Logic analyzers

[Phil Hobbs]

We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
knockoff of a Saleae Logic 8. It's been pretty handy for characterizing
the ADC in a bathymetric lidar chip we've been helping design.

It'll stream data forever at 16 MS/s. However, at higher sample rates
it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)

The USB form factor is nice, because it's the sort of instrument you
want to pull out and attach down in the guts of some gizmo, and with USB
you can just use a longer cable, which is convenient.

Sooo, I'm considering buying something a bit better than our $8
Chinesium knockoff. Suggestions?

Thanks

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

[Ricky]

On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:
> We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
> knockoff of a Saleae Logic 8. It's been pretty handy for characterizing
> the ADC in a bathymetric lidar chip we've been helping design.
>
> It'll stream data forever at 16 MS/s. However, at higher sample rates
> it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)
>
> The USB form factor is nice, because it's the sort of instrument you
> want to pull out and attach down in the guts of some gizmo, and with USB
> you can just use a longer cable, which is convenient.
>
> Sooo, I'm considering buying something a bit better than our $8
> Chinesium knockoff. Suggestions?

A customer has this unit.

https://www.pctestinstruments.com/

He is very happy with it. Had it for a decade or so.

--

Rick C.

- Get 1,000 miles of free Supercharging
- Tesla referral code - https://ts.la/richard11209

[Three Jeeps]

On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:

Up front note: I am a believer in high end logic analyzers after I was part of a team to evaluate (at HP's request) their two products (1602A, 1610 a )logic analyzers back in the day. Fast fwd to 2023....
Without knowing some of your requirements (speed, channels, etc.) hard to say. Still, I'd point you to Keysight 16861A - a 34 channel LA or the 16862a (currently have one in my lab). May not be in the budget you have. I'd consider a older used unit.

If you are drawn to PC-USB configurations, I've used the Saleae Pro 16 and thought it was pretty good - a bit higher in price than some of the chinesium products. Fairly good sampling rates: 500 MS/s and 100 MHz for digital, and analog of 50 MS/s and 5 MHz.

I have seen but not used a Digilent product, which intrigues me:
Digilent Digital (or Analog) discovery.
https://digilent.com/shop/digital-discovery-portable-usb-logic-analyzer-and-digital-pattern-generator/
https://digilent.com/search.php?search_query_adv=%22discovery%22

A colleague of mine recently got this for his lab, which he likes (after doing a fair bit of searching). It may be more functionality than you need but the DA +LA functions are intriguing...
https://www.dreamsourcelab.com
In general, I am not a big fan of having test gear tied to a PC ....You are tied into continuous vendor support to keep up with OS updates and changes, unless one freezes their environment.

Good luck

[Lasse Langwadt Christensen]

onsdag den 11. januar 2023 kl. 03.02.11 UTC+1 skrev jjhu...@gmail.com:
> On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:
> > We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
> > knockoff of a Saleae Logic 8. It's been pretty handy for characterizing
> > the ADC in a bathymetric lidar chip we've been helping design.
> >
> > It'll stream data forever at 16 MS/s. However, at higher sample rates
> > it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)
> >
> > The USB form factor is nice, because it's the sort of instrument you
> > want to pull out and attach down in the guts of some gizmo, and with USB
> > you can just use a longer cable, which is convenient.
> >
> > Sooo, I'm considering buying something a bit better than our $8
> > Chinesium knockoff. Suggestions?
> >
> > Thanks
> >
> > Phil Hobbs
> >
> > --
> > Dr Philip C D Hobbs
> > Principal Consultant
> > ElectroOptical Innovations LLC / Hobbs ElectroOptics
> > Optics, Electro-optics, Photonics, Analog Electronics
> > Briarcliff Manor NY 10510
> >
> > http://electrooptical.net
> > http://hobbs-eo.com
>

> If you are drawn to PC-USB configurations, I've used the Saleae Pro 16 and thought it was pretty good - a bit higher in price than some of the chinesium products.

