Are low/lower cost USB Oscilloscope's any good? [Design]

Prev: 74C925 alternative?
Next: 12 and 16-bit oscilloscopes

From: PeteS on 13 Dec 2006 15:43

Joerg wrote:
> Hal Murray wrote:
>
>>> I've been impressed with these guys: http://www.cleverscope.com/ --
>>> they seem to have a solid understanding of what terms like "noise"
>>> and "jitter" mean, unlike many of the cheap scopes out there.
>>
>>
>> A friend has one. He's very happy with it.
>>
>> One of the advantages of a PC based scope is that it doesn't take
>> up much room on your desktop.
>>
>
> But why are they so slow? At least the ones I have seen were. We
> designed 400MSPS converter boards even back in the late 80's and early
> 90's. It ain't rocket science.
>

The company I work for doesn't have much of a budget for test equipment,
but we have a budget for boards (yeah I know) so I was considering
making a pretty fast scope / analyser using a quad set of 500MS/s A-Ds
and a couple of fast FPGAs and some memory.

It won't be perfect, but it might do the job and save 'that budget'.

Cheers

PeteS

From: Joerg on 13 Dec 2006 15:56

PeteS wrote:

> Joerg wrote:
>
>> Hal Murray wrote:
>>
>>>> I've been impressed with these guys: http://www.cleverscope.com/ --
>>>> they seem to have a solid understanding of what terms like "noise"
>>>> and "jitter" mean, unlike many of the cheap scopes out there.
>>>
>>>
>>>
>>> A friend has one. He's very happy with it.
>>>
>>> One of the advantages of a PC based scope is that it doesn't take
>>> up much room on your desktop.
>>>
>>
>> But why are they so slow? At least the ones I have seen were. We
>> designed 400MSPS converter boards even back in the late 80's and early
>> 90's. It ain't rocket science.
>>
>
> The company I work for doesn't have much of a budget for test equipment,
> but we have a budget for boards (yeah I know) so I was considering
> making a pretty fast scope / analyser using a quad set of 500MS/s A-Ds
> and a couple of fast FPGAs and some memory.
>
> It won't be perfect, but it might do the job and save 'that budget'.
>

That's a strange way of looking at the financial bottomline. Look at the
(realistic) number of paid hours you'll be working on that one-off
project and then compare that to a nice used Tek scope. Now copy your
boss on that.

I have opted for the used Tek scope for my lab :-)

--
Regards, Joerg

http://www.analogconsultants.com

From: PeteS on 13 Dec 2006 17:32

From: a7yvm109gf5d1 on 13 Dec 2006 17:22

Joerg wrote:

> That's a strange way of looking at the financial bottomline. Look at the
> (realistic) number of paid hours you'll be working on that one-off
> project and then compare that to a nice used Tek scope. Now copy your
> boss on that.

Seriously. If it were that easy to make a scope, everyone would be
doing it...

From: Charlie Edmondson on 13 Dec 2006 17:25

Joerg wrote:

> PeteS wrote:
>
>> Joerg wrote:
>>
>>> Hal Murray wrote:
>>>
>>>>> I've been impressed with these guys: http://www.cleverscope.com/ --
>>>>> they seem to have a solid understanding of what terms like "noise"
>>>>> and "jitter" mean, unlike many of the cheap scopes out there.
>>>>
>>>>
>>>>
>>>>
>>>> A friend has one. He's very happy with it.
>>>>
>>>> One of the advantages of a PC based scope is that it doesn't take
>>>> up much room on your desktop.
>>>>
>>>
>>> But why are they so slow? At least the ones I have seen were. We
>>> designed 400MSPS converter boards even back in the late 80's and
>>> early 90's. It ain't rocket science.
>>>
>>
>> The company I work for doesn't have much of a budget for test
>> equipment, but we have a budget for boards (yeah I know) so I was
>> considering making a pretty fast scope / analyser using a quad set of
>> 500MS/s A-Ds and a couple of fast FPGAs and some memory.
>>
>> It won't be perfect, but it might do the job and save 'that budget'.
>>
>
> That's a strange way of looking at the financial bottomline. Look at the
> (realistic) number of paid hours you'll be working on that one-off
> project and then compare that to a nice used Tek scope. Now copy your
> boss on that.
>
> I have opted for the used Tek scope for my lab :-)
>
Back when I work at UC Santa Barbara, we had the same problem.
Equipment budgets needed to be approved, and were severely limited, so
often it was decided to build things rather than buy them. Why buy a
$5000 control system when you can get the parts for $500 and just have a
guy put it together and then go install it?

