)
So if you want to analyze a 400kbs i2c bus you need 4MS/sec sampling
If you analyze a 1Mbps CAN bus then you need 10MS/sec sampling.
I personally have the 3403D MSO and have never had a bus that I could not capture in streaming mode.
P.S. A other important item the 2000 and 3000 scopes are all 8 bit scopes (as most scopes from other vendors). That means if you are measuring 1V/div (-5V to +5V) the smallest signal you can see is about 40-50 mv and that is usually noise so practically you can only see differences of about 0.1V in the -5V to +5V range.
If you want higher resolution samples you need to look at the 4000 or 5000 series scopes.
Benno
Re: Memory and USB speed confusion
Thanks bennog,
I need to analyse 20M bandwidth (SPI) so was aiming for 200MS (since i actually need proper analysis, otherwise it would defeat my purpose of getting the scope, and i could just get a logic analyser). But this data can't fit over the 125MS, thus that need for the 400M data. 0.1V seems actually a lot and might be not enough resolution as well...
But the 4/5 series seems quite specialized. Is there no 10-12bit 50MS scope with 400M memory?
Ill take a look around once more, but it seems thats apparently a lot to ask from picoscope without breaking the bank, while other brands seem to come closer, which is odd cause i always regarded picoscope as the golden standard. But i guess my workplace pays a lot more for their scopes...
I need to analyse 20M bandwidth (SPI) so was aiming for 200MS (since i actually need proper analysis, otherwise it would defeat my purpose of getting the scope, and i could just get a logic analyser). But this data can't fit over the 125MS, thus that need for the 400M data. 0.1V seems actually a lot and might be not enough resolution as well...
But the 4/5 series seems quite specialized. Is there no 10-12bit 50MS scope with 400M memory?
Ill take a look around once more, but it seems thats apparently a lot to ask from picoscope without breaking the bank, while other brands seem to come closer, which is odd cause i always regarded picoscope as the golden standard. But i guess my workplace pays a lot more for their scopes...
Re: Memory and USB speed confusion
Do you have examples of other brands.
I have not seen many others with that amount of sample memory.
Benno
I have not seen many others with that amount of sample memory.
Benno
Re: Memory and USB speed confusion
I might have not been clear, what i meant was if i don't use the buffer because it is too small then the other brands become similar if not cheaper in price for example:
Pico 3403D MSO in streaming (buffer too small for my application):
1165€
Analog: 100MS/4ch = 25MS = 2.5MHz 8-bit on 4 channels
Digital: 200MS on 16 channels unsure if streaming or buffer
1 output channel
Saleae Logic Pro 8 (doesn't use buffer mode either way):
999€
Analog: 50MS = 5MHz 12-bit on 8 channels
Digital: 500MS on 8 channels
8 output channels
So you can see, if the buffers aren't big enough, i don't see much difference anymore to similar priced devices. Sure the analog quality might be worse, and the software might be more tuned for digital. But at least i can analyze every sample that was captured until my pc ram ran out. I am in fact not sure what to get yet... I am willing to spend on a good device, but i don't want to have to buy two to then leave one untouched... I have taken a look at the 5x series only to see the devices with lots of buffer get very expensive and the usb still being the same slow usb3 sadly.
Pico 3403D MSO in streaming (buffer too small for my application):
1165€
Analog: 100MS/4ch = 25MS = 2.5MHz 8-bit on 4 channels
Digital: 200MS on 16 channels unsure if streaming or buffer
1 output channel
Saleae Logic Pro 8 (doesn't use buffer mode either way):
999€
Analog: 50MS = 5MHz 12-bit on 8 channels
Digital: 500MS on 8 channels
8 output channels
So you can see, if the buffers aren't big enough, i don't see much difference anymore to similar priced devices. Sure the analog quality might be worse, and the software might be more tuned for digital. But at least i can analyze every sample that was captured until my pc ram ran out. I am in fact not sure what to get yet... I am willing to spend on a good device, but i don't want to have to buy two to then leave one untouched... I have taken a look at the 5x series only to see the devices with lots of buffer get very expensive and the usb still being the same slow usb3 sadly.
Re: Memory and USB speed confusion
1 correction on the Saleae Logic Pro 8 here you need to devide the sample rate also by the number of channels in use.
So for single channel use PS can do streaming at 100 - 125 MS/sec.
The Saleae can do 50 MS/sec (if you are lucky)
The picoscope uses the internal memory (32 / 64 MS) as buffer memory if the USB or Windows has a hiccup and you don't lose data.
The Saleae has very little onboard memory for buffering. Could not find how much it has. So if USB or Windows has a hiccup of 100mS you will for sure mis a lot of samples here.
That is the difference between the 2.
Also the picoscope has probably better protection on the input, but they don't give specs on input protection (nor PS nor Saleae)
Benno
So for single channel use PS can do streaming at 100 - 125 MS/sec.