"a bit", about 100x ...

[Dan Purgert]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

On 2023-01-10, Phil Hobbs wrote:
> [...]

> Sooo, I'm considering buying something a bit better than our $8
> Chinesium knockoff. Suggestions?

I'm quite pleased with my Saleae Logic. Definitely a pricey tool, but
I'm very glad I spent the money on it...


-----BEGIN PGP SIGNATURE-----
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=I7N7
-----END PGP SIGNATURE-----

--
|_|O|_|
|_|_|O| Github: https://github.com/dpurgert
|O|O|O| PGP: DDAB 23FB 19FA 7D85 1CC1 E067 6D65 70E5 4CE7 2860

[Fred Bloggs]

On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:

> We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
> knockoff of a Saleae Logic 8. It's been pretty handy for characterizing
> the ADC in a bathymetric lidar chip we've been helping design.
>
> It'll stream data forever at 16 MS/s. However, at higher sample rates
> it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)

Looks like that thing has a mind of its own with lots of interplay between available internal resources and your acquisition settings. I don't like the sound of that loop buffer for the sample data where it sounds like it just writes over old data without even telling you. Then the spec says it has a "10+ billion" ( somethings they don't say) sample memory depth, which leads you to believe they can buffer a significant amount of sample time, but then it goes on to say your computer is the buffer memory.

All logic analyzer designs are illogical.

Maybe look for a fast USB disc you can plug directly into the Saleae to take in the stream, and read that out later.

[Lasse Langwadt Christensen]

onsdag den 11. januar 2023 kl. 15.01.44 UTC+1 skrev Fred Bloggs:
> On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:
> > We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
> > knockoff of a Saleae Logic 8. It's been pretty handy for characterizing
> > the ADC in a bathymetric lidar chip we've been helping design.
> >
> > It'll stream data forever at 16 MS/s. However, at higher sample rates
> > it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)
> Looks like that thing has a mind of its own with lots of interplay between available internal resources and your acquisition settings. I don't like the sound of that loop buffer for the sample data where it sounds like it just writes over old data without even telling you. Then the spec says it has a "10+ billion" ( somethings they don't say) sample memory depth, which leads you to believe they can buffer a significant amount of sample time, but then it goes on to say your computer is the buffer memory.

it streams input data directly to the PC, there is no buffer

>
> All logic analyzer designs are illogical.
>
> Maybe look for a fast USB disc you can plug directly into the Saleae to take in the stream, and read that out later.

that's not how USB works

[Fred Bloggs]

On Wednesday, January 11, 2023 at 9:20:35 AM UTC-5, lang...@fonz.dk wrote:
> onsdag den 11. januar 2023 kl. 15.01.44 UTC+1 skrev Fred Bloggs:
> > On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:
> > > We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
> > > knockoff of a Saleae Logic 8. It's been pretty handy for characterizing
> > > the ADC in a bathymetric lidar chip we've been helping design.
> > >
> > > It'll stream data forever at 16 MS/s. However, at higher sample rates
> > > it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)
> > Looks like that thing has a mind of its own with lots of interplay between available internal resources and your acquisition settings. I don't like the sound of that loop buffer for the sample data where it sounds like it just writes over old data without even telling you. Then the spec says it has a "10+ billion" ( somethings they don't say) sample memory depth, which leads you to believe they can buffer a significant amount of sample time, but then it goes on to say your computer is the buffer memory.
> it streams input data directly to the PC, there is no buffer

Then what is that sample memory depth about in the spec sheet?

> >
> > All logic analyzer designs are illogical.
> >
> > Maybe look for a fast USB disc you can plug directly into the Saleae to take in the stream, and read that out later.
> that's not how USB works

That's how a USB stick works, and this analyzer says it has a USB output. Sounds like a match to me.

