PC Oscilloscope/LA anyone?

Hi Folks,
Apologies on a unrelated question.. I have been looking for a USB based
scope/LA which I could use on my ubuntu system - something less than
100$ yet should get around 20Mhz data if possible and upto 4 channels..
Ok, a bit of searching around, [1] is the best around I think, but a bit
out of my "project budget"...

Any recommendations folks?

Regards,
Nishanth Menon
Ref:
[1]http://www.bitscope.net/store/?p=list&i=cat+scope

You can make a decent LA for cheap using an FPGA.
I use an FPGA board from digilent.com. It dumps a output to the
serial port and I change it to a vcd for viewing in gtkwave, a
waveform viewer. Mine is single edge clocked but I remember an open
source project using a dual edge clock to get over 180 MHz.

Put simply, using the onchip ram, I've been able to get plenty of
samples for my needs, but theres always the onboard SRAM. Once the
trigger happens and the buffer is filled, I dump the data to the
serial port.

The newer FPGA boards even have USB ports and they're cheap too!

As for an O-scope, there are projects using fast 8-bit dacs on fpga4fun.com

g'day nishanth

i work with a beagleboard and gumstix verdex/overo COMs.
just purchased a BitScope BS325U - what an awesome piece of equipment
for the price.
it's well over the USD$100 limit you quoted, however, i just don't need
anything else for the basic electronics i am working with (100MHz
bandwidth limit).

regardz
guy

Hi Nishanth

Nishanth Menon wrote:

Hi Folks,
Apologies on a unrelated question.. I have been looking for a USB based
scope/LA which I could use on my ubuntu system - something less than
100$ yet should get around 20Mhz data if possible and upto 4 channels..
Ok, a bit of searching around, [1] is the best around I think, but a bit
out of my "project budget"...

Any recommendations folks?

Have a look to

It's USB based and ~150$. Community is waiting for Linux SW for it since several month, though. At the moment it's Windows only :frowning:

If you don't like this, maybe

http://elinux.org/Logic_Analyzers

can help you.

Best regards

Dirk

Take care, The Saleae device doesn't support 1.8V logic. According to their
homepage (http://www.saleae.com/logic/specs/) minimum input high is 2V and
isn't adjustable.

Anybody with experience in: http://www.pctestinstruments.com/? I have tested
their SW in off-line mode, at it seems great, but I would like to hear if
anyone have experience with the HW as well?

Best regards
  Søren

Take care, The Saleae device doesn't support 1.8V logic. According to their
homepage (http://www.saleae.com/logic/specs/) minimum input high is 2V and
isn't adjustable.

Anybody with experience in: http://www.pctestinstruments.com/? I have tested
their SW in off-line mode, at it seems great, but I would like to hear if
anyone have experience with the HW as well?

Sadly only windows.

But I picked up one anyway to debug my FPGA work and it is working
well for me so far.

Philip

Folks,
thanks for your feedback, I am just putting together a summary below of
the options:
a) Brian recommended doing one with an fpga from digilent.com and also
recommended looking at 8-bit dacs on fpga4fun.com
b) Guy has had good experience with Bitscope BS325U -
http://www.bitscope.net/
    + Has already s/w to work on linux
    + 100Mhz
    - cheapest one looks to be around 475 USD

c) Soren suggests http://www.pctestinstruments.com/ and Philip is
already using it and is happy with it..
    - windows only s/w
    + upto 500Mhz?
    - 389 USD

Thinking a little more, a whackier idea popped up: tlc5540[1] throws out
8 bit parallel data in sync with a clock.. if i could serialize this
data, I should be able to interface it over McBSP Dr -> the fastest data
transfer I can think of - esp if I hook up DMA to SDRAM.. i could (in
theory again) create a small interface board which could plug into
beagleboard!!! Beagle could interface over DVI or over USB to a PC for
showing the data -> in theory run the same s/w[e.g. 2] from PC or on
beagle.. hmm.. mebbe another 20$ more add-on to beagle and I should be
able to get 100Mhz scope rt?

Ref:
1: http://focus.ti.com/docs/prod/folders/print/tlc5540.html i saw
somewhere this was a 5USD chip.
2: http://xoscope.sourceforge.net/

Hi Nishanth,

Nishanth Menon wrote:

Folks,
thanks for your feedback, I am just putting together a summary below of
the options:
a) Brian recommended doing one with an fpga from digilent.com and also
recommended looking at 8-bit dacs on fpga4fun.com
b) Guy has had good experience with Bitscope BS325U -
http://www.bitscope.net/
    + Has already s/w to work on linux
    + 100Mhz
    - cheapest one looks to be around 475 USD

c) Soren suggests http://www.pctestinstruments.com/ and Philip is
already using it and is happy with it..
    - windows only s/w
    + upto 500Mhz?
    - 389 USD

Thinking a little more, a whackier idea popped up: tlc5540[1] throws out
8 bit parallel data in sync with a clock.. if i could serialize this
data, I should be able to interface it over McBSP Dr -> the fastest data
transfer I can think of - esp if I hook up DMA to SDRAM.. i could (in
theory again) create a small interface board which could plug into
beagleboard!!! Beagle could interface over DVI or over USB to a PC for
showing the data -> in theory run the same s/w[e.g. 2] from PC or on
beagle.. hmm.. mebbe another 20$ more add-on to beagle and I should be
able to get 100Mhz scope rt?