The Saleae can do 50 MS/sec (if you are lucky)
The picoscope uses the internal memory (32 / 64 MS) as buffer memory if the USB or Windows has a hiccup and you don't lose data.
The Saleae has very little onboard memory for buffering. Could not find how much it has. So if USB or Windows has a hiccup of 100mS you will for sure mis a lot of samples here.
That is the difference between the 2.
Also the picoscope has probably better protection on the input, but they don't give specs on input protection (nor PS nor Saleae)
Benno
Re: Memory and USB speed confusion
You are right, the Logic pro is also not the solution, but its cheaper as well, so what do you expect...
Don't get me wrong, if pico would release a model with a USB3 5gbps chip that could actually transfer 300MS 16-bit or 600MS 8-bit in streaming which should be possible with 5gbps, or even better a 10gbps version that could do 600MS (60MHz) 16-bit streaming and full 1GS in 8-bit and put it in a cheaper/lower Mhz model (which are 40$ including shipping btw so its not to save costs), they could replace the need for logic analyzers and scopes side to side for smaller Mhz signals and really take a big portion of the market rn. Could you imagine a tool that could not only automatically verify your signal integrity being within specification limit but also decode it in software while its at it? I would buy that in a heartbeat and replace all my other tools with it.
But right now it feels like the technology is really lagging behind. Laptops are coming with USB4 40gbps now. And real-time scopes that could really benefit from higher bandwidth is still struggling with the usb3 standard from 2008? Ill hold off from buying one for now, really sad that i would have to buy a 4.5k 6000 series to be able to capture a longer period of 50MHz data. I just cannot justify that budget.
Don't get me wrong, if pico would release a model with a USB3 5gbps chip that could actually transfer 300MS 16-bit or 600MS 8-bit in streaming which should be possible with 5gbps, or even better a 10gbps version that could do 600MS (60MHz) 16-bit streaming and full 1GS in 8-bit and put it in a cheaper/lower Mhz model (which are 40$ including shipping btw so its not to save costs), they could replace the need for logic analyzers and scopes side to side for smaller Mhz signals and really take a big portion of the market rn. Could you imagine a tool that could not only automatically verify your signal integrity being within specification limit but also decode it in software while its at it? I would buy that in a heartbeat and replace all my other tools with it.
But right now it feels like the technology is really lagging behind. Laptops are coming with USB4 40gbps now. And real-time scopes that could really benefit from higher bandwidth is still struggling with the usb3 standard from 2008? Ill hold off from buying one for now, really sad that i would have to buy a 4.5k 6000 series to be able to capture a longer period of 50MHz data. I just cannot justify that budget.
Re: Memory and USB speed confusion
Just to get some facts straight:bennog wrote: ↑Thu Feb 22, 2024 9:31 am1 correction on the Saleae Logic Pro 8 here you need to devide the sample rate also by the number of channels in use.
So for single channel use PS can do streaming at 100 - 125 MS/sec.
The Saleae can do 50 MS/sec (if you are lucky)
The picoscope uses the internal memory (32 / 64 MS) as buffer memory if the USB or Windows has a hiccup and you don't lose data.
The Saleae has very little onboard memory for buffering. Could not find how much it has. So if USB or Windows has a hiccup of 100mS you will for sure mis a lot of samples here.
That is the difference between the 2.
Also the picoscope has probably better protection on the input, but they don't give specs on input protection (nor PS nor Saleae)
Benno
I just read the saleae datasheet which states 50MS 12-bit on 3 channels, this is most likely due to the logic pro is capable of using 3gbps/5gbps available in usb3 while PS seems to only use 1gbps (125*8=1gbps). I also very much doubt usb will have so many hiccups, and that the logic pro won't have enough memory... Once your usb speed is high enough it just doesn't matter since they don't work with block modes, so its no wonder they don't list it.
Protection on the input Saleae lists as +-25V, as they are 10V inputs that seems reasonable. PS for 5000 series lists 50V for 20V inputs.
I mean i probs will buy the cheapest scope or maybe a physical one at this point and just a fast logic analyser, do the measurements at the same time with 2 devices, not as handy but i guess people have been doing that for a long time now...
Re: Memory and USB speed confusion
If that is really the case then it should be a better option for your problem.
I have seen some YouTube reviews and it looks like a nice device.
Benno
I have seen some YouTube reviews and it looks like a nice device.
Benno
Re: Memory and USB speed confusion
For future readers: i decided to buy the Picoscope 2207B and will be attempting to use the SDK for my purposes, if it doesn't work i won't have spent a huge amount and can still buy the logic analyser.
Re: Memory and USB speed confusion
You know the 2000 series is USB-2 scope and you wont get the USB-3 speeds.
Benno
Benno