[Lasse Langwadt Christensen]

onsdag den 11. januar 2023 kl. 15.25.55 UTC+1 skrev Fred Bloggs:
> On Wednesday, January 11, 2023 at 9:20:35 AM UTC-5, lang...@fonz.dk wrote:
> > onsdag den 11. januar 2023 kl. 15.01.44 UTC+1 skrev Fred Bloggs:
> > > On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:
> > > > We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
> > > > knockoff of a Saleae Logic 8. It's been pretty handy for characterizing
> > > > the ADC in a bathymetric lidar chip we've been helping design.
> > > >
> > > > It'll stream data forever at 16 MS/s. However, at higher sample rates
> > > > it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)
> > > Looks like that thing has a mind of its own with lots of interplay between available internal resources and your acquisition settings. I don't like the sound of that loop buffer for the sample data where it sounds like it just writes over old data without even telling you. Then the spec says it has a "10+ billion" ( somethings they don't say) sample memory depth, which leads you to believe they can buffer a significant amount of sample time, but then it goes on to say your computer is the buffer memory.
> > it streams input data directly to the PC, there is no buffer
> Then what is that sample memory depth about in the spec sheet?

they have to write something, they could have written unlimited except by resources on PC

> > >
> > > All logic analyzer designs are illogical.
> > >
> > > Maybe look for a fast USB disc you can plug directly into the Saleae to take in the stream, and read that out later.
> > that's not how USB works
> That's how a USB stick works, and this analyzer says it has a USB output. Sounds like a match to me.

no, that is not how USB works

[Phil Hobbs]

Fred Bloggs wrote:
> On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:
>> We've been using a little 8-bit, 48-MHz USB logic analyzer--a
>> Chinese knockoff of a Saleae Logic 8. It's been pretty handy for
>> characterizing the ADC in a bathymetric lidar chip we've been
>> helping design.
>>
>> It'll stream data forever at 16 MS/s. However, at higher sample
>> rates it quits after ~100k samples. Sheesh, and we spent eight
>> bucks on it. ;)
>
> Looks like that thing has a mind of its own with lots of interplay
> between available internal resources and your acquisition settings. I
> don't like the sound of that loop buffer for the sample data where it
> sounds like it just writes over old data without even telling you.
> Then the spec says it has a "10+ billion" ( somethings they don't
> say) sample memory depth, which leads you to believe they can buffer
> a significant amount of sample time, but then it goes on to say your
> computer is the buffer memory.
>
> All logic analyzer designs are illogical.
>
> Maybe look for a fast USB disc you can plug directly into the Saleae
> to take in the stream, and read that out later.

This one is just a dongle with a single chip in it, a copy of Saleae's
original Logic (originally $149), which they no longer sell. So it
relies pretty heavily on the USB bandwidth and latency of the computer.

Simon is using it with Sigrok, I think. Certainly we're not using the
Saleae software.

[Fred Bloggs]

I don't give a d___ how USB works. I do know there's an infinitude of versatility with those thumb drives. Find one that works.

[whit3rd]

It's not that simple; write speed to flash is limited, and there's little RAM to buffer.
A spinning-rust hard drive, with a few dozen MB of RAM cache, outpaces thumb drives
for speed, and a short RAID stack can go faster yet. Your logic analyzer must deliver
specified write speeds that match a logic device, cannot generally tolerate the slow approximations of
a USB data stream.

[Lasse Langwadt Christensen]

[Phil Hobbs]

Lasse Langwadt Christensen wrote:
> onsdag den 11. januar 2023 kl. 03.02.11 UTC+1 skrev jjhu...@gmail.com:
>> On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:
>>> We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
>>> knockoff of a Saleae Logic 8. It's been pretty handy for characterizing
>>> the ADC in a bathymetric lidar chip we've been helping design.
>>>
>>> It'll stream data forever at 16 MS/s. However, at higher sample rates
>>> it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)
>>>
>>> The USB form factor is nice, because it's the sort of instrument you
>>> want to pull out and attach down in the guts of some gizmo, and with USB
>>> you can just use a longer cable, which is convenient.
>>>
>>> Sooo, I'm considering buying something a bit better than our $8
>>> Chinesium knockoff. Suggestions?

