Author Topic: Double-checking before ordering - my BO RED build (now with E-field!)  (Read 14869 times)

0 Members and 1 Guest are viewing this topic.

Offline Cutty Sark Sailor

  • WxElement panel
  • Forecaster
  • *****
  • Posts: 3393
    • Frankfort Weather - TwinHollies WeatherCenter
Re: Double-checking before ordering - my BO RED build (now with E-field!)
« Reply #125 on: September 08, 2014, 07:36:40 AM »
Kevin, remember to use the shortest possible CAT6 from E amp to controller!  It can be right next to the controller, unlike the H field amp! Keeps delays and power loss to a minimum for E field. Mine is on a one foot CAT6  e.g, and 100' of coax... preamp/probe is 21' above ground.
 


Offline corwyyn

  • Forecaster
  • *****
  • Posts: 328
  • Far, far east Mesa, AZ
    • Sleeping Saguaros Weather  Mesa, AZ
Re: Double-checking before ordering - my BO RED build (now with E-field!)
« Reply #126 on: September 08, 2014, 07:42:30 AM »
Ok I'll switch that out and get a shorter run for the E-field.  Guess I'll have an extra of the fifty footers if someone local needs one. I did get 100' of coax to run out to the antenna so at least that won't be a problem.
Kevin
AD5VG
Blitzortung station 1042

Offline corwyyn

  • Forecaster
  • *****
  • Posts: 328
  • Far, far east Mesa, AZ
    • Sleeping Saguaros Weather  Mesa, AZ
Re: Double-checking before ordering - my BO RED build (now with E-field!)
« Reply #127 on: September 08, 2014, 08:16:03 AM »
Mike, is there a discussion on the length of the cable from the amp to the controller?  Just wondering as the Latest version of the BO document has the following on connecting the amp to the controller:

Quote
The RJ45 modular connector is used to connect the amplifier to the controller board. You should use a shielded one-to-one CAT 5e network cable. The length of the cable can be up to 30 meters. This allows us to place the amplifier near the antennas far away from noisy environments.

So I guess I'm a bit perplexed on the need to use a shorter cable? 
Kevin
AD5VG
Blitzortung station 1042

Offline DaleReid

  • Forecaster
  • *****
  • Posts: 2002
    • Weather at Eau Claire, WI
Re: Double-checking before ordering - my BO RED build (now with E-field!)
« Reply #128 on: September 08, 2014, 08:31:15 AM »
Kevin,
Ah, the excitement.  Those of us who trod there earlier can only be a wee bit jealous with a new kit to build, locating and mounting decisions and the fun it brings, so please help us relive that vicariously through your postings.

Good luck,
Sorta fun, isn't it!

Dale
ECWx.info
&
ECWx.info/t/index.php

Offline JonathanW

  • Engineer
  • Forecaster
  • *****
  • Posts: 323
Re: Double-checking before ordering - my BO RED build (now with E-field!)
« Reply #129 on: September 08, 2014, 09:34:40 AM »
Mike, is there a discussion on the length of the cable from the amp to the controller?  Just wondering as the Latest version of the BO document has the following on connecting the amp to the controller:

Quote
The RJ45 modular connector is used to connect the amplifier to the controller board. You should use a shielded one-to-one CAT 5e network cable. The length of the cable can be up to 30 meters. This allows us to place the amplifier near the antennas far away from noisy environments.

So I guess I'm a bit perplexed on the need to use a shorter cable?

I think it's important to be clear regarding the nature of the delay concerns.

On the one hand, you have the concern that the probe (E- or H-field) is located far from the GPS antenna, thus potentially introducing locating errors due to differences between the GPS-derived signal time and the time the signal actually hit the probe.

On the E-field side, you have the additional potential problem of the signal being broken out on the E-field amp by filters into three channels.  Delay skew in the cat6 cable (delay of signals in one signal pair with respect to another) might introduce additional errors from one channel to the next.  Cat6 spec indicates 50 ns max delay skew over 100 meters of cable, which would amount to the equivalent of about 4.5 meters of TOA locating error for a 30 meter cable, or a maximum phase error of about 1 degree at 200 kHz.

Copper Cabling Standards and Technical Aspects

If the developers intend to use the differently-filtered channels on the E-field side for something besides just providing options for interference mitigation, delay skew might become a serious concern at longer cable lengths.

At any rate, personally, I'd go with the maximum cable length specs in the BO document.  Of course, shorter is always better, for keeping both delay and attenuation low (which all transmission lines suffer from).

Best,
Jonathan
« Last Edit: September 08, 2014, 09:38:24 AM by n0ym »

Offline Cutty Sark Sailor

  • WxElement panel
  • Forecaster
  • *****
  • Posts: 3393
    • Frankfort Weather - TwinHollies WeatherCenter
Re: Double-checking before ordering - my BO RED build (now with E-field!)
« Reply #130 on: September 08, 2014, 10:30:31 AM »
All this discussion on cable lengths and delays...  Let's keep it real simple for simple minded folks like me
H field... up to 10M shielded CAT 5e/6 is best. Cat6 can be longer... (some reference say either can be 25m... but experience recommends shortest possible shielded CAT5e/6. Period. Amp mounts right at antennas, minimum lead.
E field... Coax integral part of filtering. Length is infinite, so to speak, preamp to amp. Suggest 1' to 1500'. But since it's part of the filter, suggest about 25' minimum....  Virtually no delay or power loss. All taken into design. CAT cable of absolutely minimum length, up to 25M... again, shortest possible. Other major factor in E probe is height above earth... mine's 7M... try to get at least 4-6M. Significant delays are from CAT  not coax. After some operational time, the server more or less knows what your delays are!  As long as they're within some predetermined maximum tolerance, they're cool.