Ref:
1: TLC5540 data sheet, product information and support | TI.com i saw
somewhere this was a 5USD chip.
2: http://xoscope.sourceforge.net/

Sorry, yes I know, always this "Dirk and his wikis" :wink: , but do you like to add something to

http://elinux.org/Logic_Analyzers

?

Best regards

Dirk

Happier?

Folks,
thanks for your feedback, I am just putting together a summary below of
the options:
a) Brian recommended doing one with an fpga from digilent.com and also
recommended looking at 8-bit dacs on fpga4fun.com
b) Guy has had good experience with Bitscope BS325U -
http://www.bitscope.net/
    + Has already s/w to work on linux
    + 100Mhz
    - cheapest one looks to be around 475 USD

c) Soren suggests http://www.pctestinstruments.com/ and Philip is
already using it and is happy with it..
    - windows only s/w
    + upto 500Mhz?
    - 389 USD

Another one for the mix is the Zeroplus unit. This is one of the few met most
of my selection criteria:

- Can handle 1.8V logic
- Can do decodes of the data
- Around 100USD.
- Linux software

The Lap-C model comes
(Logic Analyzers-Zeroplus) closests with the
biggest failure being a lack of Linux software. $120 seems to be the street
price and was discounted down to $109 a few weeks ago.

The 1.8V and around 100USD criteria eliminated a lot of the things out there
as the lower priced ones tend to have a fixed logic threshold. I wanted some
kind of decodes as it gets old fast trying to reassemble data in your head
while looking things up in a chart. This unit claims it can directly decode a
USB 1.1 bus or even do a direct connection to a RS-232 port.

No relations to the unit; this is just what I found during a recent hunt for a
1.8V friendly test tool after deciding a voltage shifter + Parallel port
leads to a whole other mess and I still have to manually decode things.

Søren Steen Christensen wrote:

Take care, The Saleae device doesn't support 1.8V logic. According to their
homepage (http://www.saleae.com/logic/specs/) minimum input high is 2V and
isn't adjustable.

I talked with Kees about this. He used Saleae for OMAP3 JTAG
debugging, and it seems that 1.8V does work. Most probably above specs
are not correct.

Additionally, Kees (thanks!) asked Saleae again regarding Linux
support. Will let you know about this.

Best regards

Dirk

We had a positive experience with Saleae at 1.8V on the Touch Book and
the Beagleboard. I don't want to take responsibility :wink: but at least,
it was working fine for us to debug SPI,

Grégoire

Hi Grégoire and Dirk,

Thanks for the info - I don't have any experience with the product, I just
read the homepage and found it incompatible with the 1.8V logic on OMAP3 :slight_smile:
- But if works it works - Thanks for the update :slight_smile:

BTW, Dirk: Kees who? - Jongenburger?

Best regards and thanks
  Søren

Søren Steen Christensen wrote:

Hi Grégoire and Dirk,

Thanks for the info - I don't have any experience with the product, I just
read the homepage and found it incompatible with the 1.8V logic on OMAP3 :slight_smile:
- But if works it works - Thanks for the update :slight_smile:

BTW, Dirk: Kees who? - Jongenburger?

Yes. 'keesj' at IRC

Dirk

Hi Nishanth,

Thinking a little more, a whackier idea popped up: tlc5540[1] throws
out 8 bit parallel data in sync with a clock.. if i could serialize this
data, I should be able to interface it over McBSP Dr -> the fastest
data transfer I can think of - esp if I hook up DMA to SDRAM..

Right, except the McBSP is limited to a 48MHz serial clock => A 8-bit 6MHz
scope which isn't that useful after all. One other possibility might be to
connect it to the expansion connector MMC interface using a FPGA as
protocol-translator and interface logic. This would give you a 48MHz 8-bit
interface (minus MMC protocol overhead). But in theory and with a little
hacking circumventing all kind of MMC stack and using DMA, I guess you could
achieve 40MB/s continuous transfer => Nearly a factor 7 more than the McBSP
approach. Of cause on the cost of a higher price and complexity due to the
FPGA...

i could (in theory again) create a small interface board which could plug

into

beagleboard!!! Beagle could interface over DVI or over USB to a PC for
showing the data -> in theory run the same s/w[e.g. 2] from PC or on
beagle.. hmm.. mebbe another 20$ more add-on to beagle and I should be
able to get 100Mhz scope rt?

Agree - Great project - Making the OMAP accept a 100MB/s data stream might
however cause some kind of a problem not having access to the GPMC- or
SDRC-bus. One way would be to do the processing and trigging in a FPGA with
some special attached memory and then afterwards transfer the data to the
OMAP - Only utilizing the OMAP for data presentation.

Although a better "idea/hack" would be using both OMAP MMC interfaces in
parallel to achieve 80MB/s continuous data transfer. If the ARM/DSP-core(s)
can keep up processing (searching for a trigger event) this data in real
time I don't know, but in case of the right optimizations (code running from
cache and DMAs transferring data from MMC interfaces to internal memory)
until a trigger is found I would assume so :slight_smile:

Best regards - God luck with the project - Please let me know in case you
take up the challenge :slight_smile:
  Søren

Hi Grégoire and Dirk,

Thanks for the info - I don't have any experience with the product, I just
read the homepage and found it incompatible with the 1.8V logic on OMAP3 :slight_smile:
- But if works it works - Thanks for the update :slight_smile:

BTW, Dirk: Kees who? - Jongenburger?

Yes, me