>>

>> If you are drawn to PC-USB configurations, I've used the Saleae Pro 16 and thought it was pretty good - a bit higher in price than some of the chinesium products.
>
> "a bit", about 100x ...
>

The Chinesium ones use a Nordic 8-bit-parallel-to-USB3 chip, and the
rest is up to Sigrok.

Turns out that you can get 16 bits by using two of them at once. You
have to stitch the decoded data files together yourself, but that's not
too hard. (Simon's a Python whiz.)

Our setup is:

Dongle 0 ($5, AliExpress) : bit 0 = clock, bits 1-7 = ADC bits 0-6

Dongle 1 ($8, Amazon) : bit 0 = clock, bits 0-8 = ADC bits 6-13

They're both looking at the clock and at bit 6, so it's easy to align
the time stamps.

The ADC clock is coming from a Highland P400 DDG, so we just wired it
up, hit the green button to start the clock, and it just works. (BTW
the Ali one has the label on straight, and the Amazon one doesn't.)

Fun.

Cheers

[Lasse Langwadt Christensen]

I have a few, some are just bare boards (probably intended as USB dev board and I think it can do 16bits )

all based on a Cypress FX2 chip which basically a 8051 with high speed USB

[Phil Hobbs]

Well, we're going to be in the market for a real 16-bit one, it looks
like. Turns out that this 12-bit ADC has a _gigantic_ glitch at the bit
10 carry--see

<https://electrooptical.net/www/sed/Glitch.png> and
<https://electrooptical.net/www/sed/GlitchDetail.png>.

These are single frames from a cell phone video summary, so they aren't
perfect, but they get the point across.

The red curve is the ADC samples, plotted against the right hand scale,
and the blue curve is the input ramp signal, plotted against the left
hand scale in volts. The horizontal scale is seconds. The two curves
were roughly aligned by eye, which accounts for the minor time scale
error. The ADC clock was 1 MHz, and the logic analyzer's was 16 MHz
(give or take).

As you can see, it's not just a normal sort of high-order carry glitch,
or even really just a DNL issue at all, because it takes many samples
to recover.

The sawtooth is pretty blameless. It's based on a bootstrap ramp
generator. (I got the idea from JL long ago.)

<https://electrooptical.net/www/sed/RampGeneratorRedrawn1.6.pdf>

Differentiating the ramp with a 20 nF / 1k ohm RC, it's linear to within
10 ppm or thereabouts. There's a small soakage tail lasting about 2 ms
after the reset, because I actually used a mylar cap rather than
polyprop, but after that the differentiated output is basically just the
noise of the TCA0372 (22 nV in 1 Hz). And of course the TCA0372 output
looks like a car battery (1.4 A).

We'll try it again with another chip, because it's possible that we blew
this one up in some artistic manner.

[Phil Hobbs]

Just noticed that the current source for the bootstrap was drawn wrong.
Fix underway.

PH

[three_jeeps]

snip

I was chatting with one of my colleagues in our lab and asked him for input on lower cost USB based LAs. He pointed to this one which he said he purchased for his personal use (The $150USD version).
https://www.dreamsourcelab.com/shop-category/logic-analyzer/
He thought the software was well done and was an updated design that has a good amount of memory.
In looking at the offerings, the U3 Pro looks like a worthwhile feature upgrade, e.g. 1GHz sampling rate/8 channels, 500MHz/16 ch, Stream mode withUSB3 @ $300 USD

Good luck
J

[Michael Schwingen]

On 2023-01-10, Phil Hobbs <pcdhSpamM...@electrooptical.net> wrote:
> We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
> knockoff of a Saleae Logic 8. It's been pretty handy for characterizing
> the ADC in a bathymetric lidar chip we've been helping design.
>
> It'll stream data forever at 16 MS/s. However, at higher sample rates
> it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)

USB may be quite fragile when running continuous at high data rates - I have
that, too, both with original Saleae and with clones. Good quality cables,
clamp-on ferrites and a different USB host may help. A separate USB BUS with
no other devices on it is also a good idea.