E field amp is virtually immune to Controller or other noise, and can be right next to controller. In design testing, there were two very rare and unusual situations, in combination, that needed correcting, but you will not face them as a combination! Well,... maybe, but extremely unlikely, and neither will exist if you've followed the cookbook!

All the figuring and thinking has been done.  You're welcome to do as you need and feel, but if you vary too much from the cookbook, down the road your signals may be rejected...
These guys know what they're doing.
Quality signal, everybody similar, and the server knows what it's about...

IMPORTANT:  Make sure your E FIELD is in the "Amp 2" controller connection!  Whether your have H field or not!  If you have NOT entered your antenna types, dimensions on your User Data Page, you should do so.  If you do not up date that info, again you might wind up being rejected a bit further down the road!  Tell the server what antennas you're running!  And you CANNOT do it with 'notes or comments' field...   ( Which I have a suggestion for, a bit later, probably over in the 'signals test tool' thread.


« Last Edit: September 08, 2014, 10:36:58 AM by Cutty Sark Sailor »
 


Offline Einar

  • Member
  • *
  • Posts: 27
Re: Double-checking before ordering - my BO RED build (now with E-field!)
« Reply #131 on: September 08, 2014, 12:23:57 PM »
Velocity factors.
The velocity factor of a transmission cable is mostly dependent upon the insulation layer(s). For the most common it is appx. 66% (of the speed of light in vacuum). This should be in the data sheet for the cable.

The velocity factor of a coax can be higher than a Cat5/Cat6 cable if you choose foam dielectric coax. But if it is solid Polyethylene it will be 66%, just as the common Cat cables.

So when delay is an issue in choosing cable type, the dielectric (inner layer of insulation) is what is of interest, not the cable type (Cat or Coax).

The velocity factor is an important parameter when we measure distance to a fault using TDR method.

Best cable choice?
Now the above is nice theory, provided the Cat cable is used as it is designed for. And that is as a twisted pair used for balanced transmission, as in an Ethernet interface. But in BO it is not. The connection between the controller and amplifiers A,B and C are not balanced. Amplifier B is the one where signal and reference (GND) is on the same pair. Amp A and B have the data lines to the AVR microcontroller on the other wire in the pair. So out the window goes any calculations based on impedance, and any calculations on delays with it. There will even be crosstalk from the serial communication to the AVR over to the signal lines. But hopefully the Tx/Rx lines to AVR will be "talking" only when a change in gain is required. Thus making crosstalk a non-issue.

This leads me to the conclusion that I would choose the Cat cable to be as short as possible, to keep these unwanted effects as small as possible. So I end up being in agreement with Cutty, although not based on the same premises.

The coax from Amp13 to the preamp have no such issues. It is powered over the center wire and braid. But this is decoupled from the signal by inductances at both ends, blocking the signal from the supply. And it will not have any influence on the signal. This is a well known method used since the last millenium, and works fine provided the right choice of components.

It is important however that you use 75Ohm coax though. And that there are no impedance breaks in it between the amp13 and preamp14. Which means that you must use 75Ohm components if there are any feedthroughs or splices. If you find any lying around from old Ethernet installations, don't use it! It is 50Ohm. Such impedance breaks can create reflections of the important pulse that is used in the calculations. This is more important the longer the coax is.

I can not see that it is part of the filtering, being 75Ohm end to end, so a short coax should work fine. Except it is a means of having the antenna (probe?) as far from disturbances as possible.

« Last Edit: September 08, 2014, 12:31:12 PM by Einar »

Offline JonathanW

  • Engineer
  • Forecaster
  • *****
  • Posts: 323
Re: Double-checking before ordering - my BO RED build (now with E-field!)
« Reply #132 on: September 08, 2014, 12:58:34 PM »
Best cable choice?
Now the above is nice theory, provided the Cat cable is used as it is designed for. And that is as a twisted pair used for balanced transmission, as in an Ethernet interface. But in BO it is not. The connection between the controller and amplifiers A,B and C are not balanced. Amplifier B is the one where signal and reference (GND) is on the same pair. Amp A and B have the data lines to the AVR microcontroller on the other wire in the pair. So out the window goes any calculations based on impedance, and any calculations on delays with it. There will even be crosstalk from the serial communication to the AVR over to the signal lines. But hopefully the Tx/Rx lines to AVR will be "talking" only when a change in gain is required. Thus making crosstalk a non-issue.

Induced signals from the data transmissions are likely quite a bit outside the bandwidth of the received signal lines.  Might see some crosstalk between received signal lines, though.

I agree with you about signal delays/velocity factor, as well as the coax impedance/termination/length.

Best,
Jonathan