Hi,

I've just had a refresher look at the software.  The A/D board has it's 
own crystal controlled timing to trigger the sampling and to control the 
time between channel samples. There is also a 4000 sample FIFO buffer to 
store the readings, so the timing of the samples is absolutely rock 
solid and unaffected by activities on the pc.

Where I deserve a wrist smack is the timestamps.  These are applied as 
each sample group (the 3 samples for each of the 2 instruments) are read 
from the FIFO buffer and therefore are affected by activities on the 
pc.  I'll add this to my todo list to use the A/D's internal timing for 
timestamps.

For communication it's just standard 100 mbit tcp/ip.  Totally 
transparent to the user and given that the data is all timestamped (sort 
of!) not something to worry about.

I'll need to measure the noise to answer the noise question.  

Looking back I can see that there are a few traces (not many) which 
flipped between channel gains +/-0.1V and +/-1.0V.  So the feature is 
giving me more gain to look into the weaker signals without being 
clobbered with saturation on the stronger traces - both the high and low 
pass filters apply a 20db gain to compensate for the attenuation of the 
filters.

I asked about your 12 Km gap as I was wondering if there was a way to 
plug it with equipment running at broadband rates.  I'm in one of the 
BT(our telco)  "black spots", condemned to never have broadband and I 
eventually had to put together our own community wireless broadband 
(http://www.gmccbroadband.org/ ).  This partly involves filling "gaps" 
across the countryside.  Our biggest gap is only 5 Km so the stuff we 
use might not help. 

We use Tranzeo 5Ghz point-to-point access points, these require line of 
sight between the two locations.  They might cover 12 Km but I'm not 
sure.  Equipment is about £450 and you'll need to pay someone to install 
it on the roof.

The other method is to rent a "private circuit" or EPS line from BT - a 
private wire between two premises sharing the same exchange -  over 
which you can connect a pair of sdsl modems.  These may not work over 
the distance, again I'm not sure.  Costs about £45/month.  Maybe that's 
what you are already doing.

Cheers

Ian
http://www.iasmith.com
http://www.festivalpreviews.com


ChrisAtUpw@....... wrote:
> In a message dated 05/01/2008, ian@........... writes:
>
>     I'm squirting the data across my own intranet to the graphing
>     computer, so am getting most of the available 100 Mb/s bandwidth
>     with minimal latency.
>
> Hi Ian,
>  
>     Communication links usually have fixed baud rates. What are you using?
>
>     For those using the internet, there shouldn't be a problem
>     provided the data are timestamped at source.  It then doesn't
>     matter how long the data takes to arrive or whether the times
>     between samples varies, the data can be properly reassembled using
>     the individual rimestamps.
>
>     ? If you are sending asynchronous data, you send a byte maybe ~11 
> cycles long overall, which has start and stop bits. You usually send 
> the signal, the receiver processes it and sends an ACK signal back. 
> Then you send the next byte. If you try simply sending at a fixed baud 
> rate, you inevitably get dropouts. You have to complete the process 
> with the time stamp data to be able to reassemble it. Your bus also 
> has a fixed interrupt repeat rate, when the CPU checks what tasks are 
> currently waiting. Only a few interrupts in a multitasking system 
> redirect the CPU instantly.
>  
>     Things though are more limited with 56K modems.  I'd be interested 
> to hear how there is a 12 Km "gap" in your system.
>  
>     Dead simple. This is the distance between my modem and the digital 
> receiver in the phone exchange.
>  
>     56 K modems rarely work at this rate. I limit mine to 38 K, 
> sometimes less. This avoids my computer having to request a lot of 
> data repeats, which can waste a lot of time.
>  
>     I note that the ADC board uses the computer supply lines. These 
> can be quite noisy. What noise do you generally see with the input 
> line to earth?
>  
>     How many times has your system had to use a restricted 
> amplification range channel? These are only common if you get local 
> quakes. I have yet to receive an out of range quake signal with my 16 
> bit +/-1/2 lsb system.
>  
>     Regards,
>  
>     Chris Chapman
__________________________________________________________

Public Seismic Network Mailing List (PSN-L)