> Sooo, I'm considering buying something a bit better than our $8
> Chinesium knockoff. Suggestions?

We have the original Salea at work (the USB3 Logic Pro 16). Not cheap, but
works quite nice.

At home, I have a chinese Logic16 clone - that is based on the USB2 FX2 chip
with a FPGA in front. Works fine with Sigrok (which the newer Logic16 Pro
did not when I bought it).

cu
Michael
--
Some people have no respect of age unless it is bottled.

[DemonicTubes]

On Tuesday, January 10, 2023 at 3:12:50 PM UTC-7, Phil Hobbs wrote:
Suggestions?

I've been using a Saleae Logic Pro 16 for the last few years (using their software).

No troubles as long as I have it hooked to a modern laptop.

I also sometimes use the logic analyzer on my Rigol DS1104Z+
...but regret it (it is nowhere near as convenient to setup and use.)

[Phil Hobbs]

Three Jeeps wrote:
> On Tuesday, January 10, 2023 at 5:12:50 PM UTC-5, Phil Hobbs wrote:
>> We've been using a little 8-bit, 48-MHz USB logic analyzer--a
>> Chinese knockoff of a Saleae Logic 8. It's been pretty handy for
>> characterizing the ADC in a bathymetric lidar chip we've been
>> helping design.
>>
>> It'll stream data forever at 16 MS/s. However, at higher sample
>> rates it quits after ~100k samples. Sheesh, and we spent eight
>> bucks on it. ;)
>>
>> The USB form factor is nice, because it's the sort of instrument
>> you want to pull out and attach down in the guts of some gizmo, and
>> with USB you can just use a longer cable, which is convenient.
>>
>> Sooo, I'm considering buying something a bit better than our $8
>> Chinesium knockoff. Suggestions?

> Up front note: I am a believer in high end logic analyzers after I
> was part of a team to evaluate (at HP's request) their two products
> (1602A, 1610 a )logic analyzers back in the day. Fast fwd to
> 2023.... Without knowing some of your requirements (speed, channels,
> etc.) hard to say. Still, I'd point you to Keysight 16861A - a 34
> channel LA or the 16862a (currently have one in my lab). May not be
> in the budget you have. I'd consider a older used unit.

I'm a huge fan of boat anchors--I have a whole lab full of
top-of-the-line HP and Tek gear from the '90s and early '00s. I bought
much of it when I first went out on my own, in 2009. That was a really
wonderful time to be buying equipment--on average I paid about 4 cents
on the dollar.

The PC-USB thing isn't something I'd consider for expensive gear,
because as you say it's brittle and not very future-proof. (Not as bad
as oscilloscopes that run Windows, but don't let me get started on that
old rabbit trail.)

> If you are drawn to PC-USB configurations, I've used the Saleae Pro
> 16 and thought it was pretty good - a bit higher in price than some
> of the chinesium products. Fairly good sampling rates: 500 MS/s and
> 100 MHz for digital, and analog of 50 MS/s and 5 MHz.
>
> I have seen but not used a Digilent product, which intrigues me:
> Digilent Digital (or Analog) discovery.
> https://digilent.com/shop/digital-discovery-portable-usb-logic-analyzer-and-digital-pattern-generator/
>
>
>
>
https://digilent.com/search.php?search_query_adv=%22discovery%22

I've seen the Analog Discovery 2, because my younger daughter had one
for her EE class at the community college. (She now does PCB layout for
Cirrus Logic in Austin.) They're pretty cool gizmos, especially for the
dough, but to my eye the signal integrity wasn't thought out as well as
the main board. Plastic BNC connectors, yecch. I'll check out the
Digital Discovery.

