Red Pitaya for a beginner

[newman89]

Has anyone gotten their hands on a Red Pitaya and put it through its paces?

I have looked around and can't find much in terms of reviews and what I can find is very basic.

Would it be worth getting one or would it only really be useful for simple projects.

For anyone that doesn't know what I'm talking about: http://redpitaya.com/

Cheers

[QuantumPhys]

Hi,

  I do not actually have the board so do not take my word for it.

  I have been following the project since its kickstarter, and I feel that even though the software seems nice, and the hardware is of good quality, the thing is still quite under-speced , making it "just another USB oscilloscope". It may be interesting if it is cheaper, but at $470, you can get a HP logic analyzer plus a Tek 2465B plus a function generator!

  Again, someone who has this board would give much better advice.

[robimarko]

They are bragging that it is open source but there is absolutely no schematics or part lists avaiable

[Dolganoff]

I've asked the same question once on their G+ page, nobody responded...

As I understand (maybe I am wrong), the Open Source part doesn't refer to hardware at all, but to the applications they would like us to write for their platform...

[jeremy]

I've got one. The HDL code is out there, and it seems to be quite well made physically. Heatsink on the Zynq is decent too.

I'm really interested in using it for its fast adc connected to an FPGA (for DSP), as well as the ability to do some linuxy stuff at the same time. I'd like to make very specific instruments, I don't really care about it being an ethernet oscilloscope.

I think you can write code for the Zynq-7010 using the free version of xilinx vivado, so that is probably what they are expecting you to do.

I guess if you use it with its stock configuration as a 50MHz scope, it probably is not a wise investment at that price. I bought it so that I don't have to solder BGAs on my own pcb Definitely not a good idea as a beginners device.

Happy to try and answer any specific questions you have.

[charlespax]

They are bragging that it is open source but there is absolutely no schematics or part lists avaiable

If development isn't done in the open, I really don't consider it open source. Even the Beagleboard folks just have releases on Github.

[Howardlong]

I noticed the Red Pitaya has been offered at a discounted price by the Elektor store. US$266  (select currency top right cnr)
http://www.elektor.com/red-pitaya

Blimey. half price, not sure what to make of that.

[jeremy]

Well, after playing with one of these for a good part of a year, I'd say that you shouldn't get one of these unless you have a decent vivado license. Also, the schematic and board layout are not released, and the system code is very unmaintained (I had to compile my own fsbl and uboot so that it would work with a reasonably recent kernel, so I ended up throwing away their code). I tried to mainline my uboot port, but I think the board was a little too proprietary to have it accepted.

However, as a zynq development kit for $266, I'm considering buying one for me (the one I have is for work). It's the closest thing to an open source oscilloscope that I've been able to find so far, and if you want to do some fast dsp then it is also one of the cheapest. It's definitely cheaper than putting an fpga and add/dac combo on the board yourself (those three chips alone add up to about $150).

Perhaps they are half price because there is a new one coming?

[janoc]

Perhaps they are half price because there is a new one coming?

Or the existing one isn't selling for that price ... It is seriously niche device.

[tggzzz]

Has anyone gotten their hands on a Red Pitaya and put it through its paces?

I have looked around and can't find much in terms of reviews and what I can find is very basic.

Would it be worth getting one or would it only really be useful for simple projects.

For anyone that doesn't know what I'm talking about: http://redpitaya.com/

Cheers

My feelings, as someone that has looked at it but doesn't own one, is that:

• it is an excellent way of getting hold of an FPGA+ADC+processor
• if you want to understand/modify the internals of an oscilloscope, it is unparalled
• if you want to investigate tight low-latency feedback from an ADC to DACs etc, then it is unparalleled
• FPGA development is possible and this provides a good starting point
• FPGA development is not for the faint-hearted; it requires time and determination
• the ground connections for the digital signals on the extension connector are poor, unnecessarily poor
• the hardware is closed-source to an unnecessary degree
• the originators are accessible through their forums
• as a general purpose scope, there are better alternatives - no surprises there!

