I spent a little more time with the Futek A/D. I also took its cover =
off to look inside.
First my observations on performance --
I think my first impressions of extremely high noise are due to a bug =
in their software. When I look at only one channel at a time I get =
results much closer to what I would expect from a chip of this type =
i.e.. about 50 counts p-p noise with the input shorted (100 ohms )
When both inputs are active the result is over 2000 counts noise but =
the display does not look right. The distribution is not at all uniform =
over all possible values. In fact it doesn't even make sense. I think =
the hardware is probably good to ~18 bits p-p or 22 bits RMS which is =
what the chip spec says it should be .=20
The included software --as it is --is not good. But some of the files =
could be useful for development of custom software.
Inside the box --
The chip is a TI MSC1200Y2. The chip includes 4 channel 24 bit A/D and =
a 8051 microprocessor with extensive IO. The USB interface is the usual =
FTDI serial to USB device. All unused pins on the MSC1200 are readily =
accessible and there is a header inside which is probably for loading =
the program.
If one has the interest and skills the board could be an excellent =
stating point for a nice system since there is quite a bit of processing =
power inside and access to all of the IO of the processor. The only =
difference between the 4 channel device and the two channel is only two =
channels of the input components are populated and the firmware only =
calls two channels . The board itself and all active parts appear to be =
the same.
Conclusion ---=20
A good starting point but the product ,as is, has some problems. The =
hardware itself has a lot of flexibility and a convenient package for =
low bucks but it would be a real project to make something of it. On the =
other hand it might be worth some effort.
I intend to make a few more tests -- if anything comes from them I will =
report on it . If you have a suggestion for a test or would like to try =
it yourself just let me know.
Dave Nelson=20
Rolling Hills Estates, California=20
From: ian=20
Sent: Wednesday, January 05, 2011 11:28 AM
To: psnlist@.................
Subject: Re: Fw: USB Data Acquisition
if you want a usb daq, then have a look at =
http://sine.ni.com/nips/cds/view/p/lang/en/nid/201987 I haven't used =
usb versions but I know their products are very good. I use a pci =
version for my system. Some software comes with the device but I don't =
know if it would do all that is needed for a station. =20
You also get them on ebay.
cheers
Ian
On 04/01/2011 23:53, Dave Nelson wrote:=20
=20
I saw this A/D on Ebay and decided to try it . I bought the two =
channel version for $99. =20
I tested it using the included software with the inputs terminated in =
100 ohms. The result was EXTREMELY POOR !!!!=20
The noise level is ~2000 counts peak to peak apparently independent =
of the sample rate. This results in an effective resolution of closer =
to 12 bits not 24 on a p-p basis .=20
This is not too surprising based on the specs for the low cost =
delta-sigma A/D chips.=20
The software makes and displays a record from the start time to the =
end time.=20
The time resolution of the real time display gets worse as the record =
length gets longer. It is not a scope type display with a variable time =
interval.
So far I have come to the conclusion this thing is useless with its =
included software. There are .dll files for windows if one could write =
new or adapt existing software for display/analysis but that may not =
improve the noise.
The apparent independence of the noise as the sample rate is varied =
leads me to the initial conclusion that the A/D is running at a fixed =
high sample rate and the lower effective sample rates are derived by =
pure decimation.=20
I need to take a closer look at it but so far I am totally unimpressed =
and don't recommend this device for any useful purpose.
24 bit A/D's in general do not perform as well as one would expect or =
hope . Read the effective resolution specs very carefully and note that =
they usually refer to RMS noise whereas the noise we want to use for =
seismic records is most likely peak to peak. The difference is 3 to 4 =
bits.=20
We are not in the audio business where RMS noise may be more =
appropriate. Frequently the effective resolution is not much better than =
16 bits at 200 sps.
Larry clearly understand the importance of the noise and goes to great =
lengths to reduce it in his A/D boards. Sometimes the very expensive =
24 bit A/Ds are not nearly as good as the manufacturers lead one to =
believe. =20
Save your money on this one.
Dave Nelson=20
Rolling Hills Estates , California =20
Do any of you have any experience with these ACD devices. They appear =
to compete with DataQ products.
=20
http://www.futekinst.com/futek/dataacquisitionsystem.asp
=20
I spent a little more time with the =
Futek=20
A/D. I also took its cover off to look inside.