> A colleague of mine recently got this for his lab, which he likes
> (after doing a fair bit of searching). It may be more functionality
> than you need but the DA +LA functions are intriguing...
> https://www.dreamsourcelab.com In general, I am not a big fan of
> having test gear tied to a PC ....You are tied into continuous vendor
> support to keep up with OS updates and changes, unless one freezes
> their environment.
>
> Good luck

That's not that hard to do with Sigrok, I don't think--we run stuff like
that in Docker containers sometimes. Dunno if that would trash the USB
speed.

[LM]

16 channels USB
https://www.dreamsourcelab.com/product/dslogic-series/

[Lasse Langwadt Christensen]

boards like this will do 16bit with sigrok/pulseview
https://www.amazon.com/Cypress-CY7C68013A-EZ-USB-USB2-0-Development/dp/B06XHP9XKF

but of course with the same USB bandwidth limitations

>
> <https://electrooptical.net/www/sed/Glitch.png> and
> <https://electrooptical.net/www/sed/GlitchDetail.png>.
>
> These are single frames from a cell phone video summary, so they aren't
> perfect, but they get the point across.
>
> The red curve is the ADC samples, plotted against the right hand scale,
> and the blue curve is the input ramp signal, plotted against the left
> hand scale in volts. The horizontal scale is seconds. The two curves
> were roughly aligned by eye, which accounts for the minor time scale
> error. The ADC clock was 1 MHz, and the logic analyzer's was 16 MHz
> (give or take).

that looks weird, that "exponential decay" is actual samples and not some artifact of the plotting?

[Phil Hobbs]

Interesting, thanks.

>
>>
>> <https://electrooptical.net/www/sed/Glitch.png> and
>> <https://electrooptical.net/www/sed/GlitchDetail.png>.
>>
>> These are single frames from a cell phone video summary, so they
>> aren't perfect, but they get the point across.
>>
>> The red curve is the ADC samples, plotted against the right hand
>> scale, and the blue curve is the input ramp signal, plotted against
>> the left hand scale in volts. The horizontal scale is seconds. The
>> two curves were roughly aligned by eye, which accounts for the
>> minor time scale error. The ADC clock was 1 MHz, and the logic
>> analyzer's was 16 MHz (give or take).
>
> that looks weird, that "exponential decay" is actual samples and not
> some artifact of the plotting?

Sure looks real to us. It's a pipelined capacitive ADC, so internal
settling stuff can come out in subsequent samples.

I don't know enough about the design to comment much on its guts.

[Michael Schwingen]

On 2023-01-12, Phil Hobbs <pcdhSpamM...@electrooptical.net> wrote:
>
> The PC-USB thing isn't something I'd consider for expensive gear,
> because as you say it's brittle and not very future-proof. (Not as bad
> as oscilloscopes that run Windows, but don't let me get started on that
> old rabbit trail.)

If you can use it with sigrok, it should be quite future-proof. USB works
fine if you don't push it to the maximum possible throughput - a USB3
connection has advantages.

What I like about the USB logic analyzers is that you can record huge
amounts of data (limited by PC memory, and the software does on-the-fly
compression) and later zoom in on the details - with a classic analyzer with
integrated memory, you need to setup triggers beforehand - buf after
triggering, those can be faster.

What is better depends on the problem you are analyzing.

cu
Michael

[Phil Hobbs]

With our stuff, there are usually three or four boat anchors in use
already, so having the logic analyzer be unobtrusive is a win.

Cheers

Phil "Still trying to figure out that weird ADC thing" Hobbs

[Lasse Langwadt Christensen]

