Comparison of the three Pyboard D models?
Comparison of the three Pyboard D models?
I see on the store pages basic descriptions of the boards, and by flipping back and forth between the tabs, I can see the differences "jump out" as visually as text differences, but I think there differences should be summarized in side-by-side table form somewhere. Just my two cents. Aside from that recommendation, can I get some clarity on the second model marked SF3W. I believe its singular difference from the first model, the SF2W, is a feature briefly indicated as "built-in high-speed USB PHY (USB HS peripheral)". What does that distinction imply precisely? Presumably all three models sync with a desktop via USB just like the old v1.1 board. Does the second board transmit and receive data faster over USB? Is it just "faster" in an arbitrary numerical sense, or is it actually a different USB protocol (is the SF2W running USB 2 and the SF3W running USB 3, just as an example?). As to the third board, the SF6W, it is described as having twice the RAM and four times the flash of the other two models, plus double-precision floating point. Great, but it doesn't actually list the USB PHY feature of the SF3W. Does that imply that the third board actually has the weaker USB performance of the first board? Again, this really isn't very clear I'm afraid. I'm sorry to come across troublesome, pestering about details I feel I should be able to gather on my own, but I'm just confused right now.
Thank you.
Thank you.
Re: Comparison of the three Pyboard D models?
The main (and only?) difference is the main IC. All the details you can possibly imagine (with a nice comparision table) can be found here: https://www.st.com/b/en/microcontroller ... tId=SS1858
And the top model (SF-6) hasn‘t got the hardware USB. I think the hardware USB will come in handy if you would like to do "real" USB Stuff, like beeing a usb host or acting like a real (not only serial) USB device like a mouse or usb stick.
Of course, part of that should be possible without the usb phy, because the ols pyboard was also capable of acting as a usb mouse, so I‘m not absolutely sure.
But the STM datasheet surely will have more infos about that.
And you are absolutely right, a nice comparrision and more Infos on the boards are needed, and I‘m sure they will follow...
And the top model (SF-6) hasn‘t got the hardware USB. I think the hardware USB will come in handy if you would like to do "real" USB Stuff, like beeing a usb host or acting like a real (not only serial) USB device like a mouse or usb stick.
Of course, part of that should be possible without the usb phy, because the ols pyboard was also capable of acting as a usb mouse, so I‘m not absolutely sure.
But the STM datasheet surely will have more infos about that.
And you are absolutely right, a nice comparrision and more Infos on the boards are needed, and I‘m sure they will follow...
Re: Comparison of the three Pyboard D models?
The 723 and 767 both have USB HS. The 722 doesnt.
The max observable transfer speed over usb fs is around 6-8Mbit/sec.
With USB HS you could peak at 400Mbit/sec but I‘m not sure how fast the pyboard can deliver data with that pace - but it will likely be much faster than FS. On the REPL you probably wont notice though so I guess it makes sense if you upload/download data a lot.
As for modelling USB devices all three have similar capabilities and are not limited to usb vcp. any usb device can be created in c code but only a subset (hid , msc and vcp profiles) are supported in uPy.
As for the difference to the big 767 board: it has all the advanced peripherals, notably the ethernet support. besides what was already mentioned (ram/flash/cache size, double prec etc).
Internal flash might be also be faster than those code parts which are mapped via qspi to memory space.
The max observable transfer speed over usb fs is around 6-8Mbit/sec.
With USB HS you could peak at 400Mbit/sec but I‘m not sure how fast the pyboard can deliver data with that pace - but it will likely be much faster than FS. On the REPL you probably wont notice though so I guess it makes sense if you upload/download data a lot.
As for modelling USB devices all three have similar capabilities and are not limited to usb vcp. any usb device can be created in c code but only a subset (hid , msc and vcp profiles) are supported in uPy.
As for the difference to the big 767 board: it has all the advanced peripherals, notably the ethernet support. besides what was already mentioned (ram/flash/cache size, double prec etc).
Internal flash might be also be faster than those code parts which are mapped via qspi to memory space.
Re: Comparison of the three Pyboard D models?
Only PYBD-SF3-xxxx has a high speed USB interface with integrated high speed PHY. Most if not every STM32F7xx MCU has a high speed USB interface but only F723/F733 have an integrated high speed PHY.
So if you really need high speed (480 MBaud) USB, PYBD-SF3 is the only choice.
So if you really need high speed (480 MBaud) USB, PYBD-SF3 is the only choice.
Re: Comparison of the three Pyboard D models?
@chuckbook you are right, sorry about the noise
Re: Comparison of the three Pyboard D models?
Okay, that's interesting (that the top-level D board actually lacks a feature of the second-level board, so that there is a "decision" to make on such matters, as opposed to each next level simply subsuming the lower levels).
Thanks.
Thanks.
Re: Comparison of the three Pyboard D models?
Between the release of the PyBoard D and the OpenMV H7, it's looking up to be a pretty remarkable spring for MicroPython projects. I'll have to decide what I want to do.
- rcolistete
- Posts: 352
- Joined: Thu Dec 31, 2015 3:12 pm
- Location: Brazil
- Contact:
Re: Comparison of the three Pyboard D models?
Anybody has measured the maximum USB transfer speed of Pyboard D SF3W ?
Does it achieve near 400-480 Mbps ?
Does it achieve near 400-480 Mbps ?
My "MicroPython Samples". My "MicroPython Firmwares" with many options (double precision, ulab, etc).
Re: Comparison of the three Pyboard D models?
Here too: what i missed on https://store.micropython.org/ is a compare between the Pyboard D-series...
The only differences are RAM/ROM size and USB HS peripheral, isn't it?
I also find the various product labels a little confusing.
In the end it results in this:
PYBD-SF2-W4F2 == SF2W model == SoC STM32F722IEK
- 512KiB ROM and 256KiB RAM
- no USB HS peripheral
- 512KiB ROM and 256KiB RAM
- built-in high-speed USB PHY (USB HS peripheral)
- 2048KiB ROM and 512KiB RAM
- no USB HS peripheral
I think it's also confusing that https://pybd.io/ is not part of the officially documentation, because there is also a section of the old pyboard: http://docs.micropython.org/en/latest/p ... ckref.html
If this is extra so wanted, it should at least be better linked together. But I don't think this is a good solution either.
Re: Comparison of the three Pyboard D models?
I'll confirm this for sure next week, but off the top of my head here are some additional points relevant to your summary:
They all have USB HS _peripherals_ but only the 723 (SF3) has the on-board HS PHY. (I realise this distinction is irrelevant for most people though as the pybd does not include an HS PHY on the other models).
The 767 (i.e. SF6) also adds double-precision floating point. And an ethernet peripheral.
Additionally, the RAM/ROM for the MCU itself, all pybd models also includes 2MiB of extra qspiflash ROM.
Agreed on the pybd.io / docs.micropython.org confusion. I think it comes down to the tiles and adaptors -- they don't really fit on docs.micropython.org, but perhaps some of the pybd-only stuff should move. And the feature matrix is a good idea.
They all have USB HS _peripherals_ but only the 723 (SF3) has the on-board HS PHY. (I realise this distinction is irrelevant for most people though as the pybd does not include an HS PHY on the other models).
The 767 (i.e. SF6) also adds double-precision floating point. And an ethernet peripheral.
Additionally, the RAM/ROM for the MCU itself, all pybd models also includes 2MiB of extra qspiflash ROM.
Agreed on the pybd.io / docs.micropython.org confusion. I think it comes down to the tiles and adaptors -- they don't really fit on docs.micropython.org, but perhaps some of the pybd-only stuff should move. And the feature matrix is a good idea.