The success story was the video distribution system for the campus.
Basically, it was a 16x256 channel switcher, with balance baseband
(6MHz) video on twinax, stereo sound and a two way intercom on each
channel. They built it in the 60s using RTL, and they were finally
going to start to replace it in 1993. The problem? They couldn't find
the RTL chips anymore for repairs, and some new buildings going in were
going to cut some of those custom multi-core twinax cables and they
couldn't replace or re-run them. The replacement would be on fiber.

The horror story I was directly involved with. There was this one
building on campus that was basically four large lecture halls arranged
in a square, operated by a facility manager that had very specific goals
for any controls. He specified what the controllers would look like: It
had to be BIG (at least 3" x 9") and had to be HARD WIRED (so the
professors couldn't walk off with them!) The lighting control had to be
done with a pot, so you could pre-set your levels. The buttons needed
to be lighted and at least 1" square...

Then, my boss got involved. He was an EE, and had been in charge of the
A/V maintanance dept before being promoted. He thought that touch tone
signalling would be just the thing, and directed one of the techs to
build the electronics. He directed another tech to run the cables for
the control. He then gave the cable tech a student helper (me!) to help
run the cables and install all the electical boxes.

Note, of course, that none of these specifications were ever written
down. The cable guy dedided that a five conductor cable would be fine.
Power, TT control signal, ground, lighting control power, and the
lighting level signal. He then got the cable and ran it all over the
building. Approximate cost, counting his and my time - $2000.

Then, the guy designing the electronics went on a bender, and never
built it, or even designed it! The cabling guy 'found other employment'
and the project needed to be passed on. So, they gave it to the new
student A/V design engineer - ME!

First realization - five conductors weren't enough. The wires weren't
even twisted! The cabling guy was an electrician, and had decided that
audio didn't really need all that fancy twisting and stuff. So,
everything was going to share that one ground wire...

I built the prototype, using parts lying around the place, and got it to
work. Only took me about 3 months, so only cost the department about
another $1000. Successfully demonstrated the prototype, installed it in
one of the lecture halls, and was given the go ahead to go into
production for the rest of the building. I designed the PC boards, but
when the EE department decided it would take 6 weeks to fab them, I was
put on to another project, and one of the techs was given the actual
installation duties.

Of course, I made two big mistakes on the boards, so the tech had to
'improvise' modifications. He built up the boards, tested them in the
lab, and started installing them. He put in the power supplies, mounted
and hooked up the boards, and powered them up. They worked for about
2-6 hours, then died giving up magic smoke!

After a week of trouble shooting ( the project was once again 'my baby')
I found the problem. I had originally used an off board +5 supply for
the electronics, but on redesign, had moved the 7805 onto the individual
boards. I had then specified 24 VDC supplies for the main power. When
installing the power supplies, the tech had found a whole bunch of
really nice unregulated 24 VDC supplies in a cabinet that he then
installed for the main power. However, since we were using only a tiny
fraction of the power needed, these supplies were a little overkill.
They were also 28 VDC supplies (for some cameras or something...) and
when unloaded, put out about 40 VDC. This overvoltage was cooking the
7805's

Finally, after about 3 years, the project was abandoned, and a real,
commercial system was installed. I figure the department spent between
$10K to $15K on trying to get that system to work. The commercial
system cost $4K.

Yes, I spec'd and installed the commercial system! I was the 'official'
campus A/V engineer by that time!

Charlie

First | Prev | Next | Last
Pages: 1 2 3 4 5 6 7
Prev: 74C925 alternative?
Next: 12 and 16-bit oscilloscopes