fredag den 13. januar 2023 kl. 16.42.51 UTC+1 skrev Phil Hobbs:
> Michael Schwingen wrote:
> > On 2023-01-12, Phil Hobbs <pcdhSpamM...@electrooptical.net> wrote:
> >>
> >> The PC-USB thing isn't something I'd consider for expensive gear,
> >> because as you say it's brittle and not very future-proof. (Not as bad
> >> as oscilloscopes that run Windows, but don't let me get started on that
> >> old rabbit trail.)
> >
> > If you can use it with sigrok, it should be quite future-proof. USB works
> > fine if you don't push it to the maximum possible throughput - a USB3
> > connection has advantages.
> >
> > What I like about the USB logic analyzers is that you can record huge
> > amounts of data (limited by PC memory, and the software does on-the-fly
> > compression) and later zoom in on the details - with a classic analyzer with
> > integrated memory, you need to setup triggers beforehand - buf after
> > triggering, those can be faster.
> >
> > What is better depends on the problem you are analyzing.
> With our stuff, there are usually three or four boat anchors in use
> already, so having the logic analyzer be unobtrusive is a win.
>
> Cheers
>
> Phil "Still trying to figure out that weird ADC thing" Hobbs

there must be bits miswired, missing, or shorted

it never recovers, after the glitch the slope/gain/offset looks different

https://electrooptical.net/static/oldsite/www/sed/Glitch.png

[Phil Hobbs]

Maybe. It's a vendor IP block that they changed around for us, to make
the floor plan long and skinny. All of the bits seem to be changing
independently, though. I'll check for supply current glitches if the
problem is reproducible on another chip.

Cheers

Phil Hobbs

[Klaus Vestergaard Kragelund]

On 10-01-2023 23:12, Phil Hobbs wrote:
> We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
> knockoff of a Saleae Logic 8.  It's been pretty handy for characterizing
> the ADC in a bathymetric lidar chip we've been helping design.
>
> It'll stream data forever at 16 MS/s.  However, at higher sample rates
> it quits after ~100k samples.  Sheesh, and we spent eight bucks on it. ;)
>

I cannot get enough instruments. Can you share a link to the one you have?

[Phil Hobbs]

It's actually 24 MHz, I misspoke. If you search Amazon for logic
analyzers and sort by price, there are about 20 very similar to these
ones. AFAICT they all have that same Nordic chip in them.

[Three Jeeps]

agreed, but a name would save a lot of inference....

[Flyguy]

You can sort out the riff-raff by searching for "64 channel". This eliminates most, if not all, of the low end logic analyzers. For instance, I found:
https://files.thelabeshop.com/2981/brochure.pdf
https://www.acute.com.tw/logic-analyzer-en/product/logic-analyzer/la3000

[Lasse Langwadt Christensen]

something like this: https://let-elektronik.dk/shop/2220-logic-analyzers/18627-usb-logic-analyzer---24mhz8-channel/

[Phil Hobbs]

Three Jeeps wrote:
> On Friday, January 13, 2023 at 6:19:46 PM UTC-5, Phil Hobbs wrote:
>> Klaus Vestergaard Kragelund wrote:
>>> On 10-01-2023 23:12, Phil Hobbs wrote:
>>>> We've been using a little 8-bit, 48-MHz USB logic analyzer--a Chinese
>>>> knockoff of a Saleae Logic 8. It's been pretty handy for
>>>> characterizing the ADC in a bathymetric lidar chip we've been helping
>>>> design.
>>>>
>>>> It'll stream data forever at 16 MS/s. However, at higher sample rates
>>>> it quits after ~100k samples. Sheesh, and we spent eight bucks on it. ;)
>>>>
>>> I cannot get enough instruments. Can you share a link to the one you have?
>> It's actually 24 MHz, I misspoke. If you search Amazon for logic
>> analyzers and sort by price, there are about 20 very similar to these
>> ones. AFAICT they all have that same Nordic chip in them.
>> Cheers

>

> agreed, but a name would save a lot of inference....
>

Dunno exactly--I didn't buy them.

[Klaus Vestergaard Kragelund]

Nice, thanks. And even a Danish site, only 100km from my location :-)

[Arie de Muijnck]

Or like this, exactly what Phil describes:
https://www.google.com/search?q=Usb+Logic+Analyzer+24Mhz+8+channel+aliexpress
https://nl.aliexpress.com/item/1005002991193528.html

They actually run a bit faster than 16 Msample/s, I get about 24 Msample/s.
They go for less than EUR 10, including shipping. I have two, just in case I blow up one.

Arie