First my observations on performance=20
--
I think my first impressions of =
extremely high=20
noise are due to a bug in their software. When I look at only one =
channel=20
at a time I get results much closer to what I would expect from a =
chip of=20
this type i.e.. about 50 counts p-p noise with the input shorted (100 =
ohms=20
)
When both inputs are active the result =
is over 2000=20
counts noise but the display does not look right. The =
distribution is=20
not at all uniform over all possible values. In fact it doesn't even =
make sense.=20
I think the hardware is probably good to ~18 bits p-p or 22 bits RMS =
which is=20
what the chip spec says it should be .
The included software --as it is --is =
not good. But=20
some of the files could be useful for development of custom=20
software.
Inside the box --
The chip is a TI MSC1200Y2. The =
chip includes=20
4 channel 24 bit A/D and a 8051 =
microprocessor=20
with extensive IO. The USB interface is the usual FTDI serial to USB=20
device. All unused pins on the MSC1200 are readily accessible and =
there is=20
a header inside which is probably for loading the program.
If one has the interest and skills the =
board could=20
be an excellent stating point for a nice system since there is quite a =
bit of=20
processing power inside and access to all of the IO of the processor. =
The only=20
difference between the 4 channel device and the two channel is only two =
channels=20
of the input components are populated and the firmware only calls two =
channels .=20
The board itself and all active parts appear to be the=20
same.
Conclusion ---
A good starting point but the product =
,as is, has=20
some problems. The hardware itself has a lot of flexibility and a =
convenient=20
package for low bucks but it would be a real project to make =
something of=20
it. On the other hand it might be worth some effort.
I intend to make a few more tests -- if =
anything=20
comes from them I will report on it . If you have a suggestion for a =
test or=20
would like to try it yourself just let me know.
Dave Nelson
Rolling Hills Estates, California =
Sent: Wednesday, January 05, 2011 11:28 AM
Subject: Re: Fw: USB Data Acquisition
if you want a usb daq, then have a look at http://sin=
e.ni.com/nips/cds/view/p/lang/en/nid/201987 =20
I haven't used usb versions but I know their products are very =
good. I use=20
a pci version for my system. Some software comes with the device =
but I=20
don't know if it would do all that is needed for a station. =
You=20
also get them on ebay.
cheers
Ian
On 04/01/2011 =
23:53, Dave=20
Nelson wrote:=20
I saw this A/D on Ebay and decided to =
try it . I=20
bought the two channel version for $99.
I tested it using the included =
software with the=20
inputs terminated in 100 ohms. The result was EXTREMELY POOR=20
!!!!
The noise level is ~2000 counts =
peak to=20
peak apparently independent of the sample rate. This results in an =
effective=20
resolution of closer to 12 bits not 24 on a p-p basis . =
This is not too surprising based on =
the specs for=20
the low cost delta-sigma A/D chips.
The software makes and displays =
a record=20
from the start time to the end time.
The time resolution of the real time =
display gets=20
worse as the record length gets longer. It is not a scope type display =
with a=20
variable time interval.
So far I have come to the =
conclusion this=20
thing is useless with its included software. There are .dll =
files=20
for windows if one could write new or adapt existing =
software for=20
display/analysis but that may not improve the =
noise.
The apparent independence of the =
noise as the=20
sample rate is varied leads me to the initial conclusion that the A/D =
is=20
running at a fixed high sample rate and the lower effective sample =
rates are=20
derived by pure decimation. =
I need to take a closer look at it =
but so far I=20
am totally unimpressed and don't recommend this device for any useful=20
purpose.
24 bit A/D's in general do not =
perform as well as=20
one would expect or hope . Read the effective resolution specs very =
carefully=20
and note that they usually refer to RMS noise whereas the=20
noise we want to use for seismic records is most likely peak to=20
peak. The difference is 3 to 4 bits.
We are not in the audio =
business where RMS=20
noise may be more appropriate. Frequently the effective =
resolution=20
is not much better than 16 bits at 200 sps.
Larry clearly understand the =
importance of the=20
noise and goes to great lengths to reduce it in his A/D =
boards. =20
Sometimes the very expensive 24 bit A/Ds are not nearly as good =
as the=20
manufacturers lead one to believe.
Save your money on =
this=20
one.
Dave Nelson
Rolling Hills Estates , California=20