[tggzzz]

I noticed the Red Pitaya has been offered at a discounted price by the Elektor store. US$266  (select currency top right cnr)
http://www.elektor.com/red-pitaya

RS has similar offers.

[jeremy]

I noticed the Red Pitaya has been offered at a discounted price by the Elektor store. US$266  (select currency top right cnr)
http://www.elektor.com/red-pitaya

RS has similar offers.

Not similar for us Australians  Down from $505+tax to $454+tax.

[charlespax]

Are there any opensource hardware alternatives to the Red Pitaya?

[jeremy]

Not that I know of. I'd definitely pay a bit for a board with a similar fast adc, fast dac and a zynq 7020, as long as it was properly open sourced.

I don't really know why they didn't open source it; if they were worried about profit, it's not like I'm going to get an 8 layer board full of BGAs fabbed myself anyway... It would cost a truckload in low volumes

[tggzzz]

I don't really know why they didn't open source it; if they were worried about profit, it's not like I'm going to get an 8 layer board full of BGAs fabbed myself anyway... It would cost a truckload in low volumes

I don't think the designers are worried about people like us!

[Xykon]

I backed the SmartScope on Kickstarter which was running at about the same time as the Red Pitaya. RP was too expensive for my taste so I went with the SmartScope which was promised to be completely open source. Well it turned out LabNation went back on their word... they're not going to publish details of the analog frontend on their scope and also the GUI won't be fully open source... even though it is heavily based on open source components. The GUI right now is absolute garbage so the thing is mostly useless. I think they could have gained a lot by publishing the sources... once again a good idea was crippled by stupidity and greed.

I finally decided to order a Red Pitaya from Elektor... I ordered it with the case though which delayed my order... they are now sending me the Red Pitaya first and will ship the case for free once it becomes available. Should arrive in the next few days and I'll be looking forward to checking it out.

I do have an old bench scope and spectrum analyser but they take a lot of space plus generate a lot of noise so I'd love to replace them with a small portable instrument. Of course I don't expect the Red Pitaya to be as powerful as these professional instruments but if it can help me with some basic tasks that would already be a big win. Can always setup the big instruments when I really need their superior power.

[miguelvp]

Hmm, seems they are releasing the information, but might take them a while:

http://wiki.lab-nation.com/images/a/ab/SmartScope_-_Schematics_-_v1.0.pdf

http://wiki.lab-nation.com/index.php/Sources

Seems the forums have more information as well.

[1design]

As I mentioned in another post Red Pitaya was not primarily designed as an instrument as such, more of a dev. platform, like the Arduino, so people can do their own things with it, not just stick to the few "downloadable" applications. Think about what you can learn by using it as a tool, modifying the FPGA and SW to suit your need for signal acquisition, processing and generation. If you add a shield with a few RF components you a get an SDR with 60MHz real time BW for example etc.

Even if you don't know how to do these things this is a good tool to learn this, with this knowledge you get additional value to your professional skillset so more jobs are now available as this is an advanced set of skills that not many have.

Best Regards,
Miha

[piranha32]

Has anyone gotten their hands on a Red Pitaya and put it through its paces?
I have looked around and can't find much in terms of reviews and what I can find is very basic.
Would it be worth getting one or would it only really be useful for simple projects.

I've got RP on Kickstarter, and I'm disappointed. The idea is very nice, but implementation - not so much. I'd have a lot to say about very crude official software, not improved much since the first release making the tool more of an educational toy than a real measurement instrument, short buffer, etc. However, the most important reason for me to leave it on a shelf is noise. The image below shows how the spectrum of the signal from ADCs looks like with no signal connected to the inputs (inputs terminated with 50ohm SMA terminators). 25dB spike at 2.3MHz, a bit lower one at second harmonic, plus quite a bit of noise at low frequencies. Of course it is well visible as noise on scope view, making RP useless for measuring weak signals.

[daveatol]

However, the most important reason for me to leave it on a shelf is noise. The image below shows how the spectrum of the signal from ADCs looks like with no signal connected to the inputs (inputs terminated with 50ohm SMA terminators). 25dB spike at 2.3MHz, a bit lower one at second harmonic, plus quite a bit of noise at low frequencies. Of course it is well visible as noise on scope view, making RP useless for measuring weak signals.

What's the range setting of the channel in your spectrum plot? If 0 dBm is the reference, the -84 dB noise floor fits in perfectly with the quantisation noise of 14-bit sampling. (Though the 2.3MHz peak shouldn't be evident, obviously)

[piranha32]

However, the most important reason for me to leave it on a shelf is noise. The image below shows how the spectrum of the signal from ADCs looks like with no signal connected to the inputs (inputs terminated with 50ohm SMA terminators). 25dB spike at 2.3MHz, a bit lower one at second harmonic, plus quite a bit of noise at low frequencies. Of course it is well visible as noise on scope view, making RP useless for measuring weak signals.

What's the range setting of the channel in your spectrum plot? If 0 dBm is the reference, the -84 dB noise floor fits in perfectly with the quantisation noise of 14-bit sampling. (Though the 2.3MHz peak shouldn't be evident, obviously)

Low voltage (+/- 1V).
 

[piranha32]

However, the most important reason for me to leave it on a shelf is noise. The image below shows how the spectrum of the signal from ADCs looks like with no signal connected to the inputs (inputs terminated with 50ohm SMA terminators). 25dB spike at 2.3MHz, a bit lower one at second harmonic, plus quite a bit of noise at low frequencies. Of course it is well visible as noise on scope view, making RP useless for measuring weak signals.

What's the range setting of the channel in your spectrum plot? If 0 dBm is the reference, the -84 dB noise floor fits in perfectly with the quantisation noise of 14-bit sampling. (Though the 2.3MHz peak shouldn't be evident, obviously)

Low voltage (+/- 1V).

I just checked with "High Voltage" setting. The peaks remain the same, what indicates that the noise is not picked up by the input circuit, but appears somewhere closer to the ADC.

[daveatol]

Low voltage (+/- 1V).

Well, ±1.095V is 0 dBm.

I just checked with "High Voltage" setting. The peaks remain the same, what indicates that the noise is not picked up by the input circuit, but appears somewhere closer to the ADC.

The ADC sample rate is 125 MHz, right? If the peak is at 2.5 MHz, then it's possible that the ADC is clocked by 2.5 MHz (which leaks into the ADC inputs) and uses a x50 PLL to get the sample rate. How did you measure the frequency of the peak?

[piranha32]

Low voltage (+/- 1V).

Well, ±1.095V is 0 dBm.

I just checked with "High Voltage" setting. The peaks remain the same, what indicates that the noise is not picked up by the input circuit, but appears somewhere closer to the ADC.

The ADC sample rate is 125 MHz, right? If the peak is at 2.5 MHz, then it's possible that the ADC is clocked by 2.5 MHz (which leaks into the ADC inputs) and uses a x50 PLL to get the sample rate. How did you measure the frequency of the peak?

The frequency is more of a "ballpark" figure. At this frequency the spectrum analyzer app detected the peak, and the exact frequency detected is 2.35MHz. According to the schematic from http://wiki.redpitaya.com/index.php?title=External_ADC_clock the system is clocked from 125MHz XO, so it's rather unlikely that the ADCs are clocked using lower frequency.

[tggzzz]

The idea is very nice, but implementation - not so much.

There's another area where the implementation is unnecessarily poor: the signals and grounds on the external connector
https://www.eevblog.com/forum/microcontrollers/red-pitaya-hd-images-for-getting-the-schematic/msg541053/#msg541053

All the "NC" pins were grounds, and they were were distributed amongst the signals - not good for signal integrity. And it is compounded by not being able to use LVDS on those pins, due to the choice of Zynq